Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 223 appraisals
IEEE transactions on cybernetics
Knowledge Distillation and Reinforcement Learning in a Human–Machine Collaboration Delivery System With a Robotic Arm
Robotic arms are widely used in various aspects of human-robot collaboration. The primary goal of this study is to explore the usability of robotic arms for delivering objects to humans in dynamic environments. Traditional robotic arms often face limitations in path planning, such as difficulties adapting to dynamic environments and complex developmental processes. To overcome these challenges, this study employs reinforcement learning (RL) to train four models-the Approach RL Model, Delivery RL Model, Decision RL Model, and Merged Model-as alternatives to conventional path planning control. Typically, there exists a significant discrepancy between simulated data and real-world features. Although image segmentation can substantially reduce the gap between virtual and real environments, notable differences remain in hand features. Therefore, to further bridge the simulation-to-reality gap, this study applies CycleGAN to transform real hand features into virtual hand features, thereby enhancing the model's transferability. Experimental results show that the Decision RL Model achieved an accuracy of 99.17%, while the Merged Model achieved 99.92%. The proposed method effectively improves the stability and accuracy of human-robot collaboration in complex scenarios. Overall, this study validates the feasibility of integrating RL, image segmentation, and image translation techniques, offering a scalable and efficient task-solving solution for robotic arms in highly dynamic application domains.
3 Aug 2026
Read appraisal →Cognitive, affective & behavioral neuroscience
Emotion detection unveiled: A cognitive-computational synthesis of physiological models, machine learning, and datasets.
This comprehensive survey synthesizes state-of-the-art advancements in emotion recognition based on physiological signals, specifically focusing on the paradigm shift occurring between 2021 and 2025. Crucially, we move beyond a technical review by establishing a novel Cognitive-Computational Synthesis Framework (CCSF). This framework explicitly maps multimodal physiological manifestations (e.g., electroencephalogram (EEG), electrocardiogram (ECG), and galvanic skin response (GSR)) to underlying cognitive processes, such as attentional allocation, arousal regulation, and perceptual bias, providing a theoretical foundation for explainable AI (XAI) in affective computing. We meticulously examine the transition from traditional machine learning to advanced deep learning architectures, highlighting how recent innovations in Transformers, self-supervised learning, and diffusion models have shattered previous performance plateaus. While earlier dimensional models were often limited to 70-75% accuracy, this survey details how modern architectures now achieve benchmarks exceeding 95% on seminal datasets like SEED and DREAMER. Furthermore, the survey provides a rigorous analysis of 40 key studies (identified via PRISMA protocols), evaluating them based on their validation strategies, cross-subject generalizability, and adversarial robustness. By bridging the gap between raw physiological data and cognitive theory, this work offers a strategic roadmap for the next generation of robust, interpretable, and real-time emotion recognition systems.
3 Aug 2026
Read appraisal →Computational biology and chemistry
The intervention of AI and ML generated Xenobots in the perspective of future technology: a review.
The evolution of synthetic mini-robots has contributed much to the field of science and technology. Still, of now, computational technology has shifted to generate AI-based robots from the living matter of Amphibian cells by utilizing evolutionary algorithms. In this aspect, we have discussed the evolution, role, applications, and ethical considerations of the Xenobots (the world's first living, programmable organisms) in this challenging artificial intelligence age. The various techniques involved in programming the xenobots were discussed, such as cell harvesting and culturing, computational design and simulation, assembly and bio-printing, cell differentiation and integration, machine learning and AI integration, etc. Moreover, they find their applications and promise significance in various fields, including medicine and health care. Finally,ethical considerations were discussed to emphasize safety, ensuring responsible development and deployment in various fields.
3 Aug 2026
Read appraisal →Science advances
Human-inspired time-series health evaluation with an adaptive multimodal electronic skin
Electronic skin powered with artificial intelligence could enable next-generation robotic and medical devices, yet integrating multimodal sensors and analyzing heterogeneous, multifrequency time series remain challenging. Most wearable machine learning architectures are time-invariant and trained for a specific task, limiting transfer across modalities and users. We present a multimodal electronic skin that captures diverse physiological signs with an adaptive learning framework that rapidly generalizes to unseen tasks with minimal labeled data. Our streamlined end-to-end framework uses a spectral variational autoencoder to denoise and compress multifrequency biosignals into a shared, unified second-wise latent space that preserves the spectral-temporal structure, followed by a transformer to capture temporal dependencies to support diverse downstream tasks with data-efficient learning. We demonstrate robust adaptation with 94.7% accuracy in activity recognition and 90.2% precision in fatigue assessment across various users and daily activities regardless of device and user variations, highlighting a scalable route to generalized physiological time-series analytics and human performance assessments.
3 Aug 2026
Read appraisal →CPT: pharmacometrics & systems pharmacology
Structured Schemas for Provenance-Rich, LLM-Assisted QSP Model Calibration
Quantitative systems pharmacology (QSP) models require calibration data from literature, yet manual curation is inconsistently documented and large language model (LLM) extraction can hallucinate values and fabricate citations. We present MAPLE (Model-Aware Parameterization from Literature Evidence), which uses structured validation schemas as a collaboration interface between LLMs and modelers. Two schemas span two scales: the SubmodelTarget schema for isolated experiments constraining individual parameters, and the CalibrationTarget schema for clinical and in vivo endpoints constraining the full model. Both separate data extraction from modeling decisions, recording every value with full provenance. Targeted validators catch characteristic LLM errors by matching values to source snippets, resolving DOIs, and executing code. For a pancreatic ductal adenocarcinoma QSP model, we used MAPLE to extract and curate 37 SubmodelTargets and 45 CalibrationTargets. Before any human review, the validators triggered 50 automated retries; every value carries a direct quote from its source and a verified citation; and 11 of 19 parameters are supported by more than one independent source. The LLM drafted usable forward models and code from context, while the modeler supplied the context and scientific judgment it cannot infer, revising forward-model choices in 65% of SubmodelTargets, priors in 46%, and source relevance in all files. This evaluation covers one model in one disease area, by a single group, so it characterizes the framework rather than establishing how broadly it generalizes. MAPLE records the modeler's reasoning in a form that can be re-run and independently checked, so it is not lost when the modeling team changes.
2 Aug 2026
Read appraisal →Neurological research
Federated deep learning model for epilepsy seizure detection using electroencephalogram (EEG) signal
OBJECTIVES: Epilepsy is a chronic neurological disorder characterized by recurrent seizures due to abnormal brain activity, which affects individuals' health and quality of life. Traditional seizure detection methods face challenges related to data privacy and security as well as difficulty in fully capturing both temporal and spatial relationships within the Electroencephalography signal. To address these limitations, a Federated Learning Enabled Unified Transformer model is proposed. METHODS: The Federated Learning with Paillier Homomorphic Encryption is deployed for preserving data privacy and enabling collaborative model training. Adaptive noise filtering and Independent Component Analysis are deployed to remove the noise and artifacts. The Multi-Scale Wavelet Coefficient is applied for signal decomposition that effectively extracts seizure-related features by decomposing Electroencephalography signals into multiple sub-bands. The Hybrid Graph-Based Attention Framework integrated Edge-Enhanced Graph Convolutional Networks for spatial feature extraction and Spectral Graph Attention for frequency-based feature selection, and refining feature representation for improving the classification accuracy. Further, the proposed technique utilizes a Unified Transformer model for seizure classification that efficiently captures temporal and spatial dependencies in Electroencephalography signals. RESULTS: The proposed model is validated on three datasets and attains 98.91% accuracy, 98.93% security, and 98.82% precision. The simulation outcomes indicate that the Federated Learning Enabled Unified Transformer model achieved outstanding performance when compared to existing models. DISCUSSION: The Federated Learning Enabled Unified Transformer model provided a superior performance by integrating Federated learning and hybrid Deep Learning models. It ensures that the model is more suitable for healthcare users.
2 Aug 2026
Read appraisal →Neural networks : the official journal of the International Neural Network Society
An approximation theory perspective on machine learning
A central problem in machine learning is often formulated as follows: Given a dataset [Formula: see text] , which is a sample drawn from an unknown probability distribution, the goal is to construct a functional model f such that f(x) ≈ y for any (x, y) drawn from the same distribution. Neural networks and kernel-based methods are commonly employed for this task due to their capacity for fast and parallel computation. The approximation capabilities, or expressive power, of these methods have been extensively studied over the past 35 years. In this paper, we will present examples of key ideas in this area found in the literature. We will discuss emerging trends in machine learning including the role of shallow/deep networks, approximation on manifolds, physics-informed neural surrogates, neural operators, and transformer architectures. Despite function approximation being a fundamental problem in machine learning, approximation theory does not play a central role in the theoretical foundations of the field. One unfortunate consequence of this disconnect is that it is often unclear how well trained models will generalize to unseen or unlabeled data. In this review, we examine some of the shortcomings of the current machine learning framework and explore the reasons for the gap between approximation theory and machine learning practice. We will then review some of recent work that achieves function approximation on unknown manifolds without the need to learn specific manifold features, such as the eigen-decomposition of the Laplace-Beltrami operator or atlas construction. In many machine learning problems, particularly classification tasks, the labels yj are drawn from a finite set of values. We summarize another recent paper that establishes a deep connection between signal separation problems and classification problems, proposing that classification tasks should be approached as instances of signal separation. We conclude by identifying several open research problems that warrant further investigation.
2 Aug 2026
Read appraisal →Ultrasound in medicine & biology
A Systematic Review: The Application of Attention Mechanisms in Medical Ultrasound Image Processing
Among numerous medical imaging modalities, ultrasound imaging is one of the most commonly used diagnostic methods in clinical practice. However, ultrasound diagnosis heavily relies on physician experience, and diagnostic results often lack reproducibility. In recent years, with the rapid development of artificial intelligence technology, which provides new impetus for the automated medical ultrasound image processing and analysis. Among the numerous deep learning approaches proposed, attention mechanisms have become a key component for improving network robustness to cope with challenges in ultrasound imaging, such as low contrast, blurred boundaries, and variable object morphologies. This paper systematically reviews the attention mechanisms employed in medical ultrasound image analysis, which can be roughly divided these mechanisms into three categories based on the differences in feature focus dimensions: channel attention, spatial attention, and hybrid attention. Most importantly, we not only summarized the application scenarios and effectiveness of various attention mechanisms but also analyzed the potential challenges faced in the future.
2 Aug 2026
Read appraisal →Science advances
Multispectral infrared-to-full-color upconversion expanding human vision
The human visual system is inherently blind to infrared radiation due to the insufficient energy of infrared photons to trigger the photoisomerization of the retinal chromophore, resulting in the loss of over half of the solar spectrum. Here, we report a colloidal quantum dot (CQD)-based infrared-to-visible upconverter that enables ultrasensitive upconversion of multispectral infrared light radiation as full-color visible vision. The photon-energy-selective excitonic transitions in quantum-confined states and the photon flux, respectively, enable infrared wavelength- and intensity-dependent number of photogenerated carriers. A dual-emissive-layer organic architecture with the strategically engineered hole-injection barrier could route these carriers into distinct color emission channels, resulting in correlated wavelength/intensity-to-color/luminance mapping. This paradigm shift yields a discrimination sensitivity for subtle infrared variations exceeding more than two orders of magnitude higher than conventional single-color modes, capitalizing on the human eye's intrinsic superiority in chromatic differentiation over only brightness contrast. The resulting upconverter exhibits broadband detection extending beyond 2 μm, luminance exceeding 700 cd m-2, and a photon-to-photon conversion efficiency of 3.85%. The upconverter could be applied as a lightweight, semi-transparent wearable eyeglass that projects multispectral infrared as full-color vision directly onto the retina. Besides, upconverters bound to light-sensitive proteins, potentially as an implantable retinal photoreceptor, confer innate infrared vision. By surpassing the evolutionary constraints of natural vision, this work establishes a versatile foundation for next-generation visual prosthetics and human-integrated sensory expansion.
1 Aug 2026
Read appraisal →Analytical chemistry
Wearable and Multimodal Electrochemical Hydrogel Sensor for Real-Time Non-Invasive Sweat Glucose Monitoring
Noninvasive sweat glucose monitoring is a promising strategy for real-time health management. In this study, we developed a flexible electrochemical sensor platform based on gold nanorods@polylysine-glucose oxidase (AuNRs@PLL-GOx) composite hydrogel, which enables noninvasive, highly sensitive, and multimode detection of sweat glucose. The polylysine (PLL) interfacial layer provides abundant amino groups for glucose oxidase (GOx) immobilization, improves the dispersion of gold nanorods (AuNRs) within the hydrogel matrix, and facilitates interfacial charge transport by maintaining close contact between GOx and the conductive AuNRs network. These effects improve the electron-transfer efficiency and analytical performance of the hydrogel sensor. Furthermore, the platform integrates differential pulse voltammetry (DPV), cyclic voltammetry (CV), and chronoamperometry (i-t) within a single hydrogel system for sweat glucose monitoring. Benefiting from synergistic and multitechnique detection, the sensor demonstrated excellent analytical performance, including wide detection range (up to 160 μM), low detection limit (3.71 μM), and strong anti-interference capability. These results indicate that this hydrogel-based platform is well suited for future wearable biosensors and smart healthcare systems.
30 July 2026
Read appraisal →Bioinspiration & biomimetics
Length control of pneumatic artificial muscles inspired by intrafusal-extrafusal muscle interactions
Pneumatic artificial muscles (PAMs) are widely used in pneumatic musculoskeletal robots due to their lightweight nature and high power output. On the other hand, their control typically relies on electro-pneumatic valves, electronic sensors, and peripheral circuits such as A/D converters. This increases the number of distributed electronic components and electrical interconnections, potentially leading to greater system mass as well as higher maintenance requirements, particularly in systems containing many PAMs. To overcome these challenges, inspired by the intrafusal-extrafusal muscle interactions in biological systems, this paper proposes a length control method for PAMs in which the pneumatic feedback loop is implemented mechanically, while the target length is commanded by a minimal electronic actuator. We first examined flow path conditions to suppress oscillations caused by pneumatic delays. Second, Experiments demonstrated a tracking error of2.01% for input frequencies0.20 Hz.
29 July 2026
Read appraisal →Biomedical physics & engineering express
Automated ventricular and midline segmentation in cranial ultrasound with metrology
Ultrasound imaging through sonolucent cranial implants is an emerging modality for post-neurosurgical monitoring of the adult brain, but quantitative interpretation remains challenging due to speckle, attenuation, shadowing, and the difficulty of consistently delineating thin anatomical landmarks. We present a deep learning system developed atLongeviti Neuro Solutionsfor segmenting key intracranial structures-the ipsilateral and contralateral lateral ventricles and the cranial midline-in coronal-plane adult cranial ultrasound images from patients withLongeviti ClearFit®Acoustic Brain Interface (ABI)TMimplants. The dataset comprises 457 proprietary, de-identified ultrasound frames with known pixel spacing, annotated in CVAT with ventricle and midline labels. We benchmark multiple encoder-decoder segmentation architectures and address severe class imbalance using class-weighted optimization with Dice and midline-focused focal-Tversky terms, followed by horizontal-flip test-time averaging. The best-performing configuration achieved a foreground macro Dice of 0.856 on a held-out test set, with Dice values of 0.926, 0.921, and 0.720 for the contralateral ventricle, ipsilateral ventricle, and midline, respectively. Finally, predicted masks are converted into geometry-based metrology overlays by estimating maximal perpendicular ventricle spans and ventricle-to-midline distances. These outputs provide standardized, millimeter-calibrated measurement visualizations for downstream review and future clinical validation.
28 July 2026
Read appraisal →Journal of neural engineering
A roadmap to navigate the future of neural engineering
A group of leaders in neural engineering collaborated to develop a roadmap to navigate the future of neural engineering. We covered a range of themes, including brain machine interfaces, neural modeling, artificial intelligence and machine learning, neural interfaces, neural imaging, augmented rehabilitation, and neuromaterials. For each topic we reviewed the current status, identified current and future challenges, and speculated on the emerging and necessary advances in science and technology to meet these challenges. Neural engineering will continue to yield the approaches and insights that advance the diagnosis and treatment of nervous system disorders, as well as provide new understanding of neural function.
28 July 2026
Read appraisal →Journal of neural engineering
Cortical-SSM: a deep state space model for motor imagery decoding from EEG signals
Objective.Classification of electroencephalogram (EEG) signals obtained during motor imagery (MI) has substantial application potential, including for communication assistance and rehabilitation support for patients with motor impairments. These signals remain inherently susceptible to physiological artifacts (e.g. eye blinking, swallowing), which pose persistent challenges. Although Transformer-based approaches for classifying EEG signals have been widely adopted, they often struggle to capture fine-grained dependencies within them.Approach.To overcome these limitations, we propose Cortical-SSM, a novel architecture that extends deep state space models to capture integrated dependencies of EEG signals across temporal, spatial, and frequency domains. We validated our method across two large-scale public MI EEG datasets containing more than 50 subjects.Main results.Our method outperformed baseline methods on the two benchmarks. Furthermore, visual explanations derived from our model indicate that it effectively captures neurophysiologically relevant regions of EEG signals.Significance.These results indicate that Cortical-SSM provides a robust and interpretable alternative to attention-based architectures for MI EEG decoding. By enabling physiologically grounded feature learning, our method advances the reliability of subject-independent EEG classification and supports the development of practical and clinically deployable brain-computer interface systems.
26 July 2026
Read appraisal →Scientific reports
Recognition of everyday activities using experiment data from wearable sensors: a deep learning-based framework
Tracking everyday activities is vital for detecting changes in older adults' health, allowing timely support to promote well-being. Wearable sensors and deep learning provide continuous monitoring, making them a supportive tool in detecting such changes. However, a more refined method is needed to recognise precise activities with a minimal set of sensors. This study aimed to develop a method to recognise everyday activities among older adults by utilising wearable sensors and a deep learning model. This is a small-scale home lab experiment to develop a method to recognise 14 everyday activities. We compared five models that recognised everyday activities with different sensor signal counts and accuracy. Our results showed that sensor placement is important. Based on the results, we proposed a two-sensor method (pelvis and right hand) to collect and correctly recognise everyday activities among older adults. This model, which utilises two sensors, classified 12 activities with an accuracy of 89.3%. Another model recognised all 14 activities with a lower accuracy of 88.2% using five sensors. We also explored a one-sensor approach, which showed low recognition performance and struggled to distinguish activity variability. The two-sensor-based system will allow for large-scale data collection on everyday activities of older adults.
26 July 2026
Read appraisal →Nanoscale
Advances in printable flexible and stretchable thin-film electrodes: materials, interfaces, technologies and bioelectronic applications
The mechanical mismatch between rigid clinical electrodes and soft biological tissues remains a primary bottleneck restricting the stability of long-term electrophysiological monitoring. Printable flexible thin-film electrodes offer a compelling solution by enabling additive, high-throughput patterning on flexible and stretchable substrates, thereby circumventing the reliance on vacuum environments and high-temperature processing typical of conventional microfabrication. This review synthesizes recent advances in functional inks, ranging from metals and carbon nanomaterials to conductive polymers and ionogels, together with high-resolution printing techniques. Addressing the critical challenge of interfacial failure in flexible devices, we explore engineering strategies to enhance adhesion at both electrode-substrate and electrode-tissue interfaces. Specifically, we analyze the pivotal roles of physical interlocking and chemical anchoring mechanisms in suppressing dynamic delamination and maintaining device integrity. Finally, the review highlights representative applications in wearable electronics, implantable systems, and emerging organoid interfaces, and outlines key translational challenges, including long-term stability and manufacturing reproducibility.
25 July 2026
Read appraisal →ACS applied materials & interfaces
Bioinspired Ionic Biogels for Conformal On-Skin Wearable Bioelectronics
Conductive gels are promising materials for on-skin wearable bioelectronics owing to their softness, ionic conductivity, and tissue-like interfacial characteristics. However, conventional preformed conductive gels often cannot fully accommodate complex and dynamically deforming skin surfaces, thereby limiting their interfacial adhesion and signal-transmission performance. Inspired by the reversible mechanical switching behavior of sea cucumbers, we developed a thermoresponsive ionic biogel (IBG) composed of gelatin, water, and 1-ethyl-3-methylimidazolium ethyl sulfate. In this system, gelatin forms a physical network with reversible sol-gel transition behavior, while the ionic liquid modulates the network through hydrogen bonding and electrostatic interactions and provides mobile ions for charge transport. This design integrates mechanical compliance, skin adhesion, water retention, and electrical functionality. The IBG can be directly coated onto skin in a flowable state and subsequently undergo mild in situ gelation, forming a soft, adaptive, and ion-conductive bioelectronic interface. The optimized IBG43 exhibits a low modulus, high stretchability, strong adhesion, and good environmental stability, together with stable strain sensing, reliable electrophysiological signal acquisition, and potential for self-powered sensing. This work presents a simple and effective strategy for constructing in situ formable skin-interfaced bioelectronic materials.
25 July 2026
Read appraisal →Nanotechnology
DFT and machine learning investigation of Au/Pt-decorated SnS2 monolayers for asthma and COPD diagnosis
Asthma and chronic obstructive pulmonary disease (COPD) are among the most prevalent chronic respiratory diseases worldwide, affecting hundreds of millions of people and contributing significantly to global morbidity and mortality. This work introduces a novel Au/Pt-decorated SnS2heterostructure for exhaled NO2detection, representing the new study to explore its role in lung disease diagnostics. It demonstrates ppb level NO2detection, a key biomarker for asthma and COPD, enabling early and differentiation of lung conditions by providing quantitative analysis of trace-level gases, which are often elevated in inflamed airways. While two-dimensional (2D) SnS2offers strong potential as a sensing platform, prior studies relied mainly on density functional theory (DFT) based gas sensing. Here, we present unprecedented integration of DFT and machine learning (ML) to investigate the gas sensing performance of pristine and Au/Pt-decorated SnS2monolayers. DFT analysis revealed enhanced adsorption and charge transfer upon noble-metal decoration, with Pt-SnS2showing optimal characteristics for asthma and COPD detection. Five ML models were trained on DFT and experimental-derived descriptors to rapidly predict the sensing behaviour of multiple gases, including NO2, among which XGBoost achievingR2= 0.9961. Both ML and DFT methods consistently identified NO2as the most sensitive analyte. This novel DFT-ML synergy not only validates fundamental adsorption mechanisms but also provides a scalable pathway for accelerated screening and design of high-performance gas sensors. Our findings establish a new prototype for integrating ML with first-principles simulations in the design of next-generation 2D material-based sensing devices.
24 July 2026
Read appraisal →Journal of materials chemistry. B
Multimodal health monitoring and theranostics based on functionalized hydrogels and artificial intelligence
Functionalized hydrogels are ideal flexible interfaces for multimodal health monitoring and integrated diagnosis-therapy systems, owing to their tissue-like mechanical properties, programmable biochemical functions, and hierarchical pores. However, practical applications are often limited by several material bottlenecks: mechanical fatigue and conductivity loss under cyclic stress, the mismatch between degradation rate and functional lifespan, and the trade-off between sensitivity and biocompatibility. To address these challenges, artificial intelligence (AI) has been applied to accelerate structural optimization and property prediction through molecular network engineering and inverse design. Meanwhile, during the collection of coupled mechanical and biochemical signals, these interfaces usually suffer from high background noise, data variability, and baseline drift. Machine learning and deep learning can process these complex datasets through noise filtering, automated feature extraction, and pattern recognition, enabling continuous monitoring and adaptive health management. This review summarizes the recent material design strategies of functionalized hydrogels, AI-driven data analysis methods, and their progress and challenges in integrated diagnosis and therapy.
23 July 2026
Read appraisal →Biomedical physics & engineering express
MELF: a multi-view ensemble learning framework for normative resting state EEG signal quality assessment
Normative resting-state electroencephalography (rsEEG), defined as standardized recordings from healthy populations, provides essential baselines for studying brain mechanisms and psychiatric disorders. However, its reliability critically depends on rigorous quality assessment, as artifacts or over-processing can distort subsequent analyses. Current approaches often rely on single-feature perspectives, which fail to capture the multidimensional nature of signal quality. We propose MELF, a quality assessment framework for normative rsEEG that systematically integrates three complementary feature domains: time domain (mobility, complexity), frequency domain (-band relative power, parallel log spectra index (PaLOSi)), and spatial domain (-band directed phase lag index, bad channel ratio, ICLabel-classified brain component proportion). A random forest-based ensemble classifier is constructed using dynamic weight allocation based on cross-validation accuracy (Acc). Experiments on four open rsEEG datasets (DSTR, TRCS, LEMON, and HBN) and expert-annotated clinical data demonstrate that: (1) MELF achieves up to 10% Acc improvement over single-view methods; (2) MELF outperforms traditional machine learning and deep learning baselines while reducing computational time, reaching 97.69% Acc (AUC = 0.9957), and maintains strong performance on an independent test set (92% Acc, AUC = 0.88); (3) the spectral view contributes the most, confirming the evaluative capability of the PaLOSi. MELF provides an interpretable, high-Acc, and robust automated framework for normative rsEEG quality assessment, with broad applications in neuroscience research and clinical quality control.
22 July 2026
Read appraisal →Current biology : CB
Looking to the brain to improve energy efficiency of AI
Modern artificial intelligence (AI) systems have achieved remarkable capabilities, but at an extraordinary energy cost. Training and running large-scale models can consume vast resources, posing environmental, economic, and societal challenges. In contrast, biological brains perform lifelong learning, adaptive control, and flexible reasoning using orders of magnitude less energy for learning and adaptation over a lifetime. What accounts for this difference - and how can it guide future AI development? In this review, we identify key biological principles that support energy-efficient capacities in biological brains, and consider how they might inform the design of more sustainable artificial systems. We organize our analysis around three domains: architectural constraints, signaling strategies, and learning algorithms. In each domain, we discuss concrete observations from biology, from cell to circuit to cognitive level, and describe how current and emerging AI systems mirror or diverge from these motifs. One striking feature of biological energy optimization is often overlooked: that brains are remarkably stable in their energy usage across heterogeneous modes, suggesting they may minimize energy needs during active environmental processing through maximizing the utility of 'rest-like' background processes. Overall, rather than advocating for biomimicry for its own sake, we argue for biologically informed engineering. Understanding how natural systems minimize energetic cost while maximizing flexibility may help us build AI that is not only powerful, but also efficient, equitable, and environmentally responsible.
21 July 2026
Read appraisal →Journal of neural engineering
Dynamic functional graph-Laplacian priors integrated with optimization for EEG source localization
Objective. Electroencephalography (EEG) source localization is an ill-posed inverse problem in which conventional methods often rely on static anatomical or smoothness assumptions and may neglect task-related dynamic functional interactions. This study aims to develop a dynamic graph-regularized EEG source localization framework that incorporates time-varying functional connectivity directly into inverse reconstruction and improves source-space motor imagery (MI) decoding.Approach. We propose DynaGraph-alternating direction method of multipliers (DG-ADMM), a source localization framework that combines linearly constrained minimum variance (LCMV) beamforming, region-of-interest-level dimensionality reduction, dynamic phase synchronization analysis, graph-Laplacian regularization, and efficient optimization. Initial source estimates are obtained using a LCMV beamformer. Region-level source signals are then extracted using principal component analysis and sliding-window phase-locking values with surrogate-based statistical testing are used to construct reliable dynamic functional graphs. The resulting graph Laplacian is mapped back to source space and embedded as a structured prior in an ADMMs optimization problem.Main results. Experiments on the MNE sample dataset showed that DG-ADMM produced spatially concentrated and physiologically plausible source patterns. On the PhysioNet MI dataset, the proposed framework achieved a binary left-versus-right MI classification accuracy of 93.52%, outperforming representative deep learning baselines. In a 320-dataset synthetic benchmark covering single-source, double-source, correlated-source, and dynamic-source conditions at two signal-to-noise ratios, DG-ADMM achieved the lowest mean center-of-mass localization error in five of eight conditions and showed its clearest advantage for dynamic sources.Significance. The results demonstrate that dynamic functional connectivity can serve as an informative graph-structured prior for EEG inverse reconstruction. DG-ADMM provides an interpretable and computationally feasible strategy for improving spatial focus, temporal consistency, and source-space decoding performance in EEG-based brain-computer interfaces.
21 July 2026
Read appraisal →Neuroinformatics
NeuroFusion: A Unified Framework for Generalized Visual Stimulus Decoding from fMRI Across Datasets and Subjects
Recent advancements in neural decoding have shown promising results in reconstructing visual experiences from brain activity. However, existing approaches focus primarily on decoding within a single dataset or subject, which limits generalization across various sources of neuroimaging. In this work, we propose a novel framework for the decoding of visual stimuli between subjects and between data sets, integrating neural recordings from multiple publicly available fMRI datasets. To address inherent intersubject and interdataset variability, we introduce a contrastive learning-based alignment strategy using image embeddings from a pre-trained IP-Adapter model. Our approach learns a shared latent space by aligning subject-specific neural representations with image features, enabling generalized decoding across both subjects and datasets. In addition, we propose a simple yet effective data augmentation method using ridge regression. This method synthesizes realistic fMRI-like signals from novel images by predicting voxel activity and injecting learned noise distributions, thus enhancing training diversity and model robustness. To the best of our knowledge, while several recent studies have explored cross-subject decoding, we extend recent cross-subject decoding efforts by training a single unified framework jointly across multiple public fMRI datasets and subjects, enabling cross-dataset transfer in addition to cross-subject generalization. We distinguish this multi-dataset unified training setting, where each dataset contributes training data, from a stricter leave-one-dataset-out transfer setting in which the target dataset is excluded from source pretraining and used only for lightweight alignment-layer adaptation. Empirically, our unified model achieves strong semantic reconstruction across datasets (e.g., up to 94.8% CLIP similarity on NSD (AUG) and 0.403 SSIM on BOLD5000 after lightweight finetuning), demonstrating robust cross-subject and cross-dataset transfer.
20 July 2026
Read appraisal →Science advances
Multistage microrobots with pH-responsive release of platelet membrane-coated nanoparticles
Targeted drug delivery in the gastrointestinal tract remains challenging because therapeutics must overcome multiple hierarchical barriers before reaching diseased tissue. Here, we present a multistage delivery platform that integrates magnetic microrobots, a pH-responsive protective coating, and platelet membrane-coated nanoparticles (PNPs) in one platform. A fillable design enables the formation of an internal magnetic layer for microrobot actuation, while the pH-responsive coating protects the cargo during transit and selectively degrades upon pH change, releasing cancer cell-targeting PNPs. In an in vitro colon cancer model that reproduces key gastrointestinal features, including flow, pH variation, and villi-like structures, this strategy increased nanoparticle retention and enhanced cancer cell cytotoxicity compared to nanoparticles administered alone. Ex vivo studies in porcine stomach and intestine further demonstrated robust locomotion on compliant and folded tissue surfaces. These results establish an environment-responsive hierarchical delivery strategy for more precise oral delivery in complex gastrointestinal settings.
20 July 2026
Read appraisal →Biomedical physics & engineering express
Benchtop validation of a low-cost wearable triaxial accelerometer
Wearable inertial sensors have become a cornerstone technology for detecting fall-related events, such as tripping while walking, in older adults. However, low-cost devices must undergo rigorous validation to ensure fidelity and stability in measurement. This study aimed to evaluate the concurrent validity of a newly developed, low-cost triaxial accelerometer by comparing its linear acceleration measurements with those obtained from a widely used commercial reference sensor (RS). A methodological validation study was conducted to assess the accuracy and precision of the proposed accelerometer. Triaxial linear acceleration data were simultaneously collected from the new device (sensor under test, SUT) and a commercial RS (Physilog®5, GaitUp, Lausanne, Switzerland). Measurements were obtained across three independent axes from 49 SUT units, with three repeated trials performed for each axis. The RS exhibited a more negative mean bias on the mediolateral (x) axis and a slightly larger mean bias on the vertical (y) axis. The SUT showed a greater mean bias (offset) on the anteroposterior (z) axis. Root mean squared error values were highly similar between thexandyaxes (<3.2% difference), while thez-axis presented moderately larger errors (∼9%). Mean absolute error was similar in thexandyaxes, whereas thez-axis was slightly higher. Visual waveform comparisons demonstrated strong overlap in mean time-series profiles, and near-unity correlations indicated high correspondence in signal variation. Bland-Altman analysis confirmed minimal bias and narrow limits of agreement for thexandyaxes, with greater variability observed along thezaxis, but no evidence of substantial systematic deviation. Overall, the low-cost accelerometer (∼60 USD) showed strong agreement with the commercial reference device (>1 000 USD), supporting its validity for three-dimensional linear acceleration measurement under controlled laboratory conditions. This low-cost device represents a promising solution for scalable motion monitoring and fall-related event detection.
20 July 2026
Read appraisal →Nature communications
Enhancing sit-to-stand transitions and walking efficiency in older adults with a soft robotic suit
Age-related declines in muscle strength and neuromuscular control make sit-to-stand transitions and walking progressively more difficult, compromising mobility and independence. Although wearable assistive technologies have been proposed to alleviate these challenges, few have demonstrated clear benefits in facilitating sit-to-stand movements for older adults who retain a degree of independent mobility. Here, we introduce a soft hip exosuit designed to assist both sit-to-stand transitions and walking activities. In a feasibility study involving ten older adults, the exosuit increased 1-minute sit-to-stand repetitions by an average of 1.8 and reduced the metabolic cost of walking by 13.6% compared with the unassisted condition. These improvements were achieved while preserving natural kinematics, lower-limb stability, and maintaining a strong sense of agency. Our findings demonstrate that soft exosuits can enhance sit-to-stand and walking performance in older adults while preserving biomechanical naturalness and user autonomy, highlighting their potential for practical home-integrated mobility assistance.
20 July 2026
Read appraisal →Science advances
Reduction of appearance artifacts in wearable on-skin electronics
Facial electrophysiological signals are crucial to human-machine interfaces and health care monitoring. Soft and skin-conformable electrodes enabled long-term and comfortable signal monitoring. However, the appearance of the electrodes affects the wearer's social interactions and self-identity, making daily usage difficult and leaving appearance artifacts. Here, we developed fully invisible and unperceivable on-skin electrodes free from appearance artifacts. Neither the wearer nor observers can detect the visual and tactile presence of the electrodes on the skin. The unperceivable property was confirmed with sensory experiments and physical characterizations of the film on skin. Furthermore, our invisible electrodes did not affect the psychological conditions of the wearers, which confirms the feasibility of artifact-free monitoring in daily lives. Last, we demonstrated the functionality of our electrode with successful monitoring of various facial electrophysiological signals, including electrooculogram (EOG), electromyogram (EMG), and electroencephalogram (EEG). Our fully invisible electrodes provide a promising direction in developing on-skin bioelectronics, seamlessly integrating health monitoring and human-computer interaction technologies into people's daily lives.
19 July 2026
Read appraisal →Biomedical physics & engineering express
Simulation-driven deep learning for the diagnosis of middle ear pathologies using wideband acoustic immittance
Wideband acoustic immittance (WAI) provides comprehensive frequency-dependent information for diagnosing middle ear pathologies. However, the scarcity of clinical data and complex response patterns significantly hinder automated diagnosis, particularly in data-limited scenarios. To address this issue, this study proposes a simulation-driven computer-aided diagnosis framework for WAI based on finite element (FE) modeling. Latin hypercube sampling was employed to systematically perturb key physiological parameters of the human ear FE models, generating a standardized virtual WAI dataset comprising 12 000 samples across the 0.2-6 kHz frequency range. The dataset includes four middle ear conditions: normal ear, ossicular chain discontinuity, ossicular chain fixation, and otitis media with effusion. Based on this dataset, a lightweight convolutional neural network tailored for multi-channel WAI inputs, termed WAIHybrid, was developed. It was benchmarked against traditional feature-based machine learning models and five representative deep learning architectures on simulated data and subsequently evaluated on an external clinical dataset comprising 206 ear-level WAI records. WAIHybrid achieved a macro-F1 of 96.30% and a balanced accuracy of 96.29% on an independent simulated test set. On the external clinical dataset, the corresponding values were 87.76% and 88.07%, respectively. Response-level comparisons, learned-representation analyses, and Integrated Gradients maps identified partial class-related correspondence between the simulated and clinical data, residual simulation-to-clinical discrepancy, and class-dependent channel-frequency attribution patterns. These findings support a simulation-driven proof of concept for automated WAI analysis in data-limited middle ear assessment. Further evaluation in larger, more balanced, and clinically heterogeneous cohorts is needed.
19 July 2026
Read appraisal →Nature communications
Sensory-guided human-machine joint learning accelerates the acquisition of motor imagery brain computer interface control
Brain-computer interfaces (BCIs) offer the potential to restore function and augment human capabilities. However, non-invasive electroencephalography (EEG)-based BCIs still face challenges in learning efficiency and control precision, particularly for naïve users performing complex tasks. Here, we present a sensory-guided joint learning framework that integrates human motor learning with adaptive machine learning to improve BCI training and performance. In 31 BCI-naïve participants, the framework enabled rapid skill acquisition, achieving average online discrete accuracies of 86.0% for one-dimensional (1D) and 77.5% for two-dimensional (2D) motor imagery tasks, along with continuous control accuracies of 77.5% (1D) and 66.9% (2D). Mechanistically, tactile guidance reduced user exploration and accelerated neural adaptation, while sample reweighting aligned decoder updates with human learning trajectories. By coupling reinforcement-driven neural plasticity with adaptive algorithmic optimization, this framework advances BCI training from passive calibration to active human-machine joint learning, enabling practical and scalable neural interfaces for communication and rehabilitation.
17 July 2026
Read appraisal →Scientific reports
Integrating skeleton based representations for robust yoga pose classification using deep learning models
Yoga is a popular form of exercise worldwide due to its spiritual and physical health benefits, but incorrect postures can lead to injuries. Automated yoga pose classification has therefore gained importance to reduce reliance on expert practitioners. While human pose keypoint extraction models have shown high potential in action recognition, systematic benchmarking for yoga pose recognition remains limited, as prior works often focus solely on raw images or a single pose extraction model. In this study, we introduce a curated dataset, "Yoga-16", which addresses limitations of existing datasets, and systematically evaluate three deep learning architectures-VGG16, ResNet50, and Xception-using three input modalities: direct images, MediaPipe Pose skeleton images, and YOLOv8 Pose skeleton images. Our experiments demonstrate that skeleton-based representations outperform raw image inputs, with the highest accuracy of 96.09% achieved by VGG16 with MediaPipe Pose skeleton input. Additionally, we provide interpretability analysis using Grad-CAM, offering insights into model decision-making for yoga pose classification with cross validation analysis.
17 July 2026
Read appraisal →Biomedical physics & engineering express
NurtureNest: an IoT-wearable predictive analytics framework for real-time maternal risk assessment
Pregnancy-related complications are increasing globally, necessitating timely and accurate risk prediction for effective clinical intervention. This paper presents NurtureNest, an internet of things-enabled machine learning framework for automated pregnancy risk assessment. The system integrates wearable sensor data from smartwatches with user-provided clinical parameters via a mobile application, enabling continuous remote monitoring. Historical clinical data are used to train multiple machine learning models for multi-class classification of pregnancy risk into low-, medium-, and high-risk categories. Ensemble learning techniques are independently trained and evaluated alongside conventional machine learning classifiers to assess their effectiveness in pregnancy risk prediction. Experimental results show that LightGBM achieves the highest performance with 92.86% test accuracy and 95.04% cross-validation accuracy. Model performance is validated using ROC analysis and feature importance evaluation. The proposed framework enables early risk detection and supports timely clinical decision-making, improving maternal healthcare outcomes.
17 July 2026
Read appraisal →Biosensors & bioelectronics
Functional wearable hydrogel microneedle platform for continuous ketone monitoring: Translating from rodents to humans.
Diabetic ketoacidosis (DKA) is a life-threatening complication of diabetes, driven by excessive ketone production; it is most common in type 1 diabetes. Continuous ketone monitoring (CKM) could enable earlier detection and prevention of DKA, yet no wearable CKM solution is clinically available. We present a wearable CKM patch that couples hydrogel microneedles (HMNs) for painless interstitial fluid (ISF) access with an enzymatic, electrochemical biosensor for on-patch detection of ketone bodies. The HMN patch and ketone biosensor were individually characterized and optimized in vitro, then integrated into a single platform. We validated the performance of the integrated CKM sensing patch across species: in vivo testing in healthy and diabetic rats, translation to a swine model, and a pilot evaluation in human participants. The CKM sensing patch reliably tracked dynamic ketone fluctuations in all models. To our knowledge, this is the first HMN-based biosensor tested in humans. These results establish a minimally invasive, wearable approach for continuous ketone tracking with the potential to transform outpatient DKA monitoring and early intervention.
16 July 2026
Read appraisal →NeuroImage
Editorial for the Special Issue on Harmonization Techniques for MRI
This editorial introduces the special issue on neuroimaging harmonization and situates its contributions within the broader methodological landscape of multi-site MRI analysis. As large-scale neuroimaging studies continue to aggregate data across scanners, protocols, and institutions, harmonization has become essential for reducing non-biological variability while preserving meaningful biological signals. We review the major classes of harmonization approaches, including statistical methods based on the ComBat family and deep learning methods that operate at the voxel level. We also review domain generalization strategies designed for previously unseen sites, and network-aware harmonization techniques that go beyond the voxel for connectivity and connectome data. Across these developments, several cross-cutting challenges emerge, including modality-specific performance, preservation of biological information, validation using traveling-subjects and large observational datasets, and the need for scalable, standardized, and privacy-preserving frameworks. Collectively, the articles in this special issue illustrate the rapid progress of the field and highlight that robust harmonization will be critical for enabling reproducible and generalizable discoveries in multi-site neuroimaging.
15 July 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Magnetically actuated microrobotic system for sequential treatment of biofilm
Biofilm-associated infections present a critical therapeutic challenge due to antibiotic resistance and impaired tissue healing. Here, we present a microrobotic system (MZ-8) that integrates real-time human-steered navigation with autonomous, microenvironment-responsive therapy to actively eradicate biofilms and promote tissue regeneration. This microrobotic system features a spine-inspired structure for mechanical biofilm disruption, a pH-responsive ZIF-8 coating for immunomodulatory Zn2+ release, and closed-loop actuation under second near-infrared fluorescence guidance. In a rat model of periprosthetic joint infection, MZ-8 achieved effective biofilm removal, induced a pro-regenerative immune response by polarizing macrophages toward the M2 phenotype, and significantly enhanced tissue regeneration. Transcriptomic analysis further revealed the activation of immunomodulatory pathways and upregulation of M2-associated genes, confirming the system's sequential shift from eradication to repair. Moreover, validation in a rabbit model and human knee joint confirmed its operational feasibility under clinical imaging guidance and excellent biosafety. This work establishes that integrating physical eradication, biochemical immunomodulation, and interactive control within a single system is essential for advancing from infection clearance to functional tissue restoration. Thus, it provides a therapeutic paradigm for biofilm-associated diseases and lays a foundation for future intelligent, clinically adaptive anti-infective systems.
15 July 2026
Read appraisal →JMIR cancer
Internet of Things-Enhanced Mathematical Oncology: Conceptual Framework for Adaptive Cancer Care Modeling
The Internet of Things (IoT) is transforming various industries, including health care. IoT-based systems are increasingly prevalent in consumer health applications, while intelligent or smart devices equipped with sophisticated sensors are gaining recognition for their potential to improve clinical care practice and decision-making. Cancer care is a particularly promising area for IoT applications, enabling real-time and personalized interventions. However, empirical research on the effects of IoT in this field is limited due to the complexities inherent in cancer as a dynamic disease and the paucity of IoT-generated data available for research. This presents an opportunity to apply mathematical modeling to understand the effects of IoT under various scenarios. These analytical and "in silico" mathematical approaches are instrumental with limited data. Such models support the analysis of treatment uncertainty and patient response while balancing patient preferences, clinical outcomes, and system-level constraints. Grounded in mathematical oncology and health informatics, this paper proposes a conceptual framework that integrates real-time IoT data as dynamic inputs into adaptive mathematical models to simulate cancer dynamics. By exploring applications across multiple levels of analysis, the study demonstrates how IoT-enhanced mathematical models could inform implementation and optimize oncology services, addressing a critical gap in current research.
15 July 2026
Read appraisal →PloS one
A dual-stream deep learning architecture for business impact scoring and alert escalation
Modern network monitoring systems generate massive volumes of telemetry data, yet most existing anomaly detection models fail to prioritize alerts according to their operational urgency and business impact. This limitation results in delayed incident responses and inefficient alert management in Network Operations Centers. To address this gap, this study proposes a Dual-Stream Predictive Alert Escalation Framework that integrates temporal failure pattern learning with business impact-aware alert prioritization. The proposed architecture consists of two key components: a bidirectional temporal encoder for modeling multivariate Key Performance Indicator (KPI) time-series data, and an auxiliary severity encoder that captures contextual metadata related to operational risk and service criticality. The outputs of these two learning streams are combined through an attention-based fusion mechanism, and a Business Impact Scoring (BIS) layer generates impact-weighted escalation decisions for proactive incident management. Experimental evaluations using real-world KPI datasets and the AI4I_2020 predictive maintenance dataset demonstrate the superior performance of the proposed framework compared to baseline methods such as LSTM, GRU, CNN-LSTM, and BiLSTM-VAE. On the combined multivariate KPI dataset, the model achieved a precision of 0.94, a recall of 0.92, and an F1-score of 0.93, along with a PR-AUC of 0.95 and a ROC-AUC of 0.94. Under impact-aware evaluation, the framework attained the highest Impact-Weighted F1 (IW-F1) of 0.85 and BIS accuracy of 0.88, resulting in an estimated 31.6% reduction in operational costs through earlier and more accurate escalation of critical events. The suitability of the selected datasets is justified by their complementary roles: publicly available KPI time-series datasets represent real-world network telemetry behavior, while the AI4I_2020 dataset provides structured severity and operational context, enabling joint evaluation of failure prediction accuracy, escalation timeliness, and business impact modeling. By prioritizing alerts based on impact-aware severity rather than raw anomaly scores, the proposed framework directly supports operational cost reduction through earlier mitigation and improved decision-making in network operations. The proposed approach bridges the gap between anomaly detection and intelligent alert management by incorporating business relevance into predictive modeling. This dual-stream architecture offers a scalable and proactive solution for AIOps-driven network reliability and automated service resilience.
14 July 2026
Read appraisal →Biomedical physics & engineering express
EEG-EMG spatiotemporal cross-attention fusion network for functional upper-limb movement classification
Objective. Electroencephalography (EEG) and electromyography (EMG) are widely used for decoding motor intentions, yet unimodal approaches often suffer from low robustness and limited representational information. EEG-EMG hybrid brain-computer interfaces (BCIs) can bridge cortical intention and muscular execution, but challenges remain in signal alignment, fusion modeling, and clinical generalization.Approach. To address these issues, we propose spatio-temporal cross-attention fusion (STCAFusion), a spatiotemporal cross-attention (CA) framework that integrates EEG and EMG through multi-band based dual-branch convolutional encoders and parallel temporal and spatial CA modules. This design enables detailed modeling of inter-modal correlations across both time and space. We evaluate STCAFusion on a newly collected dataset of synchronous EEG-EMG recordings from 12 subjects, where the data were acquired under two paradigms (Reaching and Lifting) designed from daily functional upper-limb activities to emphasize directional and strength control.Main results. With leave-one-run-out cross-validation, STCAFusion achieves average accuracies of 84.15% and 95.22% in the two paradigms, outperforming the strongest competing EEG-EMG fusion baselines by 3.4% in the Reaching paradigm and 1.8% in the Lifting paradigm. Visualization of learned attention weights further reveals meaningful spatiotemporal EEG-EMG coupling patterns, offering insights into neural-muscular coordination patterns relevant to rehabilitation-oriented BCI design.Significance. These results highlight the potential of CA-based multimodal physiological signal fusion in building reliable hybrid BCI and wearable devices for upper-limb control and rehabilitation.
14 July 2026
Read appraisal →ACS biomaterials science & engineering
A Review of Biomass-Derived Nanomaterials as Functional Interfaces for Advanced Electrochemical Hormone Detection
The study of high-performance sensing platforms is essential to achieve a high sensitivity for hormonal analytes, short response times, and cost-effectiveness. Nevertheless, standard analytical procedures, such as high-performance liquid chromatography, inductively coupled plasma mass spectrometry, and gas chromatography mass spectrometry (HPLC, ICP-MS, and GC-MS), are rather limited because their sensitivity is low at pM/nM concentrations, and the costs of these processes are too high. Moreover, modern sensor technology can be ill-equipped to selectively respond to intricate biological sample types, and there is no single standard used to quantify sensor signals. This review addresses these knowledge gaps by providing a comprehensive discussion on the role of biomass-derived materials as useful interfaces in electrochemical hormone sensing systems. To achieve improved results, we critically investigated the recent advances in the electrode modification process through biomass-based nanomaterials to promote electron transfer kinetics and interfacial engineering. The present article also discusses contemporary issues such as problems with stability, reproducibility, and selectivity in complex matrices. Furthermore, we highlighted some of the new approaches, such as the development of artificial intelligence, Internet of Things, and wearable point-of-care devices, to create the next generation hormone sensors. Overall, this current review is a comprehensive overview of biomass-derived materials, which have the prospect of enhancing surface modification at the advanced phase, AI integration, and the development of point-of-care diagnostics.
14 July 2026
Read appraisal →Biomacromolecules
Challenges and Vision for Standardization of Biopolymer Data Sets for Machine Learning
Machine learning (ML) is transforming materials research, yet potential for biopolymer discovery remains constrained by fragmented data and nonstandardized reporting. Biopolymers differ significantly from synthetic polymers, requiring specialized approaches to represent their biosynthetic origins, hierarchical structures, and application-specific metrics. In this Perspective, we identify three core challenges limiting biopolymer representation: information encoding, data quality, and data sharing. We describe the most pressing issues and propose commensurate approaches to address each key challenge. Recommendations include the design and adoption of biopolymer-specific fingerprinting and representation frameworks, development of hybrid human-large language model (LLM) data extraction strategies, and expanding Findable, Accessible, Interoperable, Reusable (FAIR)-compliant repositories. We propose a robust foundation to define interoperable, high-quality data sets that capture the full context of biopolymer materials. Standardized metadata, shared ontologies, and community-driven infrastructure would enable scalable, reproducible workflows and accelerate the ML-driven development of biopolymers.
14 July 2026
Read appraisal →Science advances
Transforming static interfaces into tactile channels with steerable transdermal foci
Grasping objects such as handles, mice, and phones plays a central role in human-machine interaction, yet the contact surfaces of the hand, despite their exquisite tactile sensitivity, are rarely exploited as channels for information transfer. Here, we introduce transdermal all-directional targeting (TADT), a static-contact interface design strategy that converts passive grip surfaces into active tactile displays, allowing users to grasp stably while receiving spatially and temporally evolving information on the fingertips. TADT enhances sub-actuator-pitch energy delivery by integrating a microtextured skin-coupling interface with phase-controlled vibrotactile actuation, thereby generating steerable three-dimensional tactile foci beneath the skin, enabling >200-hertz vibration and sub-actuator-pitch spatial super-resolution. Psychophysical experiments show that TADT lowers detection thresholds by 30% and reduces power consumption by 80% compared with designs without TADT. Using static grip alone, users reliably identified six directional cues and completed nonvisual indoor navigation and virtual reality interaction tasks, demonstrating a compact, energy-efficient modality for expressive tactile communication.
12 July 2026
Read appraisal →Science advances
Motor-free hip exosuit via high-output fibrous dielectric elastomer actuators
Exosuits can assist gait and reduce fatigue for both healthy and pathological populations, yet their bulky, rigid actuators (usually motors or pneumatic actuators) hinder natural, comfortable movement. Dielectric elastomer actuators (DEAs) provide a lightweight, compliant alternative, but are constrained by insufficient force, energy output, and integration challenges. Herein, we propose a motor-free hip exosuit driven by high-output fibrous DEAs, offering a previously unexplored paradigm for lower-limb assistance. We develop high-aspect-ratio fibrous DEAs that deliver high blocked stress (381.6 mN·mm-2), energy density (260 J/kg), and power density (1,664 W/kg), enabled by a dual-polar molecular design of the elastomer to overcome the intrinsic trade-offs between dielectric and mechanical properties. A Lego-like integration strategy is established to efficiently bundle fibers for force amplification. The resulting exosuit reduces the walking metabolic cost by 13.9% compared to no assistance, surpassing most hip exoskeletons. These findings advance DEAs toward practical wearable robotics for real-world human assistance.
12 July 2026
Read appraisal →PloS one
NLOS/LOS identification with LightGBM ensemble
Non-Line-of-Sight (NLOS)/Line-of-Sight (LOS) identification is crucial to accurate Ultra-Wideband (UWB) positioning. The current Machine Learning solutions to this problem have either too many parameters to tune or too simple features to input, which lead to unsatisfactory performance. To address this issue, this paper proposed a novel binary classifier called LightGBM Ensemble which integrates multiple LightGBMs in parallel with multi-scale patch extraction. The heterogeneous LightGBM ensemble architecture boosts the prediction power of individuals. The multi-scale patch extraction scheme extracts informative features from time-frequency domains. Extensive experiments on an open-source dataset were conducted to evaluate the proposed approach, which proves its superior classification performance and generalization performance with feasible complexity compared to the state-of-the-art Deep Learning and Decision Trees methods.
12 July 2026
Read appraisal →Physiological measurement
A comprehensive inference-time augmentation framework in physiological signals: application to PPG-based AF detection
Objective.Accurate classification of physiological signals in real-world deployments is challenged by sensor noise, motion artifacts, and distribution shifts between training and deployment data. Inference-time augmentation (ITA), which applies augmentations during inference rather than retraining, offers a simple, model-agnostic mechanism to improve robustness. However, ITA application to physiological signals has remained narrow in scope, relying on limited augmentation methods with fixed, unoptimized parameters. This work proposes a unified ITA framework to address that gap.Approach.The framework incorporates 13 augmentation methods spanning time-domain, amplitude-domain, frequency-domain, and artifact-injection transformations, with hyperparameters systematically optimized via Bayesian optimization. We evaluate the framework on atrial fibrillation (AF) detection from 30 s photoplethysmography (PPG) signals using two deep learning architectures, generative pre-trained transformer (GPT)-PPG (in two sizes) and ResNet, across five datasets comprising more than 400 patients and ∼9800 h of PPG recording. Two evaluation strategies are assessed: standard ITA applied to all inputs, and selective ITA applied to initially positive predictions.Main results.Standard ITA consistently improved the area under the receiver operating characteristic curve (AUROC, up to 8.5% for GPT-PPG and 0.7% for ResNet) and the area under the precision-recall curve (AUPRC, up to 10.6% for GPT-PPG and 0.8% for ResNet) across all model-dataset combinations. Selective ITA further reduced average FPR by up to 4.4% (GPT-PPG) and 1.3% (ResNet) on non-AF PPG datasets.Significance.These findings establish ITA as a practical, model-agnostic approach for improving the reliability of PPG-based AF classification in deployment settings where retraining is not feasible, with broader applicability to physiological signal analysis.
9 July 2026
Read appraisal →Nanoscale horizons
Multidimensional additive manufacturing micro/nanorobots: from elaborate design to smart cargo delivery
Micro/nanorobots have attracted much attention because of their potential to perform complex tasks with high precision under controlled actuation within the human body, such as targeted cargo delivery, lesion exploration, and minimally invasive surgery. Among the various emerging fabrication technologies, multidimensional additive manufacturing (MAM) technology can enable the design and fabrication of complex structures with multifunctional characteristics. Compared with traditional manufacturing methods, MAM significantly reduces production complexity and time while enhancing design flexibility and customization. This review provides a comprehensive overview of MAM technologies for constructing micro/nanorobots, along with their applications and associated challenges in the biomedical field. In addition, emerging MAM approaches, including 4D, 5D, and 6D printing assisted by physical intelligence, machine learning, and artificial intelligence show great potential for designing and fabricating more sophisticated and intelligent micro/nanorobots, thereby advancing their clinical translation in the near future.
8 July 2026
Read appraisal →PLoS computational biology
Predictive coding explains asymmetric connectivity in the brain: A neural network study
Seminal frameworks of predictive coding propose a hierarchy of generative modules, each attempting to infer the neural representation of the module one level below; the predictions are carried by top-down feedback projections, while the predictive error is propagated by reciprocal forward pathways. Such symmetric feedback connections support visual processing of noisy stimuli in computational models. However, neurophysiological studies have yielded evidence of asymmetric cortical feedback connections. We investigated the contribution of neural feedback in visual processing for computing grasp parameters, by utilizing convolutional neural network models that had been augmented with predictive feedback and were trained to compute grasp positions for real-world objects. After establishing an ameliorative effect of symmetric feedback on grasp detection performance when evaluated on noisy stimuli, we characterized the performance effects of asymmetric feedback, similar to that observed in the cortex. Specifically, we tested model variants extended with short-, medium-, long- and longer-range feedback connections (i) originating at the same source layer or (ii) terminating at the same target layer. We found that the performance-enhancing effect of predictive coding under adverse conditions was optimal for medium-range asymmetric feedback. Moreover, this effect was most prominent when medium-range feedback originated at a level of representational abstraction that was proximal to the input layer, in contrast to more distal layers. To conclude, our simulations show that introducing biologically realistic asymmetric predictive feedback improves model robustness to noisy visual stimuli in a neural network model optimized for grasp detection.
8 July 2026
Read appraisal →Clinical oral investigations
Predicting occlusal force distribution from intraoral scans using deep learning: a pilot study
OBJECTIVES: This pilot study investigated whether occlusal force distribution measured by T-Scan can be predicted from intraoral scan images using deep learning. METHODS: Paired intraoral scan and T-Scan images from 26 subjects were analyzed. Following quality control, data were randomly partitioned at the subject level into training (n = 18), validation (n = 4), and independent test (n = 4; 8 image pairs) sets. Five deep learning architectures (U-Net, ResNet-UNet, Vision Transformer, CNN-Attention, and Pix2Pix) were trained to predict T-Scan force maps from occlusal scan images. Structural Similarity Index (SSIM) was predefined as the primary endpoint. Nonparametric statistical comparisons (Friedman test with Holm-corrected Wilcoxon post hoc analysis) and bootstrap confidence intervals were used. RESULTS: A significant difference in SSIM was observed among models (χ² = 19.3, p = 0.00069). CNN-Attention achieved the highest structural similarity (mean SSIM 0.501 ± 0.030), significantly outperforming U-Net, Vision Transformer, and ResNet-UNet (adjusted p < 0.05), while not differing significantly from Pix2Pix. Pix2Pix demonstrated the lowest mean absolute error (0.237 ± 0.018), indicating superior pixel-level intensity approximation. Validation curves revealed distinct optimization behaviors, with adversarial training showing late-stage improvements in SSIM. CONCLUSIONS: Deep learning models approximated T-Scan occlusal force maps from intraoral scan images in this pilot dataset. These findings suggest that artificial intelligence may help estimate functional occlusal force distribution from morphological intraoral scan data. Further studies with larger datasets are required to validate clinical applicability. CLINICAL SIGNIFICANCE: Computational prediction of functional occlusal force maps from intraoral scans may represent a future step toward integrating morphological and functional data within digital dentistry workflows; however, larger validation studies are required before clinical application.
7 July 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Analytical modeling for suction cup designs for skin-interfaced wearable devices
Stable mounting is a central requirement for skin-interfaced wearable biomedical devices, because accurate and long-term measurements with clinical utility typically demand intimate contact with the skin, whereas practical use also requires gentle removal to minimize skin irritation and damage. Existing mounting strategies often struggle to satisfy these competing requirements simultaneously, especially under prolonged wear or in the presence of sweat and moisture. Suction-based mounting has recently emerged as a promising alternative because it can provide strong, reversible, and adhesive-free attachment, yet its underlying mechanics remain insufficiently understood. Here, we establish analytical models for the deformation and force of suction cups in a fully explicit form, covering both the cone suction cup and an optimized ring suction cup design. Unlike previous approaches that rely on indirect quantities such as the pressure difference and contact radius, which are not available before experiments and therefore cannot serve as controllable design variables, the present framework yields direct relations between suction performance and geometry parameters, material properties, and loading conditions, including the maximum push down displacement and the subsequent pull up displacement. The resulting formulas agree closely with accurate numerical solutions and lead to compact scaling laws that clearly identify how geometry and material parameters govern suction performance. These results provide a quantitative and physically transparent foundation for the design of suction-based mounting strategies in wearable devices.
7 July 2026
Read appraisal →Journal of theoretical biology
Likelihood-free parameter inference for spatiotemporal stochastic biological models using neural posterior estimation
Cell migration is a key biological process underlying wound healing, tissue development, and cancer metastasis, yet calibrating mathematical models of migration to experimental data remains a major challenge. Scratch and barrier assays are widely used to study collective cell spreading, and agent-based random walk models provide a natural stochastic description of these experiments. However, parameter inference for such models is hampered by intractable likelihoods, forcing researchers to rely on Approximate Bayesian Computation, which introduces biases and tuning difficulties, or surrogate models that require potentially erroneous noise model specifications. Here, we overcome these limitations using neural posterior estimation, a simulation-based inference framework that learns the full posterior distribution directly from stochastic simulations without surrogate approximations or explicit noise model specifications. We deploy this framework on four progressively complex random walk models of barrier assay experiments describing in vitro cell migration: an isotropic baseline, a model with directional bias (chemotaxis), a model with cell proliferation, and a combined model incorporating both bias and proliferation. For each model, we demonstrate inference in two settings: using one-dimensional summary statistics (column counts), and using a convolutional neural network that enables inference directly from raw two-dimensional spatial data. Neural posterior estimation performs well across all four models, recovering biologically interpretable parameters (e.g. cell motility, directional bias, proliferation rates) from cases where classical surrogate-based methods are adequate through to the combined model where the interplay of multiple mechanisms renders surrogate approximations unreliable. We validate all posteriors using simulation-based calibration diagnostics and provide an open-source implementation of our pipeline to facilitate its adoption and extension to more complex, spatially-structured biological models.
7 July 2026
Read appraisal →Biomedical physics & engineering express
A verification-oriented review of reaction-enhanced radiotherapy: linking nuclear data, transport modeling, and measurable observables
Reaction-enhanced radiotherapy (RER) aims to improve the therapeutic ratio of particle therapy by exploiting isotope-specific nuclear reactions that generate localized high-linear energy transfer secondary radiation. Absorbed dose remains the central metric for radiotherapy planning and evaluation; however, RER introduces additional verification requirements related to nuclear-reaction data, mixed-field transport, isotope delivery, and measurable secondary-field observables. This review frames RER as a verification-oriented physics problem built on three interdependent pillars: covariance-aware nuclear data, high-fidelity Monte Carlo transport, and reaction-specific secondary-field monitoring. Boron neutron capture therapy (BNCT) is treated as the clinically mature benchmark for RER, whereas PBCT and the emerging15N(p,α)12C-based proton-carbon-alpha therapy (Proton-CAT) are discussed as early-stage comparative cases. BNCT has the strongest clinical evidence base, while PBCT remains constrained by narrow resonances, low reaction yield, and limited observability. Proton-CAT provides a candidate 4.439 MeV prompt-γobservable whose verification value still requires detector-level validation and delivery/selectivity assessment. Overall, clinically credible RER requires coordinated progress in uncertainty-aware nuclear data, detector-integrated transport modeling, isotope delivery and tumor-to-normal biodistribution assessment, and high-sensitivity radiation monitoring for therapy verification.
7 July 2026
Read appraisal →Science advances
Toward autonomous robotic-assisted and microrobotic surgery
Autonomous robotic-assisted surgery (RAS) has emerged as a promising objective in biomedical technology, further enhanced by miniaturization toward microrobotic-assisted surgery (μ-RAS). This reduction in scale promises minimally invasive, partially or fully automated surgical procedures, with the potential to reduce patient recovery times, lower medical costs, and enable previously unavailable procedural options. This perspective highlights the specific advances in RAS that potentially map to the microscale (μ-RAS), organized across five surgical domains: endovascular, endoluminal, laparoscopic, ophthalmic, and orthopedic. We examine both clinical demands and technological advances in surgical robotics and identify the key innovations required for progress across these surgical fields. Our contribution is distinct in combining the perspectives of both surgical experts and bioengineering innovators, outlining a roadmap for the advancement and eventual integration of autonomous RAS and μ-RAS into mainstream surgical practice.
6 July 2026
Read appraisal →JMIR medical informatics
Graph Network Feature Space Fusion for Predicting Irregularly Sampled Medical Time-Series Data: Deep Learning Model Development and Validation Study
BACKGROUND: Irregularly sampled data, as a common data structure in the medical field, is frequently observed in emergency clinical datasets. It poses problems such as unequal sampling time intervals and frequencies, making it difficult to align the data without losing information for input into models. Meanwhile, due to information loss in the dataset, it is also difficult for the model to effectively analyze the overall data variation and predict the complex and dynamic conditions of emergency patients in the future. OBJECTIVE: This study aimed to address the impact of issues such as missing values and irregularly sampled time intervals on prediction results. This paper proposes a method that converts time-series data into a graph network structure and develops a prediction model for graph data to better support complex tasks. METHODS: In this paper, feature channels and measurement times in time-series data are constructed as nodes, and feature measurements at time points are constructed as edge weights. The point feature space and edge weight space are extracted by a graph convolutional neural network and a gated convolutional attention mechanism network, respectively, and feature fusion is carried out under the action of the learnable head. RESULTS: The proposed method was tested on 4 datasets and compared with both irregularly sampled time-series predictors and state-of-the-art methods for complete data. In addition, a dedicated analysis was conducted leveraging data from 5 key clinical treatment phases. In single-target classification and single-target regression tasks, the method proposed in this study outperforms other comparative methods in most experiments (accuracy=0.74-0.88, area under the curve=0.70-0.87, and mean absolute error=0.038-0.061). Meanwhile, this study constructs clinical phenotypes during patients' hospitalization as multitarget prediction tasks, and the proposed model also outperforms comparative methods in multitarget classification tasks (mean accuracy=0.75-0.91 and mean area under the curve=0.70-0.75). Ablation experiments verify the effectiveness of fusion between the node feature space and the edge feature space. The interpretability analysis of time points identifies key features that require focused attention for patients across different target tasks and time periods. CONCLUSIONS: The proposed model can effectively analyze medical data with missing values and uneven sampling. It achieves high prediction accuracy in both regression prediction tasks and classification tasks. The graph network-based model structure endows the model with better interpretability, enabling effective exploration of factors that trigger changes.
5 July 2026
Read appraisal →Neuroscience and biobehavioral reviews
Anxiety detection using neural and physiological signals and artificial intelligence: A comprehensive review.
Anxiety disorders are a significant challenge to global health, yet current diagnostic methods rely mainly on subjective and episodic assessment. The application of artificial intelligence (AI) in combination with neural and physiological signals is a promising pathway toward objective and continuous monitoring. This review provides a comprehensive evaluation of machine learning (ML) and deep learning (DL) techniques applied to electroencephalography (EEG), electrocardiography (ECG), Photoplethysmography (PPG), and electrodermal activity (EDA) for anxiety detection. A systematic review of the literature covering the period 2015-2025 indicates that the field has evolved from traditional ML systems based on hand-crafted features to modern end-to-end deep learning schemes. Our review indicates that while classical models remain effective, hybrid models such as CNN-LSTM and more sophisticated architectures, like Transformers, consistently deliver state-of-the-art results, particularly in multimodal data integration scenarios. Despite the impressive accuracies reported, a critical examination identifies several key challenges including reliance on narrow, laboratory-trained datasets, a lack of standardized validation procedures, and limited transparency in complex models, collectively impeding clinical translation. Progress in this field requires the creation of large-scale, clinically-validated datasets; the development of fault-tolerant, generalizable, and interpretable schemes; and a transition from basic state classification towards longitudinal, personalized monitoring in support of just-in-time intervention. As a comprehensive and up-to-date narrative synthesis of AI-driven anxiety detection using neural and physiological signals, this review consolidates a fragmented literature, articulates the translational gaps limiting real-world deployment, and establishes a clear research agenda for the development of clinically impactful digital mental health technologies.
3 July 2026
Read appraisal →Physics in medicine and biology
Deep learning-based dose prediction to enhance planning efficiency in cervical brachytherapy with hybrid applicators
Objective.This study aimed to develop a dose prediction model for hybrid applicators in cervical cancer brachytherapy.Approach.A total of 216 previous treatment plans were collected and divided into three datasets: training (161 plans), testing (40 plans), and clinical evaluation (15 plans). The treatment plans included three hybrid applicator types: tandem and ovoids combined with interstitial needles, tandem and ring combined with interstitial needles, and tandem and cylinder combined with interstitial needles. The model was validated on the testing dataset using mean absolute error (MAE) for voxel-wise dose evaluation and on the clinical evaluation dataset using dice similarity coefficient (DSC) and dose statistic differences.Main results.The model achieved a voxel-wise MAE of 0.45 ± 0.27 Gy and showed no statistically significant differences (p> 0.05) in all key dose-volume histogram parameters when compared to clinical plans. The DSC values ranged from 0.74 to 0.77. Small dosimetric differences between the optimized and manually generated plans were observed for the high-risk clinical target volume (HR-CTV) D98%, HR-CTV D90%, and bladder D2cc.Significance.The developed dose prediction model demonstrates significant potential for application in hybrid applicator brachytherapy. This model can provide guidance for dose delivery during dwell time optimization. As a reference for optimization, it has the potential to streamline the planning process, reduce planning time, and improve consistency in adaptive workflows.
3 July 2026
Read appraisal →Journal of the American Medical Informatics Association : JAMIA
Engineering biomarker representations of vital signs data enhances deep learning mortality prediction
OBJECTIVES: We evaluated bidirectional long short-term memory models for predicting inpatient mortality using different approaches to processing vital signs data collected during the initial 24 h of intensive care unit (ICU) admissions. MATERIALS AND METHODS: We compared 3 vital-sign representations: (1) raw data recorded every 5 min, (2) preprocessed data averaged hourly, and (3) preprocessed data using biomarker representations that extends a digital oximetry biomarker toolbox of PhysioZoo software, applied to blood pressure, heart rate, temperature, respiratory rate, and SpO2. RESULTS: Across 2 large ICU datasets, HiRID and eICU, models trained on the frequency-normalized representation achieved higher discrimination and lower Brier scores than those trained on raw 5-min and hourly averaged data. DISCUSSION: The use of biomarker representations of vital signs yielded the largest improvements in discrimination and overall probabilistic performance reflected by lower Brier scores for predicting inpatient mortality by deep learning. CONCLUSION: Thus, we recommend using a similar approach to vital signs preprocessing for time-series predictive models.
3 July 2026
Read appraisal →Nano letters
Wearable Self-Powered Biomedical Smart Sensors Deriving from E-Waste
Power requirements represent a critical challenge for wearable sensors. Self-powered sensing systems enabled by miniaturized energy-storage devices (MESDs) offer a promising solution. However, the proliferation of MESDs inevitably generates electronic waste (e-waste), which causes environmental concerns. Transient electronics that degrade into eco-friendly residuals provide opportunities for the development of green power sources. Herein, flexible e-waste-friendly power sources based on degradable MXene films were developed for the integration of wearable, self-powered biomedical sensors. The proposed transient MXene film-based supercapacitors (TMFSCs) possess good energy storage capability and mechanical flexibility and can be completely degraded into eco-friendly residuals within minutes. Furthermore, a wearable self-powered biomedical smart sensor was designed for real-time monitoring of pulse signals in real-life scenarios, and the obtained pulse rate is conducive to early evaluation of human health. Collectively, TMFSCs are considerably competitive for future eco-friendly flexible MESDs toward next-generation sustainable wearable and portable sensing electronics.
3 July 2026
Read appraisal →Physics in medicine and biology
Deep learning based time series analysis for breathing phase prediction in phase-gated proton therapy
Objective. Phase gating is a critical technique to mitigate tumor motion during radiotherapy, particularly in spot-scanned particle therapy where internal motion can interfere with dynamic spot scanning patterns and, simultaneously, introducing substantial range uncertainties. However, the current commercial state of the art in real-time phase prediction is challenged by patient-specific breathing variability as well as detection and delivery system latencies. This leads to suboptimal efficiency, mis-timed radiation delivery and requires frequent manual intervention. This study aims to improve phase prediction accuracy using deep learning (DL)-based time series forecasting to enable more accurate dose delivery.Approach. Retrospective breathing waveforms from 69 proton therapy patients, sampled at 30 Hz, were labeled with inspiratory peaks and assigned subjective regularity scores (four levels). DL models with various architectures were trained using waveform amplitude to predict current or future breathing phases. Model performance was evaluated using mean squared error, phase binning accuracy, and timing deviation for radiation on/off events. Bayesian optimization was used for hyperparameter tuning. Results were compared between models and against a commercial algorithm currently in clinical use.Main Results. The curated waveform dataset included 165 242 s for training, 24 057 s for validation, and 30 322 s for testing, with an additional 40 604 s from separate patients for extended validation. The long short-term memory and temporal fusion transformer models significantly outperformed the commercial algorithm, improving phase prediction accuracy by nearly 20% and reducing timing deviations across all regularity levels.Significance. DL-based time series forecasting may substantially improve breathing phase prediction accuracy over current clinically available methods, offering a more precise and reliable approach to phase-gated radiation delivery.
2 July 2026
Read appraisal →The Journal of prosthetic dentistry
A simulation-based study of 3D printing angle optimization by integrating deep learning and NSGA-III for prosthesis and retainer manufacturing
STATEMENT OF PROBLEM: Current dental 3-dimensional (3D) printing workflows lack automated tools for optimizing build orientations that preserve critical clinical surfaces. Existing slicing software programs often require manual adjustment and may place supports in functional areas. PURPOSE: The purpose of this simulation-based study was to develop and evaluate a computational framework by integrating deep learning and Non-dominated Sorting Genetic Algorithm III (NSGA-III) to optimize build orientations for dental crowns and retainers by minimizing the support area (SA) and printing height (PH). A secondary objective was to assess the effects of angular resolution and mesh simplification on optimization outcomes. MATERIAL AND METHODS: A 3-stage framework was developed using PointNet++ for cast classification, MeshSegNet for feature recognition, and NSGA-III for dual-objective optimization. Mesh simplification was applied to enhance computational efficiency. Twenty-four dental casts were tested in a repeated-measures design at 1, 5, and 10 degrees angular resolutions. Nonparametric statistical analyses were performed using the Friedman test (α=.05) with post hoc Wilcoxon signed-rank tests and Bonferroni correction. RESULTS: Mesh simplified reduced computation time by 87.0% without altering optimal orientation distributions. Reduced angular resolution significantly affected PH and SA outcomes under most conditions (P<.001), except for SA in the PH-optimized group (P=.004). Post hoc analysis confirmed the largest degradation in accuracy between 1 and 10 degrees. Compared with commercially available software programs, the proposed framework consistently avoided the placement of supports on critical surfaces. CONCLUSIONS: The proposed framework provides an automated and efficient approach to optimizing build orientations in dental 3D printing. Although simulation results demonstrated improved optimization performance, further experimental validation is required before clinical application.
2 July 2026
Read appraisal →IEEE transactions on bio-medical engineering
EMI Cancellation for Shielding-Free Ultra-Low-Field MRI
OBJECTIVE: Ultra-Low-Field Magnetic Resonance Imaging (ULF MRI) offers low cost and portability but suffers from electromagnetic interference (EMI) in unshielded environments. This study developed a deep learning-based active EMI suppression method to overcome these limitations. METHODS: Using a 68mT ULF MRI system, human body coupling was identified as a primary EMI pathway. We proposed EMIC-Net, a U-Net architecture incorporating Transformer and hybrid attention mechanisms, to learn the data-driven nonlinear mapping from sensing coil signals to radio-frequency (RF) receiver coil interference. Acquired data underwent phase and gain compensation prior to model training. The model's efficacy was validated through in vivo human brain imaging, comparing its performance with EDITER and standard CNN methods, and by assessing the impact of varying EMI coil numbers and training data volumes. RESULTS: EMIC-Net effectively suppressed complex dynamic EMI. It restored image SNR from 2.35 dB to 17.63 dB, with significant PSNR and SSIM improvements. Image quality neared shielded acquisitions and surpassed comparative methods. Three EMI coils provided optimal balance, and the model showed data efficiency, requiring a small dataset for effective training. CONCLUSION: The EMIC-Net method accurately predicts and efficiently removes EMI in unshielded ULF MRI, offering superior performance and practicality. SIGNIFICANCE: This research promotes portable, low-cost ULF MRI for primary healthcare and bedside diagnosis. It also offers insights for mitigating complex EMI issues in other RF sensing domains.
2 July 2026
Read appraisal →Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics (AIFB)
Advancing radiotherapy with deep learning: A review of dose prediction models
BACKGROUND: Artificial intelligence (AI), especially deep learning algorithms, holds significant promise in transforming radiation therapy by enabling more precise, personalized, and efficient treatment planning. In particular, deep learning models have demonstrated their utility in critical tasks such as dose prediction, helping to streamline complex processes and improve patient outcomes. METHODS: This paper provides a comprehensive review of recent research on deep learning applications for radiotherapy dose prediction, highlighting the architectures, datasets, and methodologies used. RESULTS: The review highlights the advancements in accuracy, efficiency, and robustness achieved by deep learning-based dose prediction models. These approaches help reduce treatment planning time and minimize human error, thereby improving clinical workflows. CONCLUSION: Deep learning offers promising solutions to enhance radiotherapy dose delivery. Future research should focus on developing the generalized dose prediction models, multi-institutional validation frameworks, explainable AI approaches for clinical trust, real-time adaptation capabilities, uncertainty quantification methods, and strategies for handling rare anatomical variations.
1 July 2026
Read appraisal →Langmuir : the ACS journal of surfaces and colloids
Wearable Sleep Detection System Based on Piezoelectric Signals and Convolutional Neural Network Analysis
Long-term and comfortable sleep monitoring is a core requirement for daily health management. However, existing sleep monitoring technologies fail to simultaneously achieve accurate sleep stage classification, wearing comfort, and direct monitoring of characteristic physiological signals. Here, we report a wearable sleep monitoring system (SMS) based on piezoelectric sensing and convolutional neural network (CNN) analysis. The system assesses sleep structure and quality by capturing characteristic eye movement signals and provides actionable recommendations for long-term sleep health management. The sensor employs a highly stereoregular polyacrylonitrile (PAN) composite fiber film modified with PDA@ZnO as the sensing unit, providing high sensitivity, excellent wearing comfort, and long-term durability. It conforms closely to the upper eyelid, enabling effective acquisition of direct physiological signals associated with sleep. Furthermore, we developed a CNN model for five-state sleep classification and validated it using original-recording-level 5-fold validation and a 1000-sample processed held-out test set, achieving an accuracy of 95.90% with a Cohen's Kappa coefficient of 0.949. The system efficiently presents sleep quality information through a simple and user-friendly human-computer interaction interface. The proposed system introduces a novel approach to convenient health monitoring and provides a feasible technical solution for accurate and long-term sleep health assessment in daily life.
1 July 2026
Read appraisal →Langmuir : the ACS journal of surfaces and colloids
An Ultrasoft Initiator-Free Eutectogel for Strain Sensing and Gesture Recognition Assisted by Machine Learning
With the rapid development of Internet of Things and artificial intelligence technologies, flexible wearable sensors have shown great potential in human-machine interaction and health monitoring fields. However, traditional hydrogel sensors face challenges such as water loss, freezing, and the use of toxic initiators during the preparation process, which lead to biological safety issues. To address these challenges, this paper proposes a green, initiator-free polymerization strategy based on the deep eutectic solvent system composed of choline chloride (ChCl) and D-sorbitol. By utilizing the property that the nitrogen-containing quaternary ammonium group in the ChCl molecule can generate free radicals upon ultraviolet irradiation, this study achieves the rapid polymerization of acrylamide in an initiator-free manner. The prepared eutectogel exhibits high transparency (≈96%), skin-fitting elastic modulus, good antifreezing performance, suitable breathability, and broad-spectrum adhesion. The flexible strain sensor constructed based on the eutectogel has high sensitivity, wide detection range, and good fatigue resistance. Combined with machine learning algorithms, this sensor system achieves a high accuracy (98.5%) of recognizing Curwen gestures. This research provides an innovative approach for developing safe, reliable, and environmentally adaptable intelligent wearable devices and has broad application prospects in intelligent music education and modern human-machine interaction.
1 July 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Mycoelectronics: Bioprinted living fungal bioelectronics for artificial sensation
The intelligence of the human biological system is enabled by the highly distributed sensing receptors on soft skin that can distinguish various stimulations or environmental cues, thus establishing the fundamental logic of sensing and physiological regulation or response. To replicate biological perception, biohybrid systems integrating living organisms with electronics have been developed to sense environmental cues. However, current eukaryote-based biohybrids face slow growth, strict culture needs, and short lifespans, limiting real-world use. Here, we introduce fungi-based printable "Mycoelectronics" which are created by additive bioprinting of living fungal mycelium networks onto stretchable electronics, as a practical living thermoresponsive sensory platform. This mycoelectronics approach leverages fung's capabilities for rapid biological responsiveness, cultivability with exponential growth, stability and self-healing in ambient conditions, bioprintability for scalable manufacturing, and mechanical flexibility for seamless integration with soft electronics. We show that the thermal responsiveness of the fungal network arises from intrinsic cellular processes-specifically, heat-induced vacuole remodeling and fusion, which modulate ionic transport and thus the electrical conductivity of the mycelial cells and networks, enabling a rapid response. By bridging the gap between cell biology and soft electronics, the mycoelectronics device, with a living mycelial network, functions as a thermal sensation system with rapid response and intrinsic self-healing properties, autonomously restoring sensing capabilities after damage and establishing sensing pathways in hard-to-reach locations. Application demonstrations in environmental and agricultural monitoring and wearable sensing systems for humans and robots highlight the versatility of this living fungal sensor platform, suggesting promising opportunities in healthcare and the environment.
30 June 2026
Read appraisal →Brain topography
Local Field Potential Recordings Using Deep Brain Stimulation: A Practical Workflow and Open-Source Signal Processing Pipeline
Deep brain stimulation (DBS) systems with electrophysiological recording capabilities offer a unique opportunity to chronically record local field potentials (LFP) from implanted electrodes, enabling real-time monitoring of subcortical neural dynamics and the development of adaptive neuromodulation strategies. However, translating raw LFP into clinically relevant biomarkers remains challenging due to numerous technical considerations. To address this, we share both a practical guide for conducting LFP experiments and BrainWave-DBS, an open-source pipeline that implements the workflow and supports analysis of patient-derived LFP data. The pipeline covers data import, preprocessing, spectral analysis, and visualization, offering a reproducible framework for LFP analysis. Furthermore, we demonstrate the functionality of the pipeline using a representative Parkinson's disease dataset and highlight key technical insights gained through this work. By improving access to neural signal acquisition and analysis, our goal is to foster broader adoption of this technology, which could advance closed-loop neuromodulation of deep brain structures.
29 June 2026
Read appraisal →Biomedical physics & engineering express
Challenges and strategies in fusion of probes and imaging equipment for neural imaging
Recent years have seen significant progress in imaging technology, which has deepened our understanding of the brain. However, challenges remain in detecting nanoscale structures like the extracellular space and mapping complex neural networks. Various chemical probes have been developed to improve the imaging precision of microstructures and functional imaging. Yet, the integration of imaging technologies and probes stays underdeveloped. Differences in disciplinary paradigms, unclear quantitative and qualitative relationships, and a lack of optimized cross-functional orchestration are the main causes of this problem. In this paper, we propose potential strategies to achieve deep integration of probes and equipment for high-quality imaging. Specifically, we advocate for three main approaches: (1) designing imaging sequences driven by probe properties, (2) constructing quantitative dose-effect models, and (3) leveraging artificial intelligence for system optimization. Importantly, these strategies establish an adaptable framework that will facilitate the seamless integration of emerging probe and equipment technologies in future brain science research. Ultimately, this enhanced integration promises to advance our knowledge of brain function and structure.
29 June 2026
Read appraisal →Biomedical physics & engineering express
Transformer imputation in CTG: a length-dependent evaluation of reconstruction methods
Objective.Computer and artificial intelligence (AI) analyses are being increasingly used in intrapartum cardiotocography (CTG). However, fetal heart rate (FHR) signal loss, which frequently occurs in clinical practice, hinders visual interpretation and reduces accuracy. Although the impact is well recognized, there is no consensus on the maximum continuous gap length that can be reliably reconstructed under clinical conditions. Therefore, we aim to identify suitable imputation methods and clarify the clinical limits of valid missing-segment lengths.Methods.Using an open FHR dataset (CTU-UHB), we extracted continuous segments and artificially introduced Removed data of varying lengths. Using performance metrics such as difference and similarity, we compared the performance among a Transformer-based model and linear and spline interpolation. Additionally, we quantified the similarity between the Pre-impute and removed data to assess task difficulty.Results.We analyzed 2727 segments from 552 cases across multiple gap lengths. In terms of numerical accuracy root mean square error (RMSE), spline consistently performed significantly worse than others. The Transformer generally maintained a better mean accuracy than linear interpolation, although significant differences were observed only under specific conditions. Conversely, for waveform preservation (correlation), the Transformer consistently outperformed linear interpolation. Notably, in highly complex imputation tasks, the Transformer proved most robust, yielding the lowest RMSE and highest correlation. However, performance systematically degraded for all methods as gap lengths increased.Conclusion.The Transformer provides an effective baseline for FHR imputation under clinical conditions, achieving a favorable balance between waveform and numerical accuracy. By clarifying the clinical limits of valid missing-segment lengths-specifically the decline in reliability beyond 30 s-this study provides guidance for standardizing preprocessing in future CTG AI research and clinical implementation.Significance.For imputing intrapartum FHR data, the Transformer generally improves waveform reproducibility over linear interpolation for short-to-moderate gaps. Defining the reliability limit provides a crucial baseline for future CTG AI.
29 June 2026
Read appraisal →ACS sensors
Lignin-Regulated Amphiphilic Janus Membrane for Eco-Friendly Electronic Skin
Eco-friendly electronic skin fabricated from sustainable natural materials represents a pivotal technology for advancing healthcare and sports-oriented wearable electronics, yet balancing comfort and electronic functionality remains challenging due to inadequate moisture management. Herein, we report an eco-friendly lignin-regulated amphiphilic Janus membrane (LAJM)-based pressure sensor that synergistically integrates unidirectional water transport with exceptional sensing performance. Through sulfonation and fluorination modifications of lignin fractions with distinct hydrophilic and hydrophobic properties, we engineer an electrospun layer with controlled hydrophobic-hydrophilic gradients, enabling unidirectional water transport while maintaining a dry and comfortable microenvironment. Moreover, LAJM demonstrates notable biocompatibility, antibacterial efficacy, and antioxidant properties, positioning it as a promising candidate for intelligent wound care systems. The LAJM-based pressure sensor achieves integrated sensing capabilities through its hierarchical structure, exhibiting a high sensitivity of 23.97 kPa-1 in the 0-20 kPa range, an extended operational range (0-200 kPa), and ultrafast response dynamics (22 ms response/31 ms recovery). These metrics enable robust continuous monitoring of vital signs (e.g., radial pulse waveforms), discrimination of voice patterns, and precise motion trajectory mapping under dynamic conditions. This technological breakthrough in lignin valorization establishes a scalable platform for next-generation breathable epidermal electronics, with transformative implications for AI-driven healthcare diagnostics and adaptive human-machine interfaces.
28 June 2026
Read appraisal →PloS one
An art style classification network integrating contrastive learning and counterfactual attention
Art style classification is a fundamental task for digital art analysis and intelligent cultural heritage management. However, existing methods often fail to deliver stable and interpretable results due to dispersed style cues, complex local textures, strong interference from background and subject content, and the difficulty of identifying regions that are causally relevant to style recognition. To address these issues, we propose a Multi Source Collaborative Style Network, termed MCS-Net, which jointly models global composition and local brushstroke textures within a unified framework. MCS-Net consists of four modules, namely a basic style feature encoding module, an attention generation module, a style contrastive learning module, and a counterfactual attention module. The attention generation module discovers multiple complementary discriminative regions. The contrastive learning module constructs a more discriminative embedding space to improve the separability of fine-grained styles. The counterfactual attention module builds counterfactual attention samples to explicitly estimate the true contribution of regions to predictions, thereby suppressing spurious activations, improving robustness, and providing counterfactual-supported evidence for model interpretation. Experiments on three public datasets, WikiArt, MultitaskPainting100k, and Pandora18k, demonstrate that MCS-Net outperforms representative baselines across standard evaluation metrics.
28 June 2026
Read appraisal →PloS one
Hybrid deep learning for mental workload classification using EEG with enhanced preprocessing and interpretability
Mental workload classification is critical in safety-sensitive fields such as healthcare and aviation. However, electroencephalography-based approaches still face challenges with generalizability, noise robustness, and interpretability. In this study, we propose an integrated hybrid deep learning framework to address these limitations and enable robust, interpretable electroencephalography-based mental workload classification. The proposed approach uses a Variational Autoencoder to enhance noise reduction and feature extraction from band-wise topographical videos, a Convolutional Block Attention Module to adaptively focus on important spatial-channel Electroencephalogram features, and a Bidirectional Long Short-Term Memory network to capture complex temporal dependencies under leave-one-subject-out cross-validation. We conducted ablation studies to identify each architecture component's contribution and sensitivity analyses to determine the optimal parameters. The model achieved the highest overall accuracy among the baselines and reached an average accuracy of 83.9% across subjects for classifying four mental workload levels. Ablation studies confirmed the added value of all three architecture components for improving performance. Sensitivity analyses identified optimal parameters, including a 10-second window length that balances temporal context and specificity. For better interpretability and neurophysiological insight, we used Gradient-weighted Class Activation Mapping to visualize key frontal-parietal brain regions and frequency bands associated with workload dynamics. Future research could explore adaptive windowing strategies, multimodal data integration, cross-dataset benchmarking, and evaluation against transformer-based and graph neural network architectures under consistent subject-independent evaluation settings to further enhance model generalizability.
28 June 2026
Read appraisal →Biomedical physics & engineering express
Intelligent integrated system for dysphagia rehabilitation and assessment
Current dysphagia assessment suffers from subjectivity and reliance on complex clinical tools, while rehabilitation lacks objective feedback. The critical mechanism of respiration-swallowing coordination (RSC) for preventing aspiration is often overlooked in existing monitoring systems. This study aimed to develop and validate an intelligent integrated system for dysphagia rehabilitation and that provides a closed-loop solution for quantitative assessment and personalized rehabilitation. The system's core innovation is the tight integration of a wireless wearable hardware subsystem for synchronous swallowing sound and respiratory waveform acquisition with an intelligent software subsystem. The software incorporates novel algorithms for automated swallowing identification, respiratory cycle analysis, and real-time classification of RSC patterns into aspiration risk levels. This enables an interactive platform that closes the loop between assessment, visual feedback, and personalized audiovisual guidance for rehabilitation training. Validation included hardware performance tests and a clinical trial with 16 participants (8 healthy participants and 8 participants with dysphagia). The hardware demonstrated reliable performance, achieving a robust swallowing sound SNR (14.07 dB in a noisy environment) and showing acceptable temporal consistency in the respiratory waveform. On a 0-100 scale, the system significantly differentiated swallowing function between healthy (82.72) and dysphagic (57.25) subjects (p< 0.01). We also conducted an exploratory experiment in two dysphagic participants using audiovisual guided training. A preliminary upward trend in swallowing-coordination scores was observed after training, suggesting a potential short-term training effect of the audiovisual guidance module. The integrated system represents an innovation by seamlessly merging wearable sensing with intelligent, coordination-focused algorithms into a unified closed-loop platform. It offers a practical and effective solution for transforming dysphagia management towards objective, personalized, and intelligent rehabilitation.
27 June 2026
Read appraisal →Analytical methods : advancing methods and applications
Microfluidic-electrochemical sweat patches for athletes: monitoring lactate, electrolytes, and hydration status
Wearable biochemical sensing is shifting sports physiology from intermittent laboratory sampling toward continuous, in situ readouts during training and competition. This review analyzes skin conformal systems that integrate capillary driven fluid routing, flexible sensor arrays, and wireless readout to capture dynamic changes in eccrine secretions relevant to performance management. We compare microchannel layouts that enable controlled filling, time resolved sampling, and evaporation suppression, and we discuss how material selection governs comfort and analytical fidelity, highlighting tradeoffs among polydimethylsiloxane, polyurethane, and polyethylene terephthalate substrates and their compatibility with scalable fabrication. Manufacturing pathways are assessed from soft lithography to laser cutting, three dimensional printing, and roll to roll processing for high throughput multilayer assembly. For transduction, we summarize enzymatic amperometric schemes based on lactate oxidase with hydrogen peroxide detection, including sensitivity, response time, and oxygen dependence, and we contrast these with emerging non enzymatic catalysts. We then detail potentiometric ion selective electrodes for sodium and potassium, focusing on ion selective membrane chemistry, solid state reference electrodes, Nernstian response, and dominant error sources such as drift and biofouling. System integration challenges, including chemical and electrical cross talk in multiplexed layouts, are linked to microfluidic isolation strategies and multiplexed electronics. Finally, we appraise validation practice, emphasizing the debated sweat to blood relationship, the need for synchronized comparative protocols, and the role of data analytics and machine learning for personalization, drift compensation, and prediction of thermoregulatory strain. Remaining barriers include long term stability, adhesion under motion, manufacturability, and regulatory evidence requirements in real world settings. Across these topics, the review emphasizes technology-specific limitations, engineering translation metrics, and realistic breakthrough directions rather than treating wearable sweat patches as a mature plug-and-play platform.
27 June 2026
Read appraisal →Biofabrication
3D printed trichome-inspired permeable bioadhesive for wearable bioelectronics
Wearable bioelectronics that adhere directly to the skin have broad applications. However, achieving optimal breathability remains a significant challenge because of sweat accumulation at the device-skin interface. Conventional approaches, such as porous structures, often limit the functional versatility of wearable bioelectronics. To address this gap, we propose a sweat-removable skin sticker (SRSS) with a hierarchical trichome-inspired channel architecture that rapidly removes sweat from the interface while maintaining robust skin adhesion. The SRSS was fabricated through a hybrid process, in which the trichome-inspired hierarchical microchannel architecture was created by direct ink writing, enabling the controlled deposition of viscous ink with high structural fidelity. Through the multi-level ribs design, the SRSS demonstrated an effective area-normalized horizontal water removal rate (25.6 ml cm-2min-1) significantly higher than the human sweat secretion rate (0.38-2.85 × 10-3ml cm-2min-1)-approximately three orders of magnitude. This feature reduces sweat accumulation in wearable bioelectronics, thereby enhancing user comfort. Unlike conventional porous materials, the SRSS relies on channel based adhesive interface design that remains compatible with attached wearable bioelectronics, such as temperature sensors in real time. This work therefore presents a structurally engineered permeable bioadhesive interface for wearable bioelectronics.
27 June 2026
Read appraisal →Scientific reports
Cross-domain transfer learning strategy enhances interpretability of deep learning model explanations
Clinical decision-making increasingly relies on deep neural networks (DNNs), yet their deployment in practice requires transparent and interpretable predictions. Explainable artificial intelligence (xAI) methods can identify input regions relevant to a model's decision, but their clinical interpretability remains limited. In this study, we investigated whether inductive transfer learning (TL) can reinforce domain-specific feature separation in xECGArch, a two-branch convolutional neural network for atrial fibrillation (AF) detection from electrocardiograms (ECGs). Each branch was pre-trained on a task aligned with its designated feature domain, P wave detection for morphology and RR interval variability prediction for rhythm, then fine-tuned on binary AF classification using an iterative layer freezing schedule. Deep Taylor decomposition (DTD) was applied to analyze explanations across all configurations. Fine-tuning accuracy ranged from 85.70% to 95.23%, remaining comparable to the original xECGArch architecture and previous TL-based approaches. However, DTD analysis demonstrated that morphology pre-training directed relevance toward P waves, whereas rhythm pre-training concentrated explanations on R peaks, with domain specificity increasing as more layers were frozen. These findings suggest that inductive TL can encourage domain-specific feature attribution in DNNs, improving the alignment of post-hoc explanations with clinically meaningful ECG regions.
26 June 2026
Read appraisal →ACS applied materials & interfaces
Biodegradable Electrohydraulic Actuators for Wearable Haptic Interfaces
Biodegradable soft actuators have gained significant attention due to their inherent conformability, adaptability, and ability to safely interact with humans, while substantially reducing long-term environmental waste. Although various biodegradable materials have been explored for fabricating soft devices and their electrical interfaces, many still suffer from limited durability and operational stability under ambient conditions. In particular, biodegradable electrohydraulic actuators often exhibit a performance degradation over time. In this work, we systematically evaluate the influence of various biodegradable materials, including compliant electrodes, dielectric films, and dielectric fluids, on actuator performance. The optimal material combination was identified and used to fabricate a biodegradable actuator that demonstrated stable operation over 10,000 cycles with no noticeable variation in actuation strain. Under industrial composting conditions, the device underwent degradation within 50 days, supporting its environmental sustainability. Furthermore, a wrist-worn biodegradable haptic device is demonstrated that is capable of generating diverse and salient tactile sensations, highlighting its potential applications in assistive technology and virtual/augmented reality.
25 June 2026
Read appraisal →JMIR medical informatics
Computational Insights Into Smart Bioelectronics in Digital Health Care (2020-2024): Topic Modeling Study
BACKGROUND: Smart bioelectronics are electronic medical devices that combine hardware and artificial intelligence (AI)-based software. These convergent medical devices analyze bio-signals measured through hardware using AI algorithms and deliver physical stimulation to enhance therapeutic effects. OBJECTIVE: This study aimed to systematically analyze recent research trends in smart bioelectronics to understand their evolving role in digital health care and to provide evidence-based insights for shaping future research and development strategies. METHODS: A total of 92 publications indexed in PubMed between 2020 and 2024 were analyzed. Latent Dirichlet allocation-based topic modeling, optimized using coherence scores, was applied to identify latent research themes. RESULTS: The results indicate a steady increase in related research over the past 5 years, along with a clear shift in research focus from bio-signal sensing and bioelectronic device materials toward AI-driven analysis and disease-oriented applications, ultimately evolving into intelligent and adaptive bioelectronic therapeutic systems. Three major research topics were identified: bio-signal-based neuromodulation (n=23, 25%), AI-driven neurological disease analysis (n=32, 34.7%), and implantable bioelectronics and biomaterials (n=37, 40.2%). CONCLUSIONS: By mapping the evolving landscape of smart bioelectronics, this study provides valuable insights into their multidisciplinary development and highlights their potential applications in clinical decision support, personalized rehabilitation, and next-generation medical device innovation.
24 June 2026
Read appraisal →PloS one
Applications of electromyography in Amyotrophic Lateral Sclerosis: A systematic review
This systematic review examined the use of surface electromyography (sEMG) for the neuromuscular assessment of individuals with Amyotrophic Lateral Sclerosis (ALS), focusing on clinical parameters, the muscle groups evaluated, acquisition protocols, technical properties of the recording systems, integration with other technologies, and signal processing strategies. We included observational studies that applied sEMG to individuals diagnosed with ALS, with or without comparison to healthy controls, and without restrictions on publication year. The analyses included signals recorded at rest and during voluntary contractions, with or without the use of biofeedback. Most studies employed conventional or high-density surface electrodes, with sampling frequencies ranging from 500 Hz to 3000 Hz. The results showed that the primary parameters assessed were muscle fatigue, fasciculation patterns, the number of motor units (MUNE/MUNIX), motor unit firing rates, and signal complexity. These parameters demonstrated sensitivity to disease progression and may contribute to early diagnosis, phenotypic stratification, and functional monitoring of ALS. Additionally, the studies highlighted the increasing use of advanced computational approaches, such as machine learning, for feature extraction and automated classification. In conclusion, sEMG is a promising tool for functional assessment in ALS, with the potential to improve diagnostic accuracy and support new therapeutic strategies based on electrophysiological biomarkers. However, despite technological advances, the included studies displayed substantial methodological heterogeneity and limited protocol standardization. Integration with other neurophysiological modalities also remains underexplored, despite its significant clinical potential.
23 June 2026
Read appraisal →Science advances
Artificial sparse neuron dendrites for visual information inference
In the human brain, dendrites exhibit nonlinear integration and sparse parallel processing capabilities, which can effectively perform visual tasks by integrating only a small subset of neuronal signals and play a crucial role in high-level information inference. However, conventional neuromorphic devices often ignore these important properties and require all neurons to perceive complete information. This makes it difficult to effectively replicate the efficient spatiotemporal processing capabilities of biological neuron dendrites. In this study, we present an artificial neuron dendrite array that integrates neurons, synapses, and dendrites, emulating the spatiotemporal spike integration properties of biological dendrites for precise parallel computation. Through multigate threshold regulation, the array enables parallel sparse spiking inference with random spatial distribution. This inference process forms a sparse dendritic spiking neural network (SD-SNN) that can perform compression, depth detection, and prediction. As a result, the SD-SNN achieves high-efficiency static and dynamic object processing while using only 0.5% of neuronal activity, slashing the power consumption by 98 and 65%, respectively. Our work reduces neural activity in the perception process by 99.5% while enhancing spatiotemporal computing capabilities and computational efficiency.
21 June 2026
Read appraisal →Physics in medicine and biology
Interlayer-aware postoperative facial appearance prediction in orthognathic surgery with bio-geometric guidance
Objective.Accurate prediction of postoperative facial appearance is essential for orthognathic surgical planning, yet remains challenging due to the nonlinear biomechanical coupling between bone and soft tissue. While deep learning methods offer a faster alternative to traditional biomechanical simulation, they typically map bony displacements directly to facial surface deformation, overlooking the intervening soft-tissue layers with distinct biomechanical properties through which displacement is progressively transmitted. Moreover, the many-to-one effect of regional bony movements on each soft-tissue point remains insufficiently captured.Approach.We propose a novel network that introduces an interlayer mixture-of-experts mechanism to decouple bone-to-surface deformation propagation into three biomechanically inspired proxy representations at different tissue depth levels. Since the contribution of each depth level varies across individuals and facial regions, a gated routing network adaptively weights each layer's contribution, providing a data-driven approximation of spatially heterogeneous, patient-specific deformation transmission. Additionally, a bio-geometric convolution module captures regional bony influences through elastic-weighted neighborhood aggregation.Main results.Evaluations on 88 orthognathic surgery patients demonstrate that the proposed method achieves a mean whole-face error of 1.39 mm, below the 2 mm clinical threshold, with statistically significant improvements over all baselines (p<0.05) and the highest clinician acceptance rate (94%).Significance.By incorporating bio-geometric priors into a deep learning framework, our approach enables a more physically grounded and interpretable prediction paradigm, supporting efficient and clinically reliable surgical planning.
21 June 2026
Read appraisal →Bioinspiration & biomimetics
A bioinspired hybrid robotic joint for safer physical human-robot interaction
In many robotic applications, such as humanoids, wearable robots or supernumerary limbs, there is a growing shift from rigid, traditional mechanisms toward softer, more compliant systems. This trend is driven by the need for safer physical human-robot interaction and the ability to operate in unstructured environments. In this work, we present a hybrid approach to controlling a single-degree-of-freedom robotic joint that combines a rigid frame with soft pneumatic actuators to enable both precise and versatile interaction. The joint is bioinspired, as it is powered by a couple of actuators mounted in an antagonist configuration, such as the human elbow. A key innovation is the design of pneumatic artificial muscles using thermoplastic polyurethane, which achieve high isometric force (400 N at 240 kPa), significant stretchability (80 mm), and low density(0.3gcm-3), making them competitive with state-of-the-art alternatives. Two model-free controllers were developed to independently regulate joint position and stiffness. Angle control achieved high precision (<2∘RMSE) with minimal overshoot (<1%) and fast response (rise time<1.3s). Stiffness control modulated joint compliance across a range of 0.054-0.076 Nm deg-1, with the expected trade-off of reduced workspace. A final proof-of-concept demonstrated the concurrent use of both controllers to modulate the joint's dynamic behavior in response to external disturbances. While future work will address multi-DOF coordination and wearable integration, this study represents a foundational step toward safe, adaptable robotic actuation through the combination of rigid structures and soft actuation.
21 June 2026
Read appraisal →Journal of robotic surgery
Progress and perspectives in soft robotics for surgery and rehabilitation
Soft robotic systems have become a promising technology in the biomedical field due to their higher safety, dexterity and adaptability compared to traditional rigid robotic systems. This review focuses on recent advances in soft robotic systems applied to surgical assistance and rehabilitation. In surgery, soft grippers and flexible endoscopes show better adaptability, dexterity, and miniaturization, enabling safer and more precise manipulation of delicate tissues. In rehabilitation, wearable soft devices show great potential to help patients with neurological injuries to regain movement. Key innovations in actuation technology are examined, along with recent advances in multifunctional, self-healing, and environmentally responsive materials. Meanwhile, sensing systems are evolving from unimodal sensing to multimodal fusion, self-perception, and sense-drive integration, enabling robots to sense the body state and external environment with higher accuracy and realize closed-loop control. Finally, this paper points out that soft robots still face key challenges such as material durability, biosafety, and stability during clinical translation. In the future, the focus should be on the construction of systems with self-diagnosis, self-adaptive adjustment and closed-loop control, and promote the efficient landing and personalized application of soft robots from experimental research to real medical scenarios.
21 June 2026
Read appraisal →Science advances
Bioinspired heliconical auxetic biofibers for intelligent biomechanical surveillance
Wearable bio-based fibers are emerging as platforms for biomechanical sensing and physiological interaction. However, achieving high sensitivity to subtle mechanical cues while maintaining flexibility, self-powered output, and distributed perception remains challenging. Inspired by the helical twining mechanics of climbing plant stems, we design a mechano-adaptive auxetic electronic bio-based fiber as an embodied intelligence sensing unit. The dual-modulus helical confinement, achieved by wrapping rigid aramid fiber/polydimethylsiloxane helices around a flexible collagen aggregate/waterborne polyurethane core, enables programmable deformation and a stable negative Poisson's ratio (ν = -0.47). Nonlinear intercomponent coupling facilitates synergistic axial-radial dynamics, amplifying interfacial contact variation under strain. The optimized fiber exhibits ultrahigh sensitivity (strain factor: 11.75), strong electrical output (8.1 volts), and remarkable power density (8.48 milliwatts per square meter) with excellent cyclic stability. Integrated into fiber-sensing arrays, it decodes microstrain patterns linked to lower-limb function, offering a scalable strategy for self-powered, adaptive, and durable biomechanical monitoring.
21 June 2026
Read appraisal →Journal of cardiovascular translational research
A Systematic Review of Modeling Platforms for Atrioventricular Valves in Atrioventricular Septal Defects
Given the high rate of re-intervention in patients with atrioventricular septal defects (AVSD), there is a continuing need to accurately model the complex valve anatomy and surgical repair strategies. This systematic review examined the current scientific landscape for modeling platforms designed to study abnormal congenital heart valves that are used or could be used for AVSD investigation. Of 1,050 sources screened, 47 were included (14 in silico; 26 in vitro; 7 in vivo). In silico models studying valve pathologies have advanced in recent years, but few targeted AVSD valves. In vitro platforms allow comparisons of repair techniques, but have yet to be applied for AVSD valve repair. In vivo studies could offer physiologically accurate platforms, but face challenges such as replicating AVSD pathologies. Multidisciplinary approaches to creating high-fidelity, reproducible, physiologically accurate AVSD valve models can transform the understanding of AVSD valve properties and repair strategies.
21 June 2026
Read appraisal →Urolithiasis
A validated custom pipeline for three-dimensional kidney stone renderings to create an open access repository
Three-dimensional (3D) rendering of urologic pathology plays an important role in simulation-based education, surgical training, and computer vision research; however, a standardized, open-access repository of high-fidelity kidney stone models stratified by chemical composition is lacking. We developed and validated a reproducible photogrammetry-based pipeline to generate realistic 3D kidney stone renderings. Chemically characterized human stones composed of calcium oxalate monohydrate (COM) (n = 11), uric acid (UA) (n = 5), cystine (n = 4), magnesium ammonium phosphate hexahydrate/carbonate apatite (MAPH/CA) (n = 2), and calcium hydrogen phosphate dihydrate (CHPD) (n = 3) were photographed using a custom-built rotating stage and dual fixed 4 K cameras. Rendered models were sent to 25 endourologists using a 5-point Likert-scale survey assessing geometric and surface texture fidelity. Successful 3D renderings were obtained for 8/11 COM stones, 5/5 UA stones, 2/2 MAPH/CA fragments, and 3/3 CHPD fragments, while all cystine stones failed to render. Across stone types, mean fidelity scores were highest for UA and COM stones (mean 3.8-3.9), intermediate for calcium phosphate stones (mean 3.6-3.8), and lowest for struvite stones (mean 3.0-3.3). Geometry scores were higher than texture scores overall, though this difference was not significant. Significant differences in geometric fidelity were observed across stone compositions (χ² = 9.30, p = 0.026). Inter-rater reliability was poor for individual evaluators (ICC = 0.10) but moderate for aggregated mean ratings (ICC = 0.67). This validated workflow enables the creation of generally realistic, open-access 3D kidney stone models (github.com/uro-glidar/3d-rendering-diverse-stones) for simulation, education, and future machine learning applications in endourology.
21 June 2026
Read appraisal →Progress in biomedical engineering (Bristol, England)
Assessing the relevance of biosignal-controlled robotic rehabilitation technologies: a systematic review.
Technology is transforming rehabilitative medicine by enhancing accessibility and personalisation. Robot-assisted rehabilitation uses robotic systems for recovery from physical and neurological impairments, enabling intensive, repetitive training with real-time feedback. These systems aim to restore motor function and mobility and to increase independence in daily life, needs that are growing in light of population ageing. Rather than revisiting control algorithms or mechanical design, this review aims to investigate the combined use of bio-signals and robotic rehabilitation systems, and to discuss their potential in rehabilitation settings outside disease-specific clinical frameworks. We reviewed studies published from 2020 through early 2025 to capture recent advances in this rapidly evolving field. Our objective was to assess the relevance of these technologies to rehabilitation outcomes, such as improvements in motor function or other clinical metrics. We considered robot-assisted systems driven by biosignals such as electromyography (EMG), electroencephalography, or electrocardiography and included only studies reporting measurable rehabilitative outcomes. Following PRISMA guidelines, we searched Scopus, PubMed, and WoS databases. Of 207 records screened, 15 met the inclusion criteria. Overall, the included studies suggest that contemporary biosignal-controlled robotic rehabilitation, particularly EMG-driven approaches, may support more intensive, personalised, and engaging therapy than conventional care alone, although the current evidence remains preliminary. Preliminary findings indicate that, by closing the loop among patient, device, and therapist, these systems may reduce muscle activity and support improvements in motor performance; however, the evidence base remains heterogeneous and generally underpowered. To support future clinical adoption, the field needs adequately sized, head-to-head trials with standardised outcome measures, longer follow-up, and transparent reporting of adherence, adverse events, user experience, and implementation barriers, including training and cost. These steps will be important to further assess efficacy and evaluate usability in real-world settings.
20 June 2026
Read appraisal →ACS nano
Learning Human-Environment Interactions via Wearable AI Interfaces
Wearable artificial intelligence (AI) interfaces are reshaping the boundaries between humans and the environment. While prior works often focus on narrow human-machine interactions, this review proposes an intact interaction information flow. It introduces a comprehensive interaction blueprint spanning local interaction and global interaction to the interaction entity, showing how humans interact with the environment. This review first examines advances in wearable form factors, sensing performance improvement strategies, and data analysis. Special emphasis is onspot on how AI interprets heterogeneous data from tactile signatures for local interaction, wearable vision for global interaction with human motion, and electrophysiological signals for the interaction entity. We then discuss the essential applications of this interaction framework, such as human-machine interaction and smart healthcare. By discussing potential barriers in device reliability, algorithm generalization, and scalable applications of wearable AI interfaces, this review provides an outlook on data-driven inverse sensor design, general intelligence strategies, and building a standard ecosystem for scalable applications. The wearable AI interfaces are toward on-body intelligence, actively perceiving, understanding, and assisting in the complex dynamic human-environment interactions.
19 June 2026
Read appraisal →Analytica chimica acta
Rapid detection of drug-resistant leukemia cell using an optofluidic chip and machine learning
BACKGROUND: Rapid detection of drug-resistant leukemia played a crucial role in formulating appropriate treatment plans for patients and improving their prognosis. In this research, an integrated optofluidic platform was developed to detect and analyze leukemia and drug-resistant leukemia cell. The detection technique was designed by embedding optical fibers coupled with photosensors and a laser source into optofluidic chip. The scattered light signals were detected when the cells pass through the detecting area. RESULTS: The platform was first validated by classifying 1 μm polystyrene microparticles and 1 μm polystyrene microparticles coated with spherical 10 nm Fe3O4 nanoparticles. After validation, the method was applied to classify leukemia and drug-resistant leukemia cells by injecting the testing sample and obtaining. The SVM classifier demonstrated the highest classification accuracy of 91.1% compared with LR, RF, and KNN classifiers for analyzing leukemia cells and drug-resistant leukemia cells. The proposed method can perform detection within 10 min with a total experimental timeframe of 20 min. SIGNIFICANCE: The presented results demonstrate the feasibility of applying microfluidics and machine learning approaches to detect and classify biological entities with slight variations based on scattered light signals. This platform holds significant potential for clinical diagnostics, offering a rapid, cost-effective, and efficient method for detecting drug-resistant leukemia cells, potentially aiding in personalized treatment strategies.
16 June 2026
Read appraisal →JMIR formative research
Automated Machine Learning Frameworks for Radiomics: Comparative Evaluation Study
BACKGROUND: Automated machine learning (AutoML) frameworks can lower technical barriers for predictive and prognostic model development in radiomics by enabling researchers without programming expertise to build models. However, their effectiveness in addressing radiomics-specific challenges remains unclear. OBJECTIVE: This study aimed to evaluate the performance, efficiency, and accessibility of general-purpose and radiomics-specific AutoML frameworks on diverse radiomics classification tasks, thereby guiding researchers and highlighting development needs for radiomics. METHODS: A total of 10 public and private radiomics datasets with varied imaging modalities (computed tomography and magnetic resonance imaging), sizes, anatomies, and end points were used. Six general-purpose and 5 radiomics-specific frameworks were tested with predefined parameters using standardized cross-validation. Evaluation metrics included area under the receiver operating characteristic curve, runtime, and qualitative aspects related to software status, accessibility, and interpretability. RESULTS: Simplatab, a radiomics-specific tool with a no-code interface, achieved the best overall balance between performance and computational efficiency, recording the highest average test area under the receiver operating characteristic curve (mean 78.46%, SD 12.22%) with a moderate runtime (1.1 h). However, its performance was not statistically superior to the most intensive general-purpose solutions. Most radiomics-specific frameworks were excluded from the performance analysis due to obsolescence, extensive programming requirements, or computational inefficiency. Conversely, general-purpose frameworks demonstrated higher accessibility and ease of implementation. CONCLUSIONS: While no single framework demonstrated absolute predictive superiority, Simplatab provides an effective balance of performance, efficiency, and accessibility for radiomics classification problems. However, continued efforts are needed to further mature AutoML solutions in the radiomics domain.
14 June 2026
Read appraisal →PloS one
Feasibility of mechanomyography-based fatigue classification for passive lower-limb exoskeleton evaluation: A pilot study
Mechanomyography (MMG) enables non-invasive monitoring of muscle mechanical activity, while its utility in time-resolved fatigue detection during dynamic human-exoskeleton interaction remains underexplored. This pilot study explored the feasibility of combining MMG with machine learning to characterize neuromuscular fatigue and evaluate passive lower-limb exoskeleton assistance during repetitive 10 kg squat-lifting tasks. MMG signals from five lower-limb muscles were extracted for time-domain, frequency-domain and nonlinear features, and fatigue identification was implemented via a spectral-based criterion and multi-muscle voting optimization. A radial basis function-enhanced random forest (RBF-RF) model integrated with data augmentation was validated through leave-one-subject-out cross-validation. The results demonstrated that the 1/5 voting rule achieved optimal performance, with mean accuracy of 0.913 ± 0.057, AUC of 0.792 ± 0.073, and a low fatigue detection error of 1.4 ± 0.8 s. Data augmentation steadily improved model robustness, and predicted fatigue levels were significantly correlated with subjective perceived exertion (ρ = 0.756, p < 0.001). This pilot study demonstrates the feasibility of MMG-based fatigue monitoring for wearable assistive systems. The proposed framework supports objective, high-temporal-resolution fatigue monitoring, and may serve as a viable tool for assessing wearable assistive systems. Further large-cohort studies are required to validate its generalizability for practical applications.
13 June 2026
Read appraisal →Nature communications
Adaptive Riemannian optimization for multi-scale diffeomorphic matching
Image matching is a fundamental task in quantitative biomedical and biological image analyses, enabling researchers to compare, integrate, and interpret imaging data across subjects, time points, modalities, and experimental conditions. Existing state-of-the-art registration methods are slow due to inefficient implementations and poor convergence rates because of the ill-conditioned nature of the optimization problem. Deep learning methods offer fast inference but require extensive training time, substantial inference memory, and fail to generalize across long-tailed distributions or diverse image modalities, necessitating costly retraining. We address these challenges by proposing FireANTs, a training-free, GPU-accelerated, multi-scale adaptive Riemannian optimization algorithm for fast and accurate dense diffeomorphic image matching. FireANTs more than doubles the speed of the community standard ANTs registration tool on a CPU, and is two orders of magnitude faster on a GPU. On the GPU, FireANTs performs competitively with deep learning methods on inference runtime while consuming up to 10 × less memory. FireANTs demonstrates robustness on a wide variety of matching problems across modalities, species, and organs, without any domain-specific training or tuning. Our framework allows hyperparameter grid search studies with less resources and time compared to traditional and deep learning registration algorithms alike.
12 June 2026
Read appraisal →Journal of neural engineering
Computational neuroelectrophysiology and artificial intelligence for drug-resistant epilepsy: recent advances, current challenges, and future directions
Drug-resistant epilepsy (DRE) affects approximately 30% of epilepsy patients, with surgical cure rates below 70%. This challenge drives a fundamental paradigm shift from localizing a discrete epileptogenic zone toward characterizing and modulating the dysfunctional brain networks that initiate and propagate seizures. This review critically synthesizes how computational neuroelectrophysiology and artificial intelligence (AI) converge to propel this shift. We chart the trajectory from quantifying local biomarkers, including high-frequency oscillations and excitation-inhibition balance metrics, to mapping the topological properties of epileptic networks through functional and effective connectivity. The integration of these network features with AI techniques, particularly spatiotemporal deep learning architectures, has demonstrated significant potential for enhancing both the localization of the epileptogenic network and the prediction of postoperative outcomes. However, the translation of this network-centric paradigm into clinical practice remains constrained by several challenges, including spatial sampling bias, the inherent label instability of surgical outcomes, and the 'black-box' interpretability crisis. Future directions must emphasize EEG foundation models, causal AI, and the development of interactive multimodal AI agents. The convergence of mechanism-driven network models and data-driven AI frameworks promises a new era of personalized, network-guided therapy for DRE.
11 June 2026
Read appraisal →Journal of controlled release : official journal of the Controlled Release Society
Toward animal-free approaches in subcutaneous drug delivery: A pain management case study.
Subcutaneous (s.c.) administration offers practical advantages for long-acting drug delivery, yet its complex tissue environment, lack of standardized models, and limited regulatory guidance pose challenges for predicting pharmacokinetics (PK). Here, we present a complementary set of experimental and computational approaches to reduce reliance on animal testing while maintaining translational relevance. Ex vivo skin models captured the influence of tissue variability, while 3D-printed constructs and molded agarose hydrogels enabled standardized, reproducible in vitro drug release studies. In parallel, machine learning models trained on curated rodent datasets predicted key PK parameters with strong agreement to in vivo data, providing an alternative that bypasses additional animal experiments and resource-intensive assays. Finally, a liposomal carprofen formulation case study demonstrated the translational potential of these methods in veterinary applications. Together, these strategies illustrate routes toward animal-free drug development: simplified experimental systems that mimic s.c. release and computational models that predict systemic PK. By combining them conceptually, we outline a framework that advances the 3Rs principles while supporting mechanistic and predictive understanding of s.c. drug delivery.
10 June 2026
Read appraisal →Journal of robotic surgery
Technical challenges in autonomous robotic ultrasound examinations: perception, planning, and control
Ultrasound is one of the most widely used medical imaging modalities in clinical practice. Robot-assisted ultrasound systems (RAUS) offer significant advantages in reducing physician workload and improving imaging standardization. However, the dynamic and complex nature of the human physiological environment severely hinders the transition of RAUS from automation to full autonomy. This paper presents a narrative review of the core research progress of RAUS between 2021 and early 2026. It comprehensively analyzes the current technological landscape of autonomous robotic ultrasound by focusing on three core pillars: perception, planning, and control. By evaluating cutting-edge breakthroughs and existing limitations across these three domains, this paper identifies the key technical bottlenecks that restrict fully autonomous RAUS. Finally, we propose an evolutionary roadmap driven by multimodal fusion, digital twins, planning-control co-design, and embodied artificial intelligence (AI) architectures. This roadmap provides a developmental reference for autonomous robotic ultrasound examinations and also lays a crucial foundation for future autonomous ultrasound-guided surgical interventions.
1 June 2026
Read appraisal →PloS one
Towards intelligent railway monitoring: A novel hybrid deep learning architecture for railway obstacle detection.
Railways are among the most efficient modes of transportation, capable of moving large quantities of goods and passengers over long distances at relatively low cost. However, accidents frequently occur due to objects or individuals present on the tracks, as trains are unable to swerve and require long braking distances. While the localization of objects within the track bed is a well-explored topic, the reliable and high-performance classification of such obstacles across all relevant categories remains an unresolved challenge. This study proposes an innovative hybrid architecture that leverages the specific visual characteristics of track bed imagery, setting a new benchmark in this domain. The hybrid design effectively leverages the strengths of ResNet50 and Swin Transformer V2, allowing the model to capture both local and global features. In addition, an Efficient Attention Module is integrated to further emphasize the most relevant features for robust obstacle classification. Using stratified five-fold cross-validation on a dataset of 2,003 images across six classes (iron bar, boulder, person, branch, canister, and barrel), the model achieved an average balanced accuracy of 99.46%. The results have implications for accident prevention, improving operational efficiency, and modernizing railway safety systems, thereby enabling the future application of automatic railway surveillance systems to ultimately enhance operational security.
1 June 2026
Read appraisal →Journal of neural engineering
An online brain-computer interface for detecting incongruity in augmented reality applications
Objective.Augmented reality can provide digital information about physical entities presented within its real-world context. However, this information might disagree with the user's expectations due to factual errors in the data or cognitive biases. Such incongruity can impair user experience and undermine trust in the AR system. To address this issue, we propose detecting inconsistencies between physical objects and digital information through hybrid brain-computer interfaces.Approach.We conducted two complementary experiments. First, we implemented a strategy that integrates eye-tracking and brain signals for incongruity detection in an offline study. Subsequently, we assessed our approach in an online study in which participants received immediate feedback on the classification.Main results.The grand average event-related potentials revealed consistent electroencephalographic responses to incongruent augmentations, specifically a centroparietal N400 effect, across both experiments. We could further distinguish between congruent and incongruent information with an average balanced accuracy of 70% in the online study.Significance.These findings demonstrate the feasibility of detecting incongruity online, allowing for autonomous system adaptation, like presenting information in a more accessible format or providing contextual support.
31 May 2026
Read appraisal →Science (New York, N.Y.)
Strain-resilient intrinsically stretchable electrochemical biointerfaces
Stretchable bioelectronics promise seamless integration with dynamic tissues, yet their electrochemical performance often collapses under strain owing to cracking, interlayer delamination, and signal distortion. We introduce an intrinsically stretchable interface for resilient electrochemical sensing (SIRES) built from a strain-resilient conductor, an electrically tunable interlayer, and a stretchable functional coating. By offsetting strain-induced resistance increases with gains in active surface area, SIRES maintains near-constant resistance and high-fidelity electrochemical readouts under strains up to 300%. The platform supports voltammetric, potentiometric, and amperometric modalities and enables multiplexed molecular monitoring across dynamically deforming tissues in wearable and implantable formats. Its unified architecture provides delamination-resistant interfaces suitable for long-term use, and the design rules generalize to affinity-based transduction, establishing a pathway toward strain-resilient molecular sensing for precision diagnostics and therapy.
31 May 2026
Read appraisal →Biomedical physics & engineering express
Stabilizing neuromorphic ECG processors via adaptive fractional fatigue
Spiking neural networks (SNNs) deployed on wearable devices can exhibit runaway firing when processing noisy electrocardiogram (ECG) signals, increasing switching activity and dynamic power and destabilising power regulation. To suppress redundant firing while preserving ECG classification performance, we introduce the adaptive fractional fatigue (AFF) SNN, which equips each hidden neuron with a power-law weighted memory buffer of its past membrane states, thereby forming a fractionally-inspired activity trace that drives an adaptive threshold: as the accumulated trace rises with sustained firing, the effective threshold increases and suppresses further spikes, implementing an intrinsic negative feedback loop. Ten independent seeds were evaluated with a fixed random train/test split of2969/743heartbeat windows (lengthT=180;3712balanced windows total;1856normal,1856arrhythmia) drawn from three publicly available Massachusetts Institute of Technology-Beth Israel Hospital records, with spikes per sample as the primary energy proxy and test accuracy and the area under the receiver operating characteristic curve as secondary endpoints. Spikes/sample decreased from2384.990±537.719(integer-order) to891.481±227.654(AFF), a62.62%reduction (bootstrap 95% confidence interval[54.00%,69.35%]), which was statistically significant (pairedt-testp=0.000015two-sided; Wilcoxon signed-rankp=0.001953two-sided; paired effect sizedz=-2.662). Accuracy was preserved (78.923±1.725%vs79.206±1.541%; pairedt-testp=0.227761), and cross-seed spike variance decreased by a factor of5.58×. AFF suppresses noise-driven hyperactivity and stabilises event traffic, delivering a statistically significant reduction in spikes/sample without accuracy loss and improving predictability for low-power neuromorphic ECG deployment.
30 May 2026
Read appraisal →ACS applied materials & interfaces
Breathable and Washable E-Textile with Readily Integrated Piezoelectric Perfluoroalkoxy Fiber for Wide Temperature Range Human Machine Interfaces
Flexible wearable sensors are widely explored for human-machine interaction and health monitoring. Electronic textiles (E-textiles) offer a promising pathway toward comfortable wearable devices by integrating sensing fibers directly into fabric structures. Among them, compared with piezoresistive and capacitive fiber sensors, piezoelectric fiber sensors require no external power supply. Compared with triboelectric fiber sensors, which are highly sensitive to humidity and temperature, piezoelectric fiber sensors provide more stable and reliable electrical signals. However, conventional piezoelectric fiber sensors still face challenges in achieving long-term wearing comfort and reliability under harsh environments. Herein, we develop a breathable, washable, and wide-temperature-tolerant based on piezoelectric perfluoroalkoxy (PFA) fibers, which can be readily integrated into E-textiles. The textile is produced via a weft knitting process that integrates coaxially structured PFA fibers with wool yarns, combining stable piezoelectric performance with wearing comfort. The fibers exhibited a charge sensitivity of (2.6 ± 0.2) pC/N at 0.2 MPa. The resulting E-textile exhibits a charge sensitivity of up to (316.6 ± 45.9) pC/N and a voltage sensitivity of (744.9 ± 73.9) mV/N. The textile maintains stable operation over a wide temperature range from -78 °C to 150 °C. Its weft-knitted structure provides breathability, while the textile also offers washability, UV resistance, and mechanical durability exceeding 10,000 stretching or bending cycles. Furthermore, the textile can effectively monitor various human motions. A smart glove constructed from five-channel PFA fibers achieved recognition accuracies above 95% for five hand gestures, demonstrating the potential offer self-powered E-textile as a wearable sensor for human-machine interaction.
30 May 2026
Read appraisal →Nano letters
Hierarchical Structural-Interfacial Engineering with Dynamic Soft-Hard Cross-Linking Enables Full-Range, Ultrasensitive MXene Piezoresistive Sensors
Human tactile perception inspires flexible piezoresistive sensors, yet simultaneously achieving high sensitivity, a wide pressure range, and mechanical robustness remains challenging. Here, we report a hierarchical stress-regulation strategy that integrates multiscale surface microstructures with a dynamically cross-linked MXene/carboxymethyl cellulose/borax sensing network. Replicated microtopographies induce progressive and spatially distributed stress localization, while heterogeneous soft-hard cross-linking regulates nanoscale deformation through adaptive hydrogen bonding and rigid borate anchoring. This coupled structural-interfacial regulation generates abundant stress-concentrated sites, stabilizes conductive pathways, and enables continuous resistance modulation across a broad pressure spectrum. Consequently, the sensor exhibits ultrahigh sensitivity (774.48 kPa-1), a wide working range (334.16 kPa), and fast response/recovery times (8.58/17.22 ms). It reliably captures both subtle physiological signals and large mechanical loads and further supports gesture recognition and robotic control when integrated with real-time feedback and machine learning. This work establishes a general framework for designing robust, full-range tactile sensors through hierarchical stress regulation.
30 May 2026
Read appraisal →Science advances
MR-AIV reveals in vivo brain-wide fluid flow with physics-informed AI
The circulation of cerebrospinal and interstitial fluid plays a vital role in clearing metabolic waste from the brain, and its disruption has been linked to neurological disorders. However, directly measuring brain-wide fluid transport, especially in the deep brain, has remained elusive. Here, we introduce magnetic resonance artificial intelligence velocimetry (MR-AIV), a framework featuring a specialized physics-informed architecture and optimization method that reconstructs three-dimensional fluid velocity fields from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). MR-AIV unveils brain-wide velocity maps while providing estimates of tissue permeability and pressure fields, quantities inaccessible to other methods. Applied to the brain, MR-AIV reveals a functional landscape of interstitial and perivascular flow, quantitatively distinguishing slow diffusion-driven transport [∼0.1 micrometers per second (μm/s)] from rapid advective flow (∼3 μm/s). This approach enables new investigations into brain clearance mechanisms and fluid dynamics in health and disease, with broad potential applications to other porous medium systems, from geophysics to tissue mechanics.
29 May 2026
Read appraisal →Progress in biomedical engineering (Bristol, England)
Soft pneumatic actuators for wearable systems: a structural classification and integration-oriented evaluation
Soft pneumatic actuators promise gentle, body-safe assistance, yet many fail at the moment of integration: inflation alters geometry, increases profile, and stresses seams and ports. This topical review reframes actuator selection through a structure-first lens that links how an actuator is built to how it deforms and how it embeds in garments. Actuators are classified by structural architecture, deformation behavior, and fabrication method, and evaluated against four integration criteria central to wearable systems: geometric compatibility, fabrication scalability, system integration readiness, and actuation simplicity. A staged selection pathway is proposed and presented as a literature-informed mapping table that links reported wearable application contexts to dominant integration priorities and a practical structural-class starting point. Planar sheet designs can preserve a thin, predictable pressurized envelope at modest pressures when seam paths, constraint layers, and attachment features are co-designed. Pouch and bladder actuators are thin at rest, simple to fabricate, and readily integrated with textiles, but commonly bulge under load without external constraint; envelope control and port or manifold routing frequently limit garment integration. Fiber-constrained actuators deliver high specific force but often require rigid end terminations and elevated pressures. Segmented elastomers provide rich kinematics through chamber layout while tending toward bulging and routing burden in multi-segment formats. Mechanical metamaterials realize geometry-programmed motion when hinge fidelity and pattern alignment are maintained; durability of compliant joints and consistent crease formation set current limits. The resulting synthesis identifies practical priorities for wearable use: preserve thin profiles under load, embed stitchable or bondable attachment features, document reproducible process windows, minimize dead volume, routing, and valve count with compact pneumatic architectures, and target low-pressure operation compatible with lightweight hardware. Framing integration in structural terms standardizes comparison across actuator classes, clarifies the narrow feasible design space for compact, body-conforming devices, and supports deliberate actuator choice for biomedical wearables.
29 May 2026
Read appraisal →ACS applied materials & interfaces
Boosting Piezoelectricity and Mechanical Properties of Silk Fibroin Films via Polyethylene Glycol Phase Manipulation
Piezoelectric silk fibroin (SF) shows great potential in fabricating sustainable wearable electronics toward health monitoring and human-machine interfacing, yet the processing difficulty and brittleness restrict its practical applications severely. This study proposes a simple strategy to achieve the boosted piezoelectricity and mechanical properties in electrospun regenerated silk fibroin (RSF) nanofiber mats by adding polyethylene glycol (PEG) with different molecular weights to induce the formation of β-sheets via hydrogen bonding and the regulation of PEG phase morphology. Specifically, RSFs consisting of PEG of 2000 and 1000 Da exhibit a significant improvement in the tensile strength (from 6.1 to 9.92 MPa) and elongation at break (from 5.72% to 117.58%), respectively. Meanwhile, RSFs consisting of PEG with a molecular weight of 2000 and 400 Da show a direct piezoelectric force sensitivity of 0.2 mV/N·μm and 0.14 mV/N·μm respectively, superior to that of pure RSF (0.04 mV/N·μm). The piezoelectric force sensors fabricated with RSFs show potential applications in the detection of diverse human motions, including finger tapping for Morse code, grip strength, and knee flexion. Thus, this work unveils a facile strategy to fabricate RSF-based biodegradable piezoelectric force sensors with a trade-off in processability, mechanical ductility, and piezoelectricity.
28 May 2026
Read appraisal →Scientific reports
Uncertainty estimation and probabilistic skull shape reconstruction using bayesian neural networks
3D shape reconstruction is an active area of research and a fundamental problem in computer vision, with growing applications in the medical domain, where it enables the recovery of missing or fine anatomical structures. Numerous approaches have been proposed for medical shape reconstruction, emphasizing accurate and anatomically plausible reconstructions. However, 3D shape reconstruction remains an inherently ill-posed problem, meaning (1) both neural network-based methods and conventional shape modeling approaches naturally introduce uncertainty in their predictions, and (2) multiple anatomically plausible reconstructions exist for a given partial or low-resolution input. While the uncertainty aspects have been widely explored in general computer vision, they remain relatively under-explored in the context of medical shape reconstruction. In this paper, we developed a 3D Bayesian U-Net and investigated its use for uncertainty estimation and probabilistic reconstructions across three key tasks: cranial reconstruction, facial reconstruction, and skull shape super-resolution. Our findings show that the Bayesian model is able to produce a range of anatomically plausible skull reconstructions while capturing natural skull variations arising from the learned weight uncertainty. Notably, these variations are primarily expressed through differences in bone thickness, which aligns with anatomical expectations, particularly relevant in real-world applications like cranial implant design. Additionally, we propose a principled framework to study the relationship between weight uncertainty and reconstruction uncertainty by analyzing the learned posterior distribution of the weights, demonstrating that our Bayesian U-Net achieves comparable reconstruction performance to a deterministic U-Net baseline while providing reliable uncertainty estimates. Our study also reveals a clear cross-task uncertainty pattern, where tasks with stronger structural constraints, like super-resolution, yield lower predictive uncertainty, while less constrained tasks, like facial reconstruction, result in higher uncertainty. Refer to the project page for more visual results https://git.zib.de/jli/uncertainty-aware-skull-reconstruction/ .
28 May 2026
Read appraisal →Biomedical physics & engineering express
Acoustic-based intraoperative assessment of femoral stem initial stability in cementless THA via sensitive frequency band identification
Intraoperative assessment of femoral stem initial stability in cementless total hip arthroplasty (THA) predominantly relies on subjective tactile and auditory feedback, which lacks standardization and increases complication risks. This study develops a quantitative framework to identify sensitive acoustic frequency bands characterizing the bone-implant fixation state. Specifically, we acquired high-fidelity acoustic signals from 118 primary cementless THA procedures. Utilizing a statistical difference analysis framework, we identified two distinct sensitive bands (640-1920 Hz and 3460-3560 Hz) to extract band power fraction features. Machine learning evaluation demonstrated that these bands significantly outperform traditional empirical partitioning, achieving a classification accuracy of 0.816±0.028, which improved to 0.834±0.042 after feature fusion. Furthermore, stratified validation confirmed high consistency across diverse bone qualities and implant designs (Summit and Corail). While robust under steady-state Gaussian noise, the system exhibited sensitivity to non-steady-state speech interference. In conclusion, the identified sensitive bands effectively reflect the mechanical evolution of the bone-implant interface, providing a physically interpretable decision-support tool for enhancing surgical safety.
27 May 2026
Read appraisal →ACS nano
All-Printed MXene-Based Springs for Concurrent Bidirectional Hand Motion Capture
Accurately capturing hand movements remains a challenge because the skin undergoes stretching, compression and bending at the same time, yet most flexible sensors respond only to tension. Achieving high sensitivity and a wide range often means making the sensor thick, which causes the sensing layer to shift under complex deformations and distorts the signal. Here, we address this by placing the strain-accommodating space within a plane parallel to the strain direction. Using a MXene/polyurethane ink with glycerol as a molecular bridge, we create a dense hydrogen-bonding network that enables high-resolution printing of a planar spring architecture. The resulting all-printed sensor achieves a gauge factor of 83.7 across a linear bidirectional range of ±60% strain. With just two sensors, we track and decode complex wrist motions with ±10° accuracy. When attached to finger joints, they enable real-time, dexterous control of a virtual hand. This work provides an alternative approach for skin strain capture through synergistic innovation in materials and architecture.
27 May 2026
Read appraisal →Journal of neural engineering
Non-invasive somatotopic proprioceptive feedback for closed-loop robotic hand aperture control
Objective.Current robotic prostheses developed for individuals with transradial amputation often lack physiological feedback, particularly proprioceptive information, which limits control precision and increases reliance on vision.Approach.This study investigates the effect of introducing non-invasive somatotopic feedback using transcutaneous electrical nerve stimulation (TENS) to convey hand aperture information, as an artificial form of proprioceptive feedback. Twenty healthy participants with intact limbs were divided into two groups: one receiving TENS-based feedback and one without feedback. Participants performed an aperture control task under visual and non-visual conditions, with some trials including a concurrent Stroop task to assess cognitive load. We hypothesized that providing non-invasive somatotopic proprioceptive feedback via TENS mitigates key control and integration challenges, leading to improved accuracy, faster learning, and reduced reliance on vision, without increasing cognitive demands.Main results.Under visual deprivation, participants receiving TENS feedback achieved significantly smaller aperture control errors than those without feedback, both under alternating visual conditions (p<0.001) and under prolonged visual deprivation conditions (p=0.017). From the very first trial, TENS-based feedback enabled control accuracy comparable to that with visual input. Although perceptual shifts affected control accuracy under dual-task conditions, the +TENS group maintained high cognitive performance and effective control toward perceived functional targets, suggesting that the artificial feedback nonetheless supported hand control in the presence of a secondary task.Significance.These findings highlight the potential of non-invasive somatotopic proprioceptive feedback delivered through TENS to provide physiologically meaningful information and support future strategies for restoring proprioceptive function in prosthetic hand users.
27 May 2026
Read appraisal →Nature communications
Measuring multi-site pulse transit time with an AI-enabled mmWave radar
Pulse Transit Time (PTT) is a measure of arterial stiffness and a physiological marker associated with cardiovascular function, with an inverse relationship to diastolic blood pressure (DBP). We present an AI-enabled mmWave system for contactless multi-site PTT measurement using a single radar. By leveraging radar beamforming and deep learning algorithms our system simultaneously measures PTT and estimates diastolic blood pressure at multiple sites. The system was evaluated across three physiological pathways - heart-to-radial artery, heart-to-carotid artery, and mastoid area-to-radial artery - achieving correlation coefficients of 0.75-0.86 compared to contact-based reference sensors for measuring PTT. Furthermore, the system demonstrated correlation coefficients of 0.90-0.91 for estimating DBP, and achieved a mean error of -0.62-0.06 mmHg and standard deviation of 4.54-5.20 mmHg, meeting the FDA's AAMI guidelines for non-invasive blood pressure monitors. These results suggest that our proposed system has the potential to provide a non-invasive measure of cardiovascular health across multiple regions of the body.
27 May 2026
Read appraisal →Science advances
Bioinspired ionic thermoreceptors with anisotropic architecture for thermotactile perception in robots
Emulating human skin's ability to perceive temperature and identify material through thermotactile perception is critical for human-machine interaction, robotics operation, and prosthetic sensory feedback systems. However, conventional artificial thermal sensors are largely limited to temperature measurement and cannot replicate the thermotactile-mediated material recognition capabilities inherent in biological systems. Here, we present a bioinspired ionic thermoreceptor with anisotropic thermal response characteristics, enabling high-fidelity material recognition and accurate temperature monitoring. The device incorporates spatially specialized sensing elements that encode thermal signals into modality-specific temporal response patterns, allowing the extraction of the thermal contact coefficient for material discrimination and the absolute temperature for precise thermal monitoring. It achieves high-accuracy material recognition (98.9%) and substantial temperature resolution (0.81 millikelvins). Furthermore, robust covalent bonding between functional layers ensures mechanical durability, signal stability, and environmental resilience. This work establishes a physically grounded and energy-autonomous thermotactile sensing platform for safe, intuitive, and intelligent human-machine interaction.
25 May 2026
Read appraisal →ACS sensors
In Situ Formation of Mechanical Interlocking Skin and Electronics Interfaces through Dual-Phase Transition of Conductive Materials
The increasing demand for daily health monitoring has accelerated the development of wearable devices. However, conventional wearables, typically based on rigid materials and external fixation methods, suffer from interfacial instability during dynamic motion and heavy sweating, leading to signal distortion and device delamination. Skin-like wearable electronics and in situ fabrication strategies have emerged to improve conformability and signal stability. However, challenges remain in achieving a stable skin-device interface and reliable performance under dynamic conditions using simplified fabrication methods. Here, we present a flexible multiparameter monitoring device that integrates with the skin through a skin-electronics interface undergoing a dual-phase transition from hydrogel to liquid and subsequently to a solid state via in situ heating and photonic crosslinking. This transition forms mechanically interlocked structures that ensure robust adhesion and stable skin contact even under humid and high-acceleration conditions. Rapid curing enabled by an integrated heating layer and optical waveguide reduces the fabrication time to 1 min, allowing efficient on-demand deployment. The device achieves stable monitoring of ECG, PPG, acceleration, and temperature signals and supports Bluetooth data transmission under accelerations up to 11 g, as well as during swimming and showering. Furthermore, machine-learning-based motion recognition and individual identification models achieve 100% accuracy. This work advances in situ-fabricated wearable electronics by improving fabrication efficiency, interfacial robustness, and environmental adaptability, providing a promising platform for portable physiological monitoring.
23 May 2026
Read appraisal →ACS sensors
Moisture-Gated Bio-Semiconductor Electronic Tattoos for Continuous and Imperceptible On-Skin Respiratory Monitoring
Continuous respiratory monitoring remains one of the most critical yet unmet needs in personal and clinical healthcare. To overcome the limitations of cumbersome systems and motion-prone wearables, we present a moisture-gated bio-semiconductor-based electronic tattoo (BSET) that enables imperceptible, skin-conformable, and wireless respiratory monitoring by leveraging the hydration-sensitive ionic-electronic conductivity of melanin and the skin-compatibility of silk. Fabricated from a silk fibroin-melanin nanofiber bio-composite, the BSET is ultrathin at 18 μm and highly breathable, achieving a water vapor transmission rate exceeding 3000 g·m-2·d-1. By detecting exhaled moisture directly on the philtrum, the sensor exhibits a rapid response time of 1 s and a recovery time between 2 and 10 s. A direct-spun nanofiber-based wiring strategy ensures robust integration, accommodating a 10 mm displacement under 10.4 MPa of stress without failure. Driven by a compact 3 g circuit with 20 mW of power consumption, the system supports continuous wireless data transmission for over 7.3 h. During vigorous exercise and sleep, the BSET reliably monitored respiratory dynamics, identifying 10-20 s apnea events and enabling multiparameter analysis of breathing frequency and exhalation intensity. This lightweight system establishes a scalable and clinically relevant solution for continuous respiratory surveillance.
23 May 2026
Read appraisal →ACS sensors
Diagnosis of Primary Open-Angle Glaucoma Using Spectral Profiling of Aqueous Humor-Derived Exosomes
Primary open-angle glaucoma (POAG) is one of the most common neurodegenerative diseases that cause irreversible optic nerve damage. POAG diagnosis requires multimodal assessments; however, current methods like intraocular pressure (IOP) and optical coherence tomography (OCT) imaging suffer from low sensitivity and inter-patient variability, respectively. Liquid biopsy provides objective molecular signatures, offering a robust alternative to overcome limitations of conventional clinical hallmarks. Here, we present an efficient and highly sensitive diagnostic platform operating on spectral profiles from aqueous humor (AH)-derived exosomes. Following morphological and compositional validation of exosome presence in AH samples, antibody-functionalized substrates were fabricated for selective capture. Immobilized AH-derived exosomes are then coated with gold nanoparticles to generate molecular fingerprint Raman signals. A total of 7600 spectra are acquired and used to train and evaluate the convolutional neural network (CNN) model for binary classification between healthy controls and glaucoma patients. The trained CNN model achieved an AUC of 0.96 and a prediction accuracy of 91%. The system enables rapid, individualized diagnosis, overcoming sample volume limitations via integrated immunoassay-based isolation and artificial intelligence (AI)-driven classification, highlighting the potential of ocular fluid-derived EVs in the diagnosis of neurodegenerative diseases.
23 May 2026
Read appraisal →ACS sensors
Urine-Based Noninvasive Detection of Prostate Cancer Using Human Olfactory Receptor-Embedded Nanodiscs
The diagnosis of prostate cancer (PCa) is limited by the low specificity of prostate-specific antigen (PSA) testing, which contributes to overdiagnosis and patient exposure to unnecessary invasive prostate biopsies. Here, we report a urine-based diagnostic platform that combines a six-member human olfactory receptor-embedded nanodisc (OR-ND) sensor array with fluorescence sensing and machine learning to detect PCa-associated volatile organic compounds (VOCs). Six ORs were reconstituted into lipid nanodiscs and evaluated using urine samples from a strictly curated subcohort of 40 PCa patients and 33 healthy controls (n = 73) to establish baseline sensor responsiveness. Subsequently, to ensure diagnostic robustness, machine learning classifiers were developed and cross-validated using an expanded data set of 290 samples, achieving an accuracy of 0.890 and an AUC of 0.964. Analysis of the full receptor data set revealed distinct response patterns, and statistical screening with correlation filtering identified OR2W1, OR51E1, and OR51E2 as the most informative features. Using these selected receptors, a random forest classifier achieved an accuracy of 0.890 and an AUC of 0.964, demonstrating high sensitivity and specificity. OR response patterns showed a closer association with Gleason score than with serum PSA levels, indicating that urinary VOC signatures capture tumor-related metabolic information complementary to conventional biomarkers. These results demonstrate that an OR-ND sensor array coupled with machine learning enables accurate and noninvasive detection and classification of PCa from urine samples. Our study provides a modular framework extensible to other VOC-associated diseases.
23 May 2026
Read appraisal →ACS sensors
Recent Advances in Electrospun Nanofiber-Based Wearable Sweat Sensors
Electrospun nanofibers, owing to their high porosity, flexibility, straightforward fabrication, and tunable physicochemical properties, have been widely employed in wearable sensors for gas detection, temperature monitoring, pulse measurement, gesture recognition, and other applications. In recent years, research on employing electrospun nanofibers in wearable sweat sensors has grown substantially. This paper provides the first comprehensive review of the research progress in electrospun nanofiber-based wearable sweat sensors, focusing on structural design, device fabrication, and potential applications. The review begins with a concise introduction to the fundamental aspects of electrospun nanofibers, including their formation principles, processing routes, material selection, and morphological characteristics. It then delves into the key components involved in constructing wearable sweat sensors based on electrospun nanofibers, such as reagent immobilization matrices, signal-enhancing materials, flexible substrates, and sweat collection components. A detailed analysis of the functional mechanisms of nanofibers in these contexts is provided, accompanied by an assessment of their advantages and limitations. Furthermore, recent applications of electrospun nanofiber-based wearable sweat sensors are systematically reviewed. Finally, the review discusses the challenges and potential future applications of electrospun nanofibers in wearable sweat sensors. This comprehensive overview aims to keep researchers informed of the latest advancements in the field and to stimulate further innovative developments.
23 May 2026
Read appraisal →ACS sensors
Artificial Intelligence-Driven Wearable Sensors for Cardiovascular Health Monitoring
Wearable sensing technologies hold great promise for continuous and non-invasive monitoring of cardiovascular health, offering new avenues for early detection and effective management of cardiovascular diseases. Despite these advances, several critical challenges remain, including inefficiencies in sensor design, the degradation of calibration model accuracy over time, and difficulties in processing complex physiological signals under real-world conditions. The integration of artificial intelligence (AI) into wearable cardiovascular monitoring systems presents transformative solutions to these limitations by improving system performance, adaptability, and clinical applicability. In this review, we provide an overview of current wearable technologies for cardiovascular monitoring, including their mechanisms, characteristics, and limitations. An in-depth discussion is subsequently presented on AI-driven wearable cardiovascular monitoring, encompassing five critical facets: sensor design, sensor calibration, signal processing, pattern recognition, and disease management. Furthermore, we analyze the challenges associated with the widespread use of AI-based wearable systems and provide insights into potential strategies for addressing these challenges. This review is expected to serve as a roadmap for future research and development in the rapidly evolving field of intelligent cardiovascular health monitoring.
23 May 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Soft, skin-interfaced electronics enable cannula-free wireless monitoring of sleep respiration.
Sleep-related breathing disorders are prevalent yet frequently underdiagnosed, in part due to limitations of conventional respiratory monitoring technologies. Standard nasal cannulas introduce airflow resistance, discomfort, and poor long-term adherence, constraining at-home and longitudinal assessment. Here, we report a soft, skin-interfaced nasal patch that enables cannula-free, wireless monitoring of respiratory activity during sleep. The device is constructed from ultrathin, elastomeric materials that conform to the nasal surface, coupling respiratory-induced tissue deformation to a strain-sensing element. The mechanics of the skin-device interface and the elastomeric response govern the sensitivity and linearity of signal transduction, enabling quantitative capture of breathing dynamics. An integrated wireless platform transmits deformation signals directly to mobile devices, eliminating the need for external tubing or tethered modules. Modular fabrication permits replacement of the strain sensor and skin-contact interface without compromising mechanical performance. Mechanical characterization under physiologically relevant deformation demonstrates high repeatability and low hysteresis, while in vivo studies confirm that the patch accurately reproduces respiratory waveforms and correlates closely with gold-standard nasal cannula measurements. By integrating soft materials mechanics, wearable strain sensing, and wireless electronics, this system provides a minimally obtrusive platform for continuous respiratory monitoring. The class of technologies presented in this work establishes design principles for skin-interfaced devices, in which elastomeric mechanics, strain transduction, and wireless integration combine to enable quantitative, unobtrusive physiological monitoring in clinical and home environments.
22 May 2026
Read appraisal →Brain topography
fNIRS-STCT: A Novel Hybrid Spatial and Temporal CNN and Transformer Network for fNIRS Signal Classification
Functional near-infrared spectroscopy (fNIRS) is a non-invasive brain imaging method that detects neural activity by monitoring blood oxygen levels and circulation changes in the brain. Some current fNIRS classification models are simply adapted from deep learning models such as Transformer and fail to effectively leverage both temporal and spatial information from fNIRS. In this paper, we develop a hybrid deep learning model for fNIRS classification, fusing spatial convolutional neural network (CNN), a temporal CNN, and Transformer, named as fNIRS-STCT. Primarily, the architecture employs dual large convolutional kernels - one for temporal feature extraction and another for spatial feature extraction. Second, spatial and temporal CNNs are respectively integrated with Transformer by feeding CNN-extracted features as input to Transformer and then concatenating CNN and Transformer features. Third, spatial and temporal features are fused by element-wise addition. Finally, we exploit cross-entropy as the loss function, which integrates both online label smoothing and flooding strategies to avoid overfitting. The model's effectiveness is demonstrated through testing on two public fNIRS datasets. Across subject-dependent, subject-semi-dependent, and subject-independent evaluation schemes, fNIRS-STCT model consistently outperforms other compared models for all given performance metrics. Specifically, on the unilateral finger- and foot-tapping dataset, our model achieves accuracies of 83.29%, 81.73%, and 79.60% across three evaluation schemes, surpassing the second-best model by margins of 3%, 2%, and 1%, respectively. The findings underscore fNIRS-STCT's effectiveness in cross-subject mental state and brain activity interpretation.
21 May 2026
Read appraisal →ACS applied materials & interfaces
Breathable Bioadhesive Electronic Skin for Precise Multimodal Electrophysiological Monitoring with AI-Assisted Physiological-State Prediction
Accurate and stable monitoring of electrophysiological (EP) signals is essential for wearable healthcare applications. However, conventional commercial electrodes suffer from poor conformal contact, limited breathability, and weak adhesion under sweaty skin and human motion condition, leading to unstable skin-electrode interface and low-quality signal acquisition. Moreover, single-modality EP signal measurements are insufficient for reliable physiological and health evaluation. Herein, we develop a multimodal electrophysiological signal monitoring electrode (MESME) featuring a fiber membrane combining thermoplastic polyurethane (TPU) and poly(acrylic acid) functionalized with N-hydroxysuccinimide ester (PAA-NHS), seamlessly integrated with a liquid metal (LM) serpentine conductive path. Compared with commercial electrodes (gel-based wet electrode), this MESME exhibits high mechanical flexibility and conformability, superior breathability, robust bioadhesion, stable electrical performance under perspiration, and motion conditions and lower interfacial impedance, enabling high-fidelity multimodal signal acquisition. Leveraging this capability, the multimodal signals generated by MESME can be further integrated with deep learning for drowsiness prediction, achieving a validation accuracy of 97.14%. This platform offers a practical strategy for multimodal electrophysiological monitoring and supports future development of data-driven health assessment models.
21 May 2026
Read appraisal →Analytical chemistry
Adaptive Deep Learning for High-Fidelity Quantitative Single-Molecule Imaging
Accurate quantification of molecular dynamics in live cells is critical for elucidating receptor signaling and guiding therapeutic strategies. Yet current deep learning methods for fluorescence imaging often distort intensity and lack robustness under diverse imaging settings, limiting quantitative utility. We present an adaptive deep learning framework that constructs in situ training data sets from ongoing sequences and integrates a self-feedback algorithm to preserve absolute molecular intensities. This approach achieves ≥ 3.9-fold gains in signal-to-noise and 1.6-fold improvements in localization precision across organic dyes, quantum dots, and fluorescent proteins. By maintaining both spatial and intensity fidelity, it enhances super-resolution reconstruction, stepwise photobleaching analysis, and live-cell single-molecule tracking. Applied to programmed death-ligand 1 (PD-L1), it uncovers density-dependent clustering, with small-molecule inhibitors inducing dimerization and reduced mobility, while antibodies increase diffusivity. Combined treatments exert complementary effects on PD-L1 membrane organization. This adaptive and intensity-preserving framework provides a broadly applicable platform for high-precision quantitative bioimaging across modalities and experimental conditions.
20 May 2026
Read appraisal →Bioinspiration & biomimetics
Designing McKibben muscles: a critical review for practical implementation
McKibben artificial muscles (AMs) are known as a prominent class of pneumatic actuators in soft robotics and biomechanical engineering, due to their unique structure and multifunctionality. This paper presents a comprehensive review of recent advancements in McKibben AMs, focusing on their performance, structural variations, and operational principles. A systematic literature search on Scopus identified 146 relevant articles, which were analysed for both performance metrics and design characteristics. Inspired by natural muscle behaviour, McKibben AMs enable complex motions such as bending, linear extension, and twisting. These actuators can be organised as individual or bundled systems: individual units are typically arranged in linear or circular patterns, while bundled systems occur in serial, parallel, braided, convergent, or pennate configurations. Recent innovations in smart actuation methods, braided sleeves and internal bladders have expanded their capabilities, enabling embedded sensing, environmental adaptability and untethered operation. Additionally, alternative manufacturing methods offer promising solutions for developing McKibben muscles with enhanced functionality and tailored properties.
19 May 2026
Read appraisal →Physiological measurement
EdgeECG: a lightweight edge-oriented network with dual criterion pruning for real-time ECG arrhythmia classification
Objective.Miniature electrocardiogram (ECG) devices can rapidly and accurately acquire real-time cardiac signals, enabling timely warnings for patients with heart disease. To achieve accurate arrhythmia classification on resource-constrained ECG edge devices, we propose EdgeECG, an ultra-lightweight neural network designed for deployment on low-power microcontrollers.Approach.EdgeECG is first designed with a compact convolutional architecture to ensure compatibility with resource-limited embedded platforms. A dual criterion pruning (DCP) strategy is then introduced to evaluate weight importance using both magnitude and median deviation for more precise model compression. In addition, quantization is applied to reduce storage cost and improve deployment efficiency on an STM32F103 microcontroller.Main results.Experimental results on the benchmark MIT-BIH arrhythmia dataset show that EdgeECG achieves an overall classification accuracy of 98.07%, outperforming several representative methods. In patient-independent record-level evaluation, the accuracy reaches 80.73%. The model contains only 2680 parameters and is successfully deployed on an STM32F103 microcontroller with 64 KB SRAM, achieving an inference latency of 0.025 s and an energy consumption of 0.156 mJ per inference. In addition, DCP reduces the number of non-zero parameters by nearly 50% while maintaining promising classification performance.Significance.EdgeECG provides an effective solution for five-class ECG arrhythmia classification on severely resource-constrained edge devices. Its compact architecture, effective pruning strategy, and successful deployment on an STM32F103 demonstrate its potential for low-power edge-based ECG analysis and wearable cardiac monitoring applications. The source code of EdgeECG is publicly available at:https://github.com/jyanhuang/EdgeECG-code.
19 May 2026
Read appraisal →Biomedical physics & engineering express
Oral squamous cell carcinoma diagnosis with time and frequency domain features from optical coherence tomography A-scan signals
Optical coherence tomography (OCT) is extremely useful in the screening and detection of oral cancers. But various challenges, such as its subjectivity, operator dependence in interpretation, and lack of quantitative outcomes, have been delaying its adoption in clinical decision-making. The objective of this research is to quantify various A-scan features embedded in the OCT signal through advanced signal processing techniques and machine learning algorithms.Ex vivoimaging of biopsied oral tissues (normal mucosa, carcinomain situ(CIS), well-differentiated, and poorly differentiated oral squamous cell carcinoma) was performed using a spectral domain OCT system. Our A-scan dataset consisted of representative 1D signals obtained from different regions of the tissue bed. A set of 12 time- and 8 frequency-domain features were computed on each A-scan, and ten machine learning models were evaluated for binary and multi-label classification. LightGBM achieved the highest performance in both binary (accuracy: 0.8847,F1: 0.8878, AUC: 0.9539) and multi-label classification (accuracy: 0.8248,F1: 0.8201, AUC: 0.964). LightGBM was selected as the final model based on its superior and consistent performance across both classification paradigms. Our proof-of-concept feasibility study demonstrates good accuracy for differentiating oral mucosal tissues, highlighting the biomarker signature of A-scans.
19 May 2026
Read appraisal →ACS applied bio materials
Next-Generation MXene-Hydrogel Hybrids: Frontiers in Biosensing and Environmental Surveillance
Hydrogels have emerged as promising soft materials for applications in sensing, energy storage, and wearable electronics due to their tunable physicochemical properties and intrinsic biocompatibility. The integration of two-dimensional transition metal carbides/nitrides (MXenes) into hydrogel matrices has enabled the development of highly conductive, flexible, and electrochemically active composites for advanced sensing platforms. MXene-hydrogel hybrids exhibit enhanced charge transport, mechanical stability, and interfacial functionality, making them particularly attractive for electrochemical sensing applications. This review provides a comprehensive overview of MXene-hydrogel composites, focusing on their design strategies, synthesis approaches, and electrochemical sensing performance. Despite these advantages, critical challenges remain, including susceptibility of MXenes to oxidation, restacking of nanosheets, limited long-term stability, and difficulties in reproducible and scalable synthesis. These limitations significantly impede their translation into practical and commercial devices. Particular emphasis is placed on identifying current bottlenecks and outlining future research directions, including the development of oxidation-resistant MXenes, advanced hybrid architectures, scalable fabrication techniques, and integration into wearable and point-of-care sensing systems. Addressing these challenges is essential for realizing the full potential of MXene-hydrogel systems in next-generation electrochemical sensing technologies.
19 May 2026
Read appraisal →Science advances
Magnetic medical microrobots with memory-capable genetic circuits
Autonomous microrobots can reach hard-to-access regions in the human body for minimally invasive therapy. However, their microscale size limits the integration of on-board memory, making their operation dependent on external controls. Here, we develop a magnetic probiotic microrobot integrated with memory-capable genetic circuit to execute autonomous antitumor treatment. Through a one-time magnetic hyperthermia trigger, the biological thermal sensor in the microrobot perceives temperature change and activates the memory module Bxb1-ssrA-attB-P7-attP, transferring the microrobots into a therapeutic state to continuously degrade fibrin and soften the tumor microenvironment. The genetic memory remains active for at least 12 days. A synergy toward deep tumor penetration is subsequently established between the memory-encoded softening and the physical penetration through magnetically controlled wave-like locomotion of microrobots. Compared with memory-absent microrobots, the proposed microrobots achieve a 6.70-fold tumor matrix stiffness reduction and boost in vivo anticancer efficacy from 21.86 to 87.52%. Beyond oncology, the proposed system establishes a generalizable framework of memory-encoded medical microrobots.
18 May 2026
Read appraisal →Computer methods and programs in biomedicine
Geometry-aware PointNet for rapid prediction of cerebral aneurysm hemodynamics
BACKGROUND AND OBJECTIVE: Cerebral aneurysms affect 2-5% of the global population and pose a significant health risk upon rupture. While computational fluid dynamics provides detailed hemodynamic information for risk assessment, its high computational demands limit routine clinical use. This study aims to develop a deep learning model to rapidly predict three-dimensional velocity fields and wall shear stress at peak systole using aneurysm geometry. METHODS: We synthesized a dataset of 984 idealized middle cerebral artery bifurcation aneurysms. For each case, computational fluid dynamics simulations were conducted with pulsatile boundary conditions to generate ground-truth data, and peak-systolic snapshots were extracted. We developed a single-input point-cloud network augmented with a distance-to-wall feature to predict both velocity fields and wall shear stress. Model performance was evaluated using mean absolute error (MAE), normalized MAE (NMAE), and relative L2 error (rL2). RESULTS: On the test set, the model achieved velocity-field accuracy of NMAE 4.05% and rL2 19.2% over the full geometry domain, and WSS accuracy of NMAE 2.59% and rL2 23.9%. The mean inference time was approximately 1.6 seconds for velocity and 0.3 seconds for wall shear stress per case. Out-of-distribution evaluation on non-idealized geometries showed substantial zero-shot degradation (NMAE 19.1%, rL2 62.4%), while leave-one-out fine-tuning improved performance (NMAE 10.8%, rL2 37.0%). CONCLUSIONS: The proposed geometry-aware point-cloud surrogate provides fast peak-systolic hemodynamic prediction on idealized aneurysm geometries. However, out-of-distribution evaluation indicates that broader patient-specific training, physiological boundary conditions, and reliability assessment are required before routine clinical application.
17 May 2026
Read appraisal →JMIR medical informatics
From Flow to Feature Using a Proof-of-Concept Spectral-Driven Machine Learning Approach Using Smart Urinary and Drainage Catheter Systems: Algorithm Development and Validation
BACKGROUND: Current urinary and drainage catheter systems collect fluids for visual inspection or manual sampling, offering limited diagnostic value while being labor-intensive and prone to error. Machine learning (ML) has the potential to automate the analysis of these fluids. However, existing methods rely on complex preprocessing steps, which hinder real-time analysis. OBJECTIVE: We aim to develop and evaluate a fully automated, real-time diagnostic approach for smart urinary and drainage catheter systems by leveraging spectral data and ML to differentiate pathological from healthy excreted fluids without the need for manual preprocessing. METHODS: This study proposes a novel, fully automated approach for smart urinary and drainage catheter systems that uses spectra and ML to extract features from excreted fluids, enabling real-time analysis directly. A total of 454 surgical drainage fluid samples (from 181 patients) and 401 urine catheter samples (from 168 patients) were analyzed using smart catheters and drains equipped with compact mini-spectrometer sensors. The collected spectral data were fed into 3 different ML models: a random forest, a partial least squares discriminant analysis regression, and a convolutional neural network (CNN). Each model aimed to extract features and differentiate between pathological and healthy urine and drainage samples based on the various biomarkers available from previously conducted laboratory analyses. RESULTS: All 3 approaches (random forest, partial least squares discriminant analysis regression, and CNN) achieved promising results, demonstrating the potential of the overall approach. In particular, the CNN models trained on the drainage biomarkers hemoglobin and bilirubin achieved the best results. Matthews correlation coefficient scores of 0.83 and 0.81 were obtained for hemoglobin and bilirubin, respectively, when differentiating between pathological and healthy samples using the extracted spectral features. CONCLUSIONS: This work demonstrates the potential of spectral-driven ML for smart urinary and drainage catheter systems. This approach offers a real-time, noninvasive method for analyzing excreted fluids, paving the way for improved diagnostics and personalized patient care. Further research will explore the optimal ML model for this application.
17 May 2026
Read appraisal →Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Unexpected benefits of self-modelling in neural systems
Self-models have been a topic of interest for decades in human cognition and more recently in machine learning. Yet what benefits do self-models confer? Here we show that when an artificial network learns to predict its internal states as an auxiliary task, the network becomes simpler, more regularized and more parameter efficient. To test the hypothesis of self-regularizing through self-modelling, we used a range of network architectures performing three classification tasks across two modalities. In all cases, adding self-modelling caused a significant reduction in network complexity. The reduction was observed in two ways. First, the distribution of weights was narrower when self-modelling was present. Second, a measure of network complexity, the real log canonical threshold (RLCT), was smaller when self-modelling was present. These results support the hypothesis that self-modelling has a restructuring effect, reducing complexity and increasing parameter efficiency. This self-regularization may help explain some of the benefits of self-models reported in recent machine learning literature, as well as the adaptive value of self-models to biological systems. In particular, these findings may shed light on the possible interaction between the ability to model oneself and the ability to be more easily modelled by others in a social or cooperative context. This article is part of the theme issue 'World models in natural and artificial intelligence'.
17 May 2026
Read appraisal →ACS synthetic biology
Prompt-Only Gene-Circuit Modeling with a Large Language Model Simulates the synNotch Spheroids
Morphogenesis emerges from gene regulatory circuits that translate biochemical cues into spatially organized cell fates; yet, constructing quantitative simulations of these processes remains technically demanding. Here, we present a prompt-only pipeline in which a large language model (LLM) automatically generates CompuCell3D (CC3D) simulations directly from natural-language biological specifications. Using the canonical synNotch spheroid assay as a benchmark, we show that the LLM-generated code captures key features of contact-dependent signaling, cadherin-mediated adhesion, and multicellular sorting dynamics without manual programming. Baseline simulations exhibited the expected qualitative ordering of inducer-responder activation but revealed fragmentation into multiple spheroids, implicating premature reciprocal signaling as a kinetic failure mode. Incorporating a cumulative contact-duration threshold for synNotch activation rescued robust single-spheroid formation, suggesting that sustained cell-cell interactions are a critical design principle for this pattern. Our results establish conversational LLM guidance as a powerful accelerator for multiscale modeling, collapsing the steep learning curve of CC3D into a dialog-driven workflow. This framework enables nonspecialists to test gene-circuit hypotheses in silico and provides a generalizable route toward rapid, mechanism-level exploration of developmental design principles.
16 May 2026
Read appraisal →Thorax
Clinical impact of customised positive airway pressure (PAP) therapy interfaces (3DPiPPIn): a single site randomised controlled trial
INTRODUCTION: Positive airway pressure (PAP) therapy is the recognised treatment for sleep disordered breathing (SDB), delivered via a tight-fitting face mask (interface). Conventional interfaces do not consider facial geometries, often resulting in poor fit and ineffective therapy. Three-dimensional (3D) printing of customised interfaces may improve comfort and outcomes. OBJECTIVES: To evaluate the clinical impact of customised versus conventional oronasal interfaces in adults with obstrcutive sleep apnoea (OSA). The primary outcome was residual Apnoea Hypopnea Index (AHI) at 6 months; secondary outcomes included interface leak, therapy concordance and patient reported symptoms. METHODS: A randomised controlled trial with 160 adults naïve to PAP therapy and diagnosed with SDB (AHI ≥15 events/hour). Randomisation was minimised by age and ethnicity. Structured light facial scans (POP2, Revopoint, China) were used to produce 3D printed moulds (Fuse 30+, Formlabs, Massachusetts, USA) for silicone injected oronasal customised interface cushions.AHI was compared using quantile regression to account for the skewed distribution of the AHI data. Secondary outcomes were compared using logistic, quantile and linear regressions. RESULTS: 160 participants were recruited (intervention: 82, control: 78). Customised interfaces were associated with a 1.5 (events/hour) increase in AHI (p=0.059), higher interface leak (difference in medians 30.0 L/min, 95% CI 7.36 to 40.14, p<0.0001) and lower compliance (difference in compliance 0.78, 95% CI 0.05 to 1.54, p=0.04) at 6 months. CONCLUSIONS: This trial did not demonstrate customised oronasal interfaces were superior to conventional interfaces. Future research should focus on addressing design and manufacturing limitations before any potential advantages can be evaluated. TRIAL REGISTRATION NUMBER: ISRCTN74082423.
16 May 2026
Read appraisal →Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
What physics offers for artificial intelligence? Lessons from the brain's inner time and its dynamics.
Computing can take advantage of physics-that is, 'whatever physics offers' (Jaeger et al. 2023 Nat. Commun. 14, 4911 (doi:10.1038/s41467-023-40533-1)). One of the key features physics provides is time: it describes rules and makes predictions about how activity patterns change over time, as governed by dynamics. We argue that computing can learn from physics in how it conceives time-namely, in terms of dynamics, i.e. the changing patterns of activity unfolding over time. In particular, we focus on the brain's intrinsic neural dynamics of spontaneous activity and how it 'uses' them for dynamic input processing and encoding in order to 'participate' in the world's physical time. By 'participate', we mean becoming part of the input dynamics: for instance, when listening or dancing to music, neural activity (via entrainment) and, consequently, mental activity follows the rhythm and dynamics of the music. This shapes the listener's experience, such as consciousness-the brain, and thus the subject, actively participates in music rhythms through temporal alignment (Northoff et al. 2023 Interface Focus13, 20220076 (doi:10.1098/rsfs.2022.0076)). Drawing on recent empirical evidence, we show that dynamic features such as the brain's scale-free activity and variability-which reflect an intrinsic temporal structure, or the brain's 'inner time'-are central to tracking and encoding input dynamics. Importantly, this processing is actively modulated rather than passively received, through the brain's own 'hidden dynamic repertoire'. Extending earlier discussions by Dreyfus and others, we argue that current computing devices, whether classical or natural (i.e. non-von Neumann machines), lack spontaneous activity and an inner time that can exert an active, rather than purely passive, influence on processing. As a result, they can neither actively process and encode input dynamics through their own inner time, nor 'use' or 'participate' in the dynamics of the world's physical time. Instead of 'being in time' and 'being in the world', current computing devices-and, by extension, artificial intelligence-are effectively 'locked out of time and world', meaning they are not directly connected to physical time. Unlike humans, they therefore cannot be characterized as 'being in time' or 'being in the world', which in turn prevents them from acquiring tacit or implicit knowledge, including the ability to navigate and behave flexibly within a continuously changing world. This article is part of the theme issue 'World models in natural and artificial intelligence'.
15 May 2026
Read appraisal →Neuroscience
Recent trends in electrodermal activity signal processing and deep learning methods for emotion recognition
Electrodermal activity (EDA) has emerged as one of the most informative physiological signals for emotion recognition due to its close association with sympathetic nervous system activation. Although research on affective computing has expanded rapidly in recent years, systematic investigations of signal-processing techniques specifically tailored to EDA remain limited. This gap is significant because the selection of decomposition and signal processing strategies directly influences the accuracy, robustness, and interpretability of emotion recognition systems. Furthermore, defining emotion itself and tracing its physiological manifestations through EDA, from raw measurements to recognized affective states, remains an open challenge that highlights the need for methodological rigour. In this review, we synthesise developments published between 2018 and 2025, with a dual focus on advanced EDA signal processing methods and deep learning (DL) approaches for emotion recognition. We review common EDA acquisition sites, discuss challenges arising from non-stationarity and inter-subject variability. The review systematically compares time-domain, frequency-domain, time-frequency, and advanced time-series analysis techniques, alongside emerging end-to-end DL architectures for affective modelling. Unlike prior surveys that emphasise system-level design, this work adopts a signal-processing-centric and physiology-informed perspective. Additionally, we provide a structured comparative evaluation of widely used EDA decomposition methods using key performance metrics, offering a unified framework for method selection and interpretation. Overall, this review serves as a comprehensive resource for researchers and practitioners while advocating hybrid approaches that integrate the interpretability of classical signal processing with the predictive power of predictive methods.
15 May 2026
Read appraisal →Heart (British Cardiac Society)
Individual hearts: computational models for improved management of cardiovascular disease
Cardiovascular disease remains a leading cause of morbidity and mortality worldwide, with conventional management often applying standardised approaches that struggle to address individual variability in increasingly complex patient populations. Computational models, both knowledge-driven and data-driven, have the potential to reshape cardiovascular medicine by offering innovative tools that integrate patient-specific information with physiological understanding or statistical inference to generate insights beyond conventional diagnostics. This review traces how computational modelling has evolved from theoretical research tools into clinical decision support systems that enable personalised cardiovascular care. We examine this evolution across three key domains: enhancing diagnostic accuracy through improved measurement techniques, deepening mechanistic insights into cardiovascular pathophysiology and enabling precision medicine through patient-specific simulations. The review covers the complementary strengths of data-driven approaches, which identify patterns in large clinical datasets, and knowledge-driven models, which simulate cardiovascular processes based on established biophysical principles. Applications range from artificial intelligence-guided measurements and model-informed diagnostics to digital twins that enable in silico testing of therapeutic interventions in the digital replicas of individual hearts. This review outlines the main types of cardiovascular modelling, highlighting their strengths, limitations and complementary potential through current clinical and research applications. We also discuss future directions, emphasising the need for interdisciplinary collaboration, pragmatic model design and integration of hybrid approaches. While progress is promising, challenges remain in validation, regulatory approval and clinical workflow integration. With continued development and thoughtful implementation, computational models hold the potential to enable more informed decision-making and advance truly personalised cardiovascular care.
14 May 2026
Read appraisal →Progress in biomedical engineering (Bristol, England)
Cross-subject generalization for EEG decoding: a survey of deep learning methods
Deep learning for cross-subject electroencephalography (EEG) decoding is hindered by high inter-subject variability, which introduces a severe domain shift between training and unseen test subjects. This survey presents a comprehensive review of deep learning methodologies specifically engineered to address this cross-subject generalization challenge. To ground this analysis, we formalize the cross-subject setting as a multi-source domain problem and delineate the rigorous, subject-independent evaluation protocols required for valid assessment. Central to this survey is a systematic taxonomy of the current literature into discrete methodological families, including feature alignment, adversarial learning, feature disentanglement, and contrastive learning. We conclude by examining three critical elements for advancing robust, real-world decoding: the theoretical limitations of current methodologies, the structural value of subject identity, and the emergence of EEG foundation models.
14 May 2026
Read appraisal →Biomedical physics & engineering express
Differentiating depression subtypes through three-electrode EEG and emotional stimuli: a machine learning approach
We propose an EEG-based framework for depression subtype assessment using emotion-modulated neural dynamics elicited by immersive virtual reality (VR). EEG was recorded from 70 participants (31 first depressive episode, FDE; 18 recurrent depressive episode, RDE; 21 control participants, HC) using a compact frontal montage (Fp1/Fpz/Fp2) during positive and negative VR conditions, focusing on the immediate post-stimulus regulation period. Subject-level features were analyzed with a linear mixed-effects model to probe group-by-condition interaction patterns, and a Bi-Emotional Siamese Network (BESN) was developed to model baseline-anchored within-subject deviations by fusing positive/negative reactivity streams with path-signature-based temporal encoding. Feature-level analyses revealed interaction signals suggestive of differential emotion-related modulation across groups. In subject-wise evaluation, BESN achieved 83.1% accuracy for HC versus FDE discrimination and 70.6% accuracy for three-class classification (HC versus FDE versus RDE), outperforming conventional machine-learning baselines. Robustness was further supported on an external public resting-state EEG dataset (MODMA), achieving 78.2% accuracy. These results suggest that baseline-anchored, emotion-modulated EEG dynamics combined with path-signature modeling provide an objective and generalizable computational approach for depression-related group discrimination.
14 May 2026
Read appraisal →Physiological measurement
Turning glasses into sensors: capacitive facial expression recognition under frame geometry constraints
Objective.Wearable facial expression recognition requires sensing modalities that are unobtrusive, illumination-independent, and compatible with everyday eyewear form factors. Existing optical and electromyographic approaches are limited by line-of-sight requirements, skin contact, or restricted action unit (AU) coverage. This work investigates the feasibility of non-contact capacitive sensing for facial muscle activity monitoring within the strict geometric constraints of standard eyewear.Approach.We developed FERGlasses, integrating eight loading-mode capacitive channels via a custom flexible PCB and active shielding to detect skin-to-frame displacement. A validation study with 40 participants was conducted to elicit seven facial action coding system-defined AUs. Signal fidelity was assessed using independent Random Forest regressors and benchmarked against commercial optical-proximity (OCOsense) and computer-vision (Py-Feat) standards.Main results.The system demonstrated high agreement for upper-face activity, achieving median cosine similarities of 0.97 for the Outer Brow Raiser (AU02) and 0.87 for the Nose Wrinkler (AU09), while lower-face actions such as the Chin Raiser (AU17) yielded a median of 0.33, consistent with mechanical signal attenuation along the frame. Channel-ablation analysis quantified the specific anatomical contributions of each electrode group, identifying eyebrow and rhinion sensors as primary signal drivers for upper-face and perinasal actions, respectively, and temple sensors as critical for cheek raiser detection.Significance.This study provides the first metrological validation of multi-channel capacitive sensing for AU-level recognition in eyewear. The results establish the modality as a robust, privacy-preserving, and illumination-independent alternative to optical methods for longitudinal physiological monitoring, with the current prototype serving as a wired research platform and full consumer-grade integration identified as a future engineering objective.
14 May 2026
Read appraisal →Angewandte Chemie (International ed. in English)
A Novel Acetylcholine Nanosensor for Single Vesicle Storage and Sub-Quantal Exocytosis in Living Neurons and Organoids
Acetylcholine (ACh) is a critical neurotransmitter that regulates diverse physiological functions, such as cognition and muscle contraction, through synaptic transmission. However, in situ quantitative chemical analysis of single-vesicle storage and release dynamics at single-cell level remains a major technical challenge, hindering mechanistic understanding of cholinergic synaptic plasticity across physiological and pathological states. To address this, we developed an Ti3C2Tx MXene/enzyme-functionalized (M@E@CF) nanosensor, enabling real-time monitoring of vesicular ACh storage and exocytotic release in primary cholinergic neurons and human spinal cord organoids. Our findings reveal a sub-quantal release mode in both mouse and human-derived neurons. To further elucidate the regulatory principles of exocytosis kinetics, we classified single-vesicle exocytosis patterns based on signal peak shapes using a 1D convolutional neural network (1D-CNN) deep learning model, uncovering significant differences in the number of released molecules and kinetic parameters across modes. Critically, in Down syndrome models, we observed significantly reduced single vesicle ACh storage and release alongside an elevated release fraction, concurrent with shortened fusion pore durations during exocytosis. Thus, the M@E@CF nanosensor platform establishes a versatile tool for spatiotemporal investigation of neurotransmitter storage and release dynamics, providing a critical technical foundation for exploring physiological functions and pathological mechanisms of the cholinergic system.
13 May 2026
Read appraisal →Neuroinformatics
Instance-Aware Pseudo-Labeling and Class-Focused Contrastive Learning for Weakly Supervised Domain Adaptive Segmentation of Electron Microscopy
Annotation-efficient segmentation of the numerous mitochondria instances from various electron microscopy (EM) images is highly valuable for biological and neuroscience research. Although unsupervised domain adaptation (UDA) methods can help mitigate domain shifts and reduce the high costs of annotating each domain, they typically have relatively low performance in practical applications. Thus, we investigate weakly supervised domain adaptation (WDA) that utilizes additional sparse point labels on the target domain, which require minimal annotation effort and minimal expert knowledge. To take full use of the incomplete and imprecise point annotations, we introduce a multitask learning framework that jointly conducts segmentation and center detection with a novel cross-teaching mechanism and class-focused cross-domain contrastive learning. While leveraging unlabeled image regions is essential, we introduce segmentation self-training with a novel instance-aware pseudo-label (IPL) selection strategy. Unlike existing methods that typically rely on pixel-wise pseudo-label filtering, the IPL semantically selects reliable and diverse pseudo-labels with the help of the detection task. Comprehensive validations and comparisons on challenging datasets demonstrate that our method outperforms existing UDA and WDA methods, significantly narrowing the performance gap with the supervised upper bound. Furthermore, under the UDA setting, our method also achieves substantial improvements over other UDA techniques.
13 May 2026
Read appraisal →Biofabrication
Resolving the multiscale design trade-off in bone tissue engineering: from generative design to digital twins
The clinical translation of bone tissue engineering is constrained by a severe multiscale design trade-off: the requirement for load-bearing mechanical strength competes directly with the need for vascular-permissive porosity. Conventional approaches address scaffold geometry, material formulation, and fabrication in a sequential, decoupled manner and therefore fail to navigate this multi-objective conflict. To overcome this limitation, we propose an 'AI-Driven Holistic Intelligence' framework. It spans the entire pipeline from generative inverse design to digital twin (DT)-based predictive maturation. The framework integrates gradient scaffold optimization, closed-loop adaptive manufacturing, and virtual clinical trial within a unified workflow. This architecture is designed to co-optimize the boundary conditions of stiffness and perfusion simultaneously, rather than in isolation. Nonetheless, realizing this vision as a clinically viable autonomous regeneration pathway will require bridging persistent translational gaps in data standardization, algorithm interpretability, and regulatory alignment. Importantly, the core innovation of this framework is not merely to navigate the empirical 'trade-off' along a fixed Pareto front, but to resolve it through multiscale co-optimization that expands the accessible design space, enabling simultaneous improvements in mechanical strength and vascular perfusion beyond the limits achievable by conventional single-objective approaches.
13 May 2026
Read appraisal →Biomedical physics & engineering express
Vitiligo state assessment based on progressive transfer learning and multimodal domain adaptation
Vitiligo is a common skin depigmentation disorder; assessing its state is crucial for the treatment outcome. Collecting multimodal data for vitiligo assessment is complex and costly in clinical practices, and the limited data size restricts the performance of deep learning models. Transfer learning can alleviate the shortage of training data in medical image recognition, however, its applications in vitiligo state assessment are constrained by feature differences between natural and medical images and insufficient generalization to different modalities. To address these challenges, this paper introduces a vitiligo state assessment method based on progressive transfer learning and multimodal domain adaptation. The scheme uses a large set of unlabeled medical images as a bridge to reduce the discrepancies between natural and medical images through multi-step fine-tuning. An adaptive parameter unfreezing strategy is then applied for the accurately labeled target data to improve the adaptability and accuracy of the model. In addition, an integrated multimodal domain adaptation approach based on CycleGAN and hue, saturation, value color space transformation is proposed to reduce the impact of modality differences on transfer learning. Experimental results demonstrate that compared to the conventional transfer learning method, the proposed method improves the accuracy by 2.4% and 3.6% on the clinical and Wood's lamp modality datasets, respectively. This accurate vitiligo state assessment method can be applied to a wide range of multimodal dermatological diseases where labeled data is limited.
13 May 2026
Read appraisal →Lasers in medical science
Machine learning enhanced optical spectroscopy for breast cancer diagnosis: A review
This literature review examines the transformative role of machine learning (ML) and deep learning (DL) in enhancing optical spectroscopy for breast cancer diagnosis. By synthesizing advancements from peer-reviewed studies (2015-2025), we evaluate how ML/DL integration improves the detection of malignancy-associated biochemical changes, enabling noninvasive, rapid, and accurate differentiation between healthy and cancerous tissues. This review highlights key spectroscopic modalities, such as Raman, fluorescence, diffusive optical spectroscopy (DOS), and photoacoustic spectroscopy (PAS), and their integration with AI-driven models, such as convolutional neural networks (CNNs), support vector machines (SVMs), and logistic regression. These techniques achieve diagnostic accuracies of up to 94% in subtype classification (e.g., luminal A, HER2-positive) by analyzing spectral biomarkers such as hemoglobin, lipids, and collagen. Challenges such as data variability, model interpretability, and clinical integration barriers are critically assessed. These findings underscore the potential of ML/DL-enhanced spectroscopy to standardize diagnostics, reduce unnecessary biopsies, and personalize treatment monitoring. Future directions emphasize the need for explainable AI (XAI), multimodal data fusion, and large-scale, diverse datasets to bridge translational gaps. By addressing technical, ethical, and regulatory hurdles, this integration promises to advance early detection, improve clinical outcomes, and reshape precision oncology.
11 May 2026
Read appraisal →Biomaterials science
Hydrogel-based wearable and implantable biosensors in health monitoring.
Hydrogel-based wearable and implantable biosensors have quickly established themselves as a paradigm-shifting platform for continuous health monitoring. They effectively reconcile the mechanical and chemical discrepancies between conventional rigid electronics and delicate biological tissues. This success is largely attributed to the hydrogel material's inherent advantages: high water content, a precisely tunable elastic modulus, and versatile functional chemistry. Collectively, these features provide structural, biological, and processing benefits that facilitate truly seamless bioelectronic integration. This Review systematically summarizes the latest advancements in hydrogel biosensors, placing emphasis on their distinctive material attributes. These include the synergistic ionic-electronic conduction, sensitive stimuli responsiveness, advanced antifouling surface chemistry, and exceptional tissue-mimicking compliance. We further delve into the fundamental physical and chemical sensing mechanisms, supported by representative applications spanning from the non-invasive, real-time analysis of sweat, tears, saliva, and interstitial fluids to the more complex domains of subcutaneous, cardiac, and neural implants. Finally, we candidly address the crucial hurdles that remain, such as long-term hydration stability, signal fidelity drift, and adverse immune responses. Concurrently, we highlight pioneering strategies to overcome these issues, including the adoption of zwitterionic designs for enhanced biocompatibility, nanocomposite reinforcement for mechanical robustness, and the utilization of transient biodegradability. By critically elucidating the intricate relationships governing hydrogel structure, chemistry, and bioelectronic function, this Review aims to chart the research trajectory toward the next generation of durable, self-healing, and fully biointegrated sensing systems.
11 May 2026
Read appraisal →JMIR formative research
At-Home Sleep Electroencephalography Assessment in Young and Older Adults Using a Novel Wireless Soft Electronics Sleep Monitoring System: Experimental Study
BACKGROUND: Sleep quality declines with age and is a known contributor to multiple chronic health conditions, including Alzheimer disease. Emerging evidence suggests that certain electroencephalography (EEG) neural signatures measured during sleep may be predictive of cognitive decline in older adults. Sleep EEG signals are traditionally measured using bulky, rigid, and uncomfortable equipment in an unfamiliar laboratory setting, which can negatively impact sleep signals. Due to these limitations, sleep EEG data acquisition is typically limited to a single night. OBJECTIVE: This study aimed to validate our recently developed portable, skin-like EEG monitoring patch for 7 nights in the home environment in a pilot sample of young and older adults by evaluating usability and acceptance, and replicating age-related differences in sleep architecture observed in the polysomnography literature. METHODS: Eighteen young adults and 18 cognitively unimpaired older adults without sleep disorders were enrolled (data from 11 young adults and 12 older adults were included in the analyses) in a 7-night study during which they wore novel, gel-free, wireless, ultrathin, skin-conforming, sleep monitoring, fabric-based patches. These patches were self-applied to the forehead and face for optimal usability and comfort. The patches incorporate laser-cut mesh electrodes with low-profile electronics (including a rechargeable battery and amplifier) and transmit EEG signals to a participant-controlled, Bluetooth-enabled, tablet-based data acquisition app. An automated algorithm was used to stage sleep and assess microarchitecture features from the EEG commonly impacted for each participant. Averages across nights were computed for these sleep features for each participant. RESULTS: Young and older adults reported that the sleep patch was easy to use and comfortable to wear. There was no loss of signal power over 7 nights of wear across participants (retained-data signal-to-noise ratio over the 7-d period: young adult, mean 20.69, SD 12.78, maximum 52.13, minimum 5.19; older adult, mean 22.10, SD 9.39, maximum 49.96, minimum 13.79). Most datasets not retained were lost due to poor reference electrode adhesion on the nose (75/101, 74% of lost datasets in young adults and 57/88, 65% in older adults). Trained sleep technologists verified that the retained datasets were of sufficient quality to be scored without difficulty. Expected age-group differences in sleep features were observed, including age-related reductions in stage N3 sleep (young adult, mean 18.55, SD 6.70; older adult, mean 10.40, SD 6.43; Mann-Whitney U=42.0; P=.01) and reduced sleep spindle density (young adult, mean 2.92, SD 2.24; older adult, mean 0.94, SD 1.33; Mann-Whitney U=45.0; P=.006). CONCLUSIONS: This study demonstrates that our novel, comfortable, wearable patch can reliably measure physiological sleep data over multiple nights at home in adults across the lifespan, thereby making multinight sleep assessment in cognitive aging studies and clinical research more accessible than traditional polysomnography. In future studies, the small, lightweight system, which is highly scalable, can be shipped inexpensively to participants' homes, making this technology and research accessible to individuals who may have difficulty traveling or who are hesitant to travel to a laboratory or clinic.
11 May 2026
Read appraisal →Journal of neural engineering
Uncovering relationships in multi-channel EEG data using principal Hessian directions and Ricci flow
Objective.The high-dimensional nature of multi-channel EEG data poses major challenges for downstream classification. Uncovering connectivity relationships and network structure between EEG channels can improve high-dimensional EEG signal recovery by identifying redundant inputs and guiding more targeted feature selection. In this study, we aim to demonstrate a systematic framework for extracting useful interchannel structure in EEG data.Approach.We present two complementary methods for inferring network structure and connectivity in EEG data: (1) a new supervised algorithm for inferring classification-relevant community structure based on principal Hessian directions (pHds), and (2) a discrete Ricci flow-based unsupervised community detection algorithm. We demonstrate these systematic methods on high-dimensional real-world EEG datasets involving classifying imagined digits versus non-digits and detecting emotional valence.Main results.We show that our pHd and Ricci flow methods-when combined-can detect interchannel relationships that meaningfully hold on unseen test classification data. Moreover, we demonstrate that this interchannel structure extracted via pHd and Ricci flow can enable subsequent improvements in downstream EEG signal classification.Significance.Our combined pHd-Ricci flow method expands the existing EEG preprocessing toolkit by offering a systematic framework for extracting meaningful network structure from high-dimensional EEG data. By facilitating more targeted and effective feature selection, our method has the potential to improve EEG signal recovery in real-world applications.
11 May 2026
Read appraisal →Analytical chemistry
TPP-Fabricated All-Fiber Nanoforce Sensor with Deep Learning Analysis Enables Ultrasensitive Cancer Cell Mechanophenotyping
Cellular mechanics provide sensitive, quantitative markers of cancer progression, yet routine nanoforce measurements remain constrained by the cost, footprint, and electromagnetic limitations of conventional platforms. Here we report an all-fiber nanoforce sensor (AFNS) that integrates a two-photon-polymerized (TPP) microcantilever on the end facet of a multimode fiber, establishing a compact and robust analytical strategy for ultrasensitive force measurements. The printed structure comprises a rigid base, a ∼2 μm-thick cantilever disk, and a probe tip. Finite-element analysis is employed to estimate the effective spring constant and to optimize mechanical compliance. Force-dependent multimode specklegrams are quantitatively decoded using an EfficientNet-based regression model, enabling highly accurate force readout (R2 = 0.9999; mean absolute error <0.2%). The sensor achieves a force resolution of 0.41 nN and maintains stable analytical performance over 100 repeated loading cycles, representing a rigorous repeatability assessment for this class of fiber-tip force sensors. The analytical capability of the AFNS is validated by quantifying the Young's modulus of polydimethylsiloxane and by mechanically phenotyping A549 and HepG2 cells at different invasion stages, showing strong agreement with atomic force microscopy. With low cost, replaceable probes, and a compact all-fiber configuration, this work provides a scalable analytical methodology for high-precision micro/nano-force characterization and quantitative cellular mechanics assays.
11 May 2026
Read appraisal →Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy
Raman spectroscopy with machine-learning classification for the prediction of stereotactic radiotherapy induced treatment toxicity in high-risk localised prostate cancer
Radiotherapy can lead to late-onset toxicity, to varying extents between individuals due to differences in radiosensitivity. Predicting which patients are most at risk is key to augmenting the therapeutic window. However, the underlying biological mechanisms remain poorly understood, and current experimental methods often lack clinical applicability. This study employs Raman spectroscopy to analyse biochemical profiles in peripheral lymphocytes and plasma, aiming to monitor radiotherapeutic response and predict intrinsic radiosensitivity in high-risk localised prostate cancer patients treated with stereotactic radiotherapy. Partial-least squares discriminant analysis classification of Raman spectra at baseline (n = 20) from post-hormone therapy (n = 19), mid-treatment (pre-4th fraction; n = 21) and 3-months after treatment (n = 18) returned mean area under the curve values ranging from 0.88 to 0.93. Ensemble classifiers applied to imbalanced late toxicity datasets (grade 0-1, n = 16; grade 2+, n = 4) yielded mean F1 scores of 0.74 (random forest, lymphocytes) and 0.69 (AdaBoost, plasma); metrics based on best performing model for minority-class. Classical least squares lymphocyte and plasma toxicity models identified major concentration differences in amino acids, proteins, lipids, DNA and related biomolecules (p < 0.05). These findings demonstrate the potential of Raman spectroscopy as a minimally invasive, objective tool for classifying blood-based biochemical profiles across radiotherapy treatment time points and distinguishing patients with late grade 0-1 and grade 2+ toxicity.
10 May 2026
Read appraisal →Biomedical physics & engineering express
Enhancing CNN regressors with contour encoding and self-supervision for improved 3D/2D x-ray to CT registration in spinal surgery navigation
With advances in deep learning, regression-based methods have shown promising results in 3D/2D medical image registration. However, strict intraoperative radiation dose constraints produce low-dose x-ray images with severe blur and reduced contrast, significantly degrading registration accuracy and limiting precise image-guided spinal interventions. We propose the Contour Encoding Regressor (CER), a novel end-to-end CNN framework that extracts highly discriminative features directly from binary contour masks of intraoperative x-rays without any restrictions on contour length, shape, or morphology. These contour features are efficiently encoded by a dedicated module and seamlessly fused into the regression pipeline to improve robustness against image degradation. To further enhance pose estimation, CER employs a dual-branch architecture that explicitly decouples rotational and translational parameters, thereby reducing mutual interference and improving overall accuracy. In addition, a self-supervised fine-tuning strategy with a tailored multi-component loss function is introduced to adapt the model to blurred low-dose conditions and minimize residual errors. On low-dose x-ray images, CER achieves a mean target registration error of 1.39 mm-a clinically acceptable threshold-while outperforming state-of-the-art methods in accuracy and enabling real-time performance (0.03-0.06 s per frame on clinically accessible GPUs). These improvements meet the stringent precision and speed requirements of intraoperative navigation, offering strong potential to enhance surgical safety and outcomes in minimally invasive spinal procedures.
10 May 2026
Read appraisal →Biomedical physics & engineering express
Dosimetric evaluation of simultaneous multi-energy (2-18 MV) and intensity optimization for IMRT and VMAT.
Objective.This study aimed to develop and evaluate a framework for the simultaneous optimization of beam fluence and multiple photon energies beyond dual-energy configurations for intensity modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT), with the goal of improving treatment plan quality and normal tissue sparing.Approach.An Elekta Versa HD linear accelerator (linac) was modeled using the BEAMnrc/EGSnrc Monte Carlo platform to simulate 2 MV, 4 MV, and 6 MV photon beams, complemented with measured beam data for 10 MV and 18 MV. A custom MATLAB-based framework was developed to perform simultaneous optimization of energy and fluence, allowing energy weighting of individual beamlets. A total of 12 various energy configurations including single (SE), dual (DE), triple (TE), and quadruple (QE)-energy combinations for IMRT and VMAT were generated for thirteen prostate cancer patients. Plans were evaluated using dose-volume histograms, homogeneity index, conformity index, and organ-at-risk (OAR) dose metrics.Main results.Gamma analysis confirmed the high degree of agreement between measured and simulated beam data, which validated our Monte Carlo linac model. Target coverage remained comparable across all energy configurations. Compared with 6 SE, multi-energy (ME) IMRT and VMAT achieved superior OAR sparing, with the greatest benefits for bladder and rectum in the 5-40 Gy range. TE-IMRT reduced bladder and rectal mean doses by up to 1.6 Gy, with V10 reductions of ∼8%; on average, mean doses were reduced by 0.9 Gy (p= 0.001) and 1.0 Gy (p= 0.001). For VMAT, the 6&18 DE-VMAT plan yielded the largest sparing, lowering bladder and rectal mean doses by ∼1.0 Gy. Relative to 10 SE, all ME plans achieved equal or lower OAR doses, with TE-IMRT and 6&18 DE-VMAT showing the most consistent benefit, reducing mean doses to the bladder, rectum, right femoral head, and left femoral head by up to 1.26, 0.77, 0.6, and 0.55 Gy, respectively.Significance.This study presents the first assessment of the feasibility and potential clinical value of simultaneous optimization of beam intensity and more than two photon energies for IMRT and VMAT plans, producing optimized ME fluence maps that require subsequent conversion into clinically deliverable treatment plans. The findings demonstrate that ME optimization can enhance OAR sparing without compromising target coverage, supporting its promise for future clinical application.
10 May 2026
Read appraisal →Carbohydrate polymers
Multifunctional starch-based eutectogels with high mechanical properties, adhesion, self-healing, fatigue resistance and degradability for flexible wearable sensing
Flexible wearable sensors have developed rapidly due to the urgent demands for human motion monitoring. Eutectogel-based soft materials have exhibited great potential for wearable sensing applications. However, it is still a challenge to develop multifunctional eutectogels with excellent mechanical properties, adhesion, self-healing, fatigue resistance and degradability. In this work, a multifunctional starch-based eutectogel was developed by a one-pot and in-situ photo-polymerization method. The eutectogel had excellent stretchability (2710%), high strength (5.43 MPa), toughness (94.38 MJ/m3), Young's modulus (10.23 MPa), adhesion, fatigue resistance (1000 cycles), degradability and a certain degree of self-healing ability. These properties were closely correlated to the synergistic effect among the amylopectin, polyacrylic acid and choline chloride via hydrogen-bond and electrostatic interactions. The mechanical properties of eutectogel can be adjusted by changing the starch content. A wearable sensor based on the eutectogel was realized accurate and stable monitoring of human joint motions. This study addressed the limitations of the dehydration and single function of the traditional hydrogels, and provides a novel way for the development of the multifunctional bio-based gels for next-generation flexible wearable devices.
9 May 2026
Read appraisal →Medical image analysis
Context-enriched contrastive auto-encoder with topology learning for medical hyperspectral image classification to diagnose tumors
Deep learning has emerged as a highly effective approach for the automatic classification of medical hyperspectral images (MedHSIs), facilitating the accurate diagnosis of diverse tumors. Most of current methods suffer from the challenges in the separability of latent representation and the preservation of topological structure. To remedy such deficiencies, we propose a novel context-enriched contrastive auto-encoder with topology learning (CCAET) based MedHSI classification for tumor diagnosis. Concretely, we initially employ an auto-encoder (AE) network as the foundational framework to achieve latent representation and data reconstruction. Subsequently, to enhance the separability of the latent representation, we integrate a context-enriched contrastive loss into the AE framework, where both individual pixels and their contextual information are used to redefine contrastive learning, so that small distances between intra-class features and large distances between inter-class features are achieved. Further, we construct a topology-preserved loss by incorporating a graph based approach into the foundational framework, which aims to maintain the topological structure during latent representation learning. Furthermore, we devise a gradient descent-based optimization scheme to optimize the overall loss of the proposed CCAET method, so as to obtain the final latent representation of MedHSI. Finally, we implement soft-max-based label prediction to obtain the classification results. Different from most existing methods, the CCAET approach considers both the separability of latent representations and the preservation of topological structure to enhance MedHSI classification performance. Extensive experimental evidence has been presented to substantiate that the proposed CCAET method outperforms several state-of-the-art techniques in terms of MedHSI classification based tumor diagnosis.
9 May 2026
Read appraisal →Epilepsy research
Machine learning in epilepsy
Epilepsy is a complex neurological disorder characterized by pathological processes that unfold across multiple biological scales, from cellular excitability and synaptic integration to large-scale network dynamics observable in electroencephalographic (EEG) recordings. While traditional analytical approaches have provided valuable insights, they often fail to capture high-dimensional and nonlinear structure of contemporary electrophysiological and clinical datasets. Consequently, machine learning (ML) has emerged as a powerful analytical framework in epilepsy research, although its rapid adoption has revealed a growing gap between algorithmic performance and biological interpretability. This review examines ML methods operating across three analytically distinct yet interconnected levels: (i) unsupervised learning for cellular-level phenotyping using high-dimensional electrophysiological data; (ii) supervised learning for EEG-based seizure detection and prediction; and (iii) multiscale modeling frameworks integrating neuronal and network dynamics. Rather than providing an exhaustive catalog of algorithms, we focus on inferential assumptions underlying ML applications, the methodological pitfalls constraining generalization and clinical relevance, and how ML-derived representations can be interpreted within established neurophysiological theory. We highlight that unsupervised ML facilitates identification of latent excitability phenotypes and trajectories obscured in traditional univariate analyses, while supervised ML has substantially advanced automated seizure detection and prediction, despite persistent challenges related to data leakage, class imbalance, and ambiguous preictal labeling. We argue that the most promising direction lies in embedding ML within multiscale mechanistic models, where data-driven inference facilitates parameter estimation and hypothesis generation rather than black-box prediction. By prioritizing interpretability, rigorous validation, and cross-scale integration, ML-enhanced multiscale frameworks offer a path toward clinically actionable models of epilepsy.
9 May 2026
Read appraisal →ChemistryOpen
Ionic Hydrogel Triboelectric Sensor for Real-Time Swimming Kinematics Monitoring and Stroke Recognition
Recently, wearable sports electronics have attracted increasing attention for their potential to enable continuous motion tracking and energy-autonomous operation in complex athletic environments. In this work, we present a flexible triboelectric nanogenerator (PAC-TENG) constructed from a PVA/Ala/CaCl2 ionic hydrogel electrode and a compliant silicone triboelectric layer. The PAC hydrogel provides high stretchability, stable ionic conductivity, and strong deformation tolerance, while the silicone surface ensures efficient and repeatable charge generation. Benefiting from this synergistic design, the PAC-TENG achieves excellent electrical outputs, including a peak voltage of 472.9 V, a short-circuit current of 37.1 µA, and a transferred charge of 104.1 nC, along with high power density under matched load. The device exhibits strong sensitivity to bending angle, applied force, and motion frequency, enabling accurate perception of human biomechanics. When mounted on swimming-related joints, the PAC-TENG can reliably monitor stroke patterns and detect abnormal motion signals, highlighting its promise for self-powered aquatic sports analytics and intelligent safety monitoring.
6 May 2026
Read appraisal →Annual review of biomedical engineering
Digital Twins for Biofluids
Digital twins-virtual representations dynamically linked to physical systems-have the potential to transform biomedical engineering by enabling real-time prediction, optimization, and personalization in health and disease. In biofluids, digital twins offer a framework for integrating physics-based models with data from clinical imaging, sensors, and physiological measurements to support diagnostics, therapeutic planning, and device design. This article reviews modeling approaches used in the construction of digital twins for biofluid applications. We survey high-fidelity numerical methods alongside emerging machine learning techniques, highlighting their respective strengths and limitations. Key requirements for digital twins are discussed, emphasizing the bidirectional interaction between physical and virtual assets and the importance of selecting modeling strategies tailored to specific biomedical contexts. While notable progress has been made over the past decade, significant challenges remain, particularly in integrating multiphysics models with data-driven methods and in establishing standardized protocols for data acquisition, interoperability, and sharing.
5 May 2026
Read appraisal →Annual review of biomedical engineering
Physics-Informed Machine Learning in Biomedical Science and Engineering
Physics-informed machine learning (PIML) is emerging as a potentially transformative paradigm for modeling complex biomedical systems by integrating parameterized physical laws with data-driven methods. Here, we review three main classes of PIML frameworks: physics-informed neural networks (PINNs), neural ordinary differential equations (NODEs), and neural operators (NOs), highlighting their growing role in biomedical science and engineering. We begin with PINNs, which embed governing equations into deep learning models and have been successfully applied to biosolid and biofluid mechanics, mechanobiology, and medical imaging, among other areas. We then review NODEs, which offer continuous-time modeling, especially suited to dynamic physiological systems, pharmacokinetics, and cell signaling. Finally, we discuss deep NOs as powerful tools for learning mappings between function spaces, enabling efficient simulations across multiscale and spatially heterogeneous biological domains. Throughout, we emphasize applications where physical interpretability, data scarcity, or system complexity make conventional black-box learning insufficient. We conclude by identifying open challenges and future directions for advancing PIML in biomedical science and engineering, including issues of uncertainty quantification, generalization, and integration of PIML and large language models.
5 May 2026
Read appraisal →Science advances
SkinECG: An orthogonal remote powering wearable skin-like sensor
The Internet of Things (IoT) era witnessed a surge in wearable devices for the human body, each requiring a power source that supports long-term operation. However, batteries are depletable and require replacement, while the effectiveness of energy-harvesting techniques is location specific. We developed the SkinECG system, an orthogonal remote powering skin-like electrocardiogram (ECG) sensor. It is powered by body-coupled powering (BCP) using power acquired optimally via suitable energy harvesters at different locations on the human body, solving the location mismatch between the energy-harvesting source and sensor. The remote Orthogonal Energy Harvesting Network (O-EHN) effectively transfers power from multiple energy-harvesting sources without internode interference and cancellation. We demonstrated the 0.71-mm-thin, flexible SkinECG system that conforms to the human chest and detects ECG signals without any cumbersome batteries or directly connected power for long-term health monitoring.
5 May 2026
Read appraisal →Science advances
Augmenting ultrasound for continuous glucose monitoring via a wearable acoustically readable microneedle patch
Continuous glucose monitoring (CGM) represents substantial advancement, yet poses a challenge in wearable health care development. Current enzymatic CGM faces limitations in stability, cost, and durability. We introduce an enzyme-free, wearable acoustically readable microneedle patch (ARMPatch) composed of glucose-responsive hydrogel. Positioned between a standard ultrasound probe and the skin, ARMPatch acts as an acoustic interface enabling CGM using conventional ultrasound. Its hydrogel microneedles minimally penetrate the epidermis, allowing interstitial fluid to trigger variable swelling of the microneedles in response to glucose fluctuations. ARMPatch delivers stable and selective glucose readings for up to 56 days. Glucose variations with millimolar resolution are obtained via ultrasound within a response time of 30 to 60 minutes. In vivo CGM in animal models for 7 days reveals a reversible correlation between microneedle swelling and glucose variations. Acting as "accessory" for standard ultrasound probes, this approach offers a minimally invasive, cost-effective, and long-lasting solution to non-enzymatic CGM, expanding the utility of ultrasound in wearable biosensing.
5 May 2026
Read appraisal →Journal of advanced research
Flexible wearable medical devices: from material innovations and data processing to intelligent healthcare applications
BACKGROUND: Flexible wearable medical devices drive healthcare transformation via non-invasive, real-time physiological monitoring and personalized management. Traditional rigid devices lack long-term comfort, while chronic disease care and telemedicine demand reliable, patient-centered solutions. Advances in materials (carbon nanomaterials, liquid metals, hydrogels) enable stretchable, biocompatible substrates adapting to bodily movements. Innovative sensing (PPG, EEG, sweat-based detection) tracks blood pressure, glucose, and neural signals accurately. Key challenges include insufficient sensor flexibility, lagging supervision, data security issues, and large-scale production difficulties, requiring interdisciplinary collaboration. AIM OF REVIEW: This review integrates cutting-edge advances in flexible wearable medical devices, covering material innovations, sensing breakthroughs, power management, and AI-driven data analysis. It evaluates strategies to enhance biocompatibility of carbon composites, liquid metals, and hydrogels, analyzing improvements in device sensitivity, biocompatibility, and real-world applicability. It critically discusses machine learning's role and limitations in chronic disease management, telemedicine, and sports optimization. Key unresolved challenges include lagging regulations, privacy risks in multi-source data fusion, mismatched sensing development, and large-scale production difficulties. The aim is to provide a scalable, secure, and clinically translatable roadmap for future research. KEY SCIENTIFIC CONCEPTS OF REVIEW: Core technologies involve carbon nanomaterials (e.g., graphene), liquid metals (structurally optimized for balanced conductivity and flexibility), and hydrogels (preferred as flexible substrates for mimicking natural soft tissues' mechanics and biocompatibility). Sensing technologies (PPG, EEG, sweat-based detection) enable non-invasive, real-time monitoring of vital parameters like blood pressure, glucose, and neural signals. AI decodes complex physiological data for predictive diagnosis, while IoT/5G integration advances telemedicine. Current limitations include poor sensor flexibility, lagging regulations, and large-scale production hurdles, contrasting with trends like advanced sensing, new materials, and hybrid diagnostic-therapeutic systems-highlighting flexible wearables' potential in intelligent healthcare.
5 May 2026
Read appraisal →Magnetic resonance in medicine
Dynamic-Guided Diffusion Probability Model for Cranial Nerves Segmentation
PURPOSE: Segmentation of cranial nerves (CNs) bundles using magnetic resonance imaging (MRI) provides a valuable quantitative approach for analyzing the morphology and orientation of individual CNs. Currently, the CN regions can be segmented directly using deep learning-based methods. However, existing methods overlook the unique characteristics of CNs, particularly their environmental features and representation in multimodal images that may lead to suboptimal segmentation outcomes. METHODS: We proposed a dynamic-guided diffusion probability model for CNs segmentation, which enhances segmentation performance by integrating the intrinsic characteristics of CNs. A dynamic-guided mechanism approach called the SE-A-NL module was proposed. The module is capable of addressing both the varying characterization abilities of multimodal images and the long-range connections of CNs within images. RESULTS: Quantitative and qualitative experiments demonstrate that the proposed method surpasses current state-of-the-art approaches, delivering accurate and effective segmentation of five pairs of cranial nerves. Notably, the method outperforms existing techniques in 16 out of the 20 evaluated metrics. CONCLUSION: The overall network model effectively integrates multimodal information and anatomical priors by combining multi-channel attention and non-local attention mechanisms, thereby improving CNs segmentation performance. Thorough comparative and ablation studies highlight the superior performance of the proposed method.
5 May 2026
Read appraisal →Biosensors & bioelectronics
Peptide-responsive photonic hydrogels integrated with deep learning assistance for early MMP-9 detection
Matrix metalloproteinase-9 (MMP-9) is crucial for extracellular matrix remodeling, and its dysregulation is associated with inflammatory diseases and different forms of cancer. Conventional MMP-9 detection methods such as enzyme-linked immunosorbent assay (ELISA) are limited by complexity, expensive equipment, and lengthy antibody incubation times. These limitations prevent their use in point-of-care testing. An MMP-9 responsive photonic crystal (PC/PEG-M9SP) hydrogel has been developed to address these challenges. The hydrogel is synthesized from 4-arm polyethylene glycol-acrylate and an MMP-9 sensitive peptide via Michael-type addition reaction. Upon MMP-9-specific enzymatic cleavage, the hydrogel undergoes a structural reconfiguration, resulting in a distinct color shift. Integrated with a deep learning-based smartphone app, this platform enables both visual and quantitative detection within 10 min, achieving high sensitivity (10.60 nm mL/ng) and a detection limit of 0.62 ng/mL. Validation in complex biological fluids demonstrated strong concordance with ELISA, confirming the analytical reliability of the hydrogel. This system provides a rapid, portable, and cost-effective solution for accurate MMP-9 detection, with strong potential for clinical and point-of-care applications.
5 May 2026
Read appraisal →Neural networks : the official journal of the International Neural Network Society
MMC-CS: Multi-branch multi-stage contrastive learning for self-supervised compressed sensing
In recent years, optimization-inspired networks that integrate optimization theory into deep neural networks (DNNs) have achieved remarkable success in image compressed sensing (ICS). However, current DNN-based ICS methods face two key challenges: (1) the difficulty of acquiring ground-truth (GT) data from labeled measurements and (2) the underutilization of measurement data. To address these issues, a novel self-supervised deep learning framework is proposed to solve the inverse problem in ICS. Building upon traditional self-supervised learning, our proposed method leverages measurement values through a multi-branch, multi-stage progressive cross-contrast structure, enabling effective learning of the common prior of the original image in the absence of GT data. Based on this framework, a novel Multi-branch Multi-stage Cross-contrast CS (MMC-CS) end-to-end DNN is designed, which unfolds the Proximal Gradient Descent (PGD) algorithm. Our approach further implements multi-scale co-optimization in the image reconstruction process, integrating both the image paths (IPs) and convolutional feature paths (CFPs). Additionally, the reconstruction performance across multiple spatial levels is enhanced by incorporating wavelet convolution (WTConv). Extensive experiments demonstrate that our proposed method outperforms existing self-supervised approaches, achieving an average PSNR improvement of 0.3-1.6 dB. Moreover, it shows strong potential to compete with state-of-the-art supervised methods.
4 May 2026
Read appraisal →Annals of biomedical engineering
Stacking Ensemble Machine Learning for Cardiac Safety Assessment Using hiPSC-CM MEA Data
PURPOSE: Prediction of Torsades de Pointes (TdP) risk using hiPSC-CM assays remains challenging, as many models fail to capture nonlinear patterns and exhibit unstable performance across different drugs. We examined whether a stacking ensemble can improve robustness when only two simple Δ F P D c -derived predictors are available. METHODS: Two electrophysiological predictors were derived from MEA-based FPDc measurements: the maximum change ( Δ F P D c ) and the interpolated Δ F P D c at Cmax. These features were used to train a stacking model comprising random forest (RF), XGB, and a shallow artificial neural network (ANN). Performance was assessed on 16 unseen CiPA reference compounds using AUC, likelihood ratios, pairwise accuracy, and classification error. Class weighting addressed class imbalance, and likelihood-based metrics assessed diagnostic consistency. RESULTS: The stacking ensemble consistently outperformed all single classifiers. The XGB-based meta-classifier achieved perfect discrimination for both AUC1 and AUC2 (1.000), with pairwise accuracy reaching 1.000 and classification error remaining below 0.125 across repeated evaluations. CONCLUSIONS: Combining a simple MEA-derived feature set with a stacking architecture provides a more reliable framework for early TdP risk assessment than single-model approaches. External validation using additional MEA datasets and the integration of interpretable modeling strategies will be important for future translational use.
4 May 2026
Read appraisal →ACS applied materials & interfaces
Fingerprint-Inspired Recyclable Electronic Skin for Non-Contact Urinary Incontinence Monitoring and Sensing
The development of high-performance flexible electronics faces a fundamental challenge: balancing sophisticated multimodal sensing and end-of-life environmental sustainability. Drawing inspiration from the human fingerprint, we present a microelectronic-printing-based arch-shaped sensor that resolves this conflict by integrating dual-mode (contact/non-contact) sensing with inherently recyclable components. The device, fabricated through high-resolution printing of liquid metal circuits (with line widths below 200 μm) on functionally tailored substrates, exhibits a sensitivity of 3.07 kPa-1 and a response time under 60 ms. It operates across a broad pressure range (0-15 kPa) in contact mode and senses proximity up to 5 cm. This robust performance enables an integrated wireless system for encrypted communication, gesture-trajectory tracking, and dynamic gesture recognition with 92.67% classification accuracy. Notably, we demonstrate a closed-loop lifecycle for the key functional materials, achieving 86.2% recovery and direct reuse of the liquid metal conductor. This work provides a versatile design strategy that simultaneously meets the demands for performance, functionality, and sustainability, offering a practical platform for future green and intelligent wearable technologies.
3 May 2026
Read appraisal →Journal of medical Internet research
Effects of a Soft Robotic Exoskeleton for Gait Training on Clinical Outcomes in Patients With Parkinson Disease: Randomized Controlled Pilot Study
BACKGROUND: Balance and gait disorders in Parkinson disease (PD) impair motor function and quality of life. OBJECTIVE: Evidence on soft exoskeleton robots (SERs) for PD rehabilitation is limited. This study evaluated the impact of SERs on motor dysfunction in PD. METHODS: A total of 56 people with PD (July 2023 to May 2024) were randomized to 2 groups: the control group (n=25, 44.6%) received conventional rehabilitation, and the experimental group (n=31, 55.4%) received conventional rehabilitation combined with SER training (ChiCTR2500111990). Training occurred 5 times per week for 20 minutes each session over 4 weeks. Primary outcomes included gait speed and stride length, while secondary outcomes assessed the percentage of swing phase, ankle joint range of motion, Unified Parkinson Disease Rating Scale total and motor scores, and Montreal Cognitive Assessment. Paired sample t tests (2-tailed) were used for within-group pre- and postintervention comparisons, and independent sample t tests (2-tailed) were used for between-group comparisons. Correlation analyses were conducted between gait parameters and improvements in ankle mobility. RESULTS: After 4 weeks, the experimental group showed significant improvements in gait and balance. Specifically, left stride length increased by a mean of 0.15 (SD 0.16; 95% CI 0.09-0.21) m (P<.001), right stride length by a mean of 0.15 (SD 0.15; 95% CI 0.10-0.21) m (P<.001), left ankle dorsiflexion by a mean of 2.84 (SD 1.46; 95% CI 2.32-3.36) degrees (P<.001), left swing phase percentage by a mean of 1.56% (SD 3.05%; 95% CI 0.44-2.68; P=.01), and right swing phase percentage by a mean of 1.6% (SD 2.72%; 95% CI 0.62-2.62; P=.002). The Unified Parkinson Disease Rating Scale Part III total score decreased by a mean of 2.80 (SD 3.98) points, and balance subscale scores decreased by a mean of 0.40 (0.58) points (P<.001). Montreal Cognitive Assessment scores increased by a mean of 1.23 (1.23; 95% CI 0.77-1.68) points (P<.01), and Barthel Index scores increased by a mean of 6.84 (7.14; 95% CI 4.22-9.46) points (P<.001). Other measures such as balance reaction time, reaction speed, maximum movement distance, and movement direction control showed significant improvement (P<.01). Compared to the control group, the experimental group demonstrated greater improvements in gait speed (P=.04), balance reaction time (P=.04), and maximum movement distance (P=.048). Correlation analysis revealed that improvements in left ankle dorsiflexion were positively correlated with improvements in gait speed, stride length, and swing phase duration (P<.05). CONCLUSIONS: SER-assisted training significantly improves gait, balance, and PD symptoms. Our work integrates multidimensional assessments (gait analysis, balance metrics, and clinical scales) and reveals that gains in ankle mobility directly correlate with gait improvements, suggesting a key mechanism. This study contributes by establishing SER as an effective adjunct to conventional therapy, supported by comprehensive quantitative data.
3 May 2026
Read appraisal →Biomedizinische Technik. Biomedical engineering
Blood pressure estimation using single photoplethysmography signal based on ResNet-BiGRU
OBJECTIVES: Conventional cuff-based blood pressure (BP) measurement provides only intermittent readings, whereas photoplethysmography (PPG)-based methods enable continuous and noninvasive monitoring. This study aims to develop a deep learning framework for accurate, cuffless BP estimation using a single PPG signal. METHODS: A hybrid deep neural network, termed ResNet-BiGRU, was developed by integrating residual convolutional blocks with bidirectional gated recurrent units to jointly capture morphological and temporal features. The UCI Cuff-Less Blood Pressure Estimation Dataset (a subset of MIMIC-II), which contains synchronized PPG and arterial blood pressure (ABP) signals from 942 subjects, was used for model training and validation. After applying a 0.5-8 Hz bandpass filter and segmenting into 5 s windows, the data were split 80/20 for training and validation. External evaluation was conducted using the VitalDB dataset, which provides synchronized PPG and ABP recordings from surgical patients under diverse physiological conditions. RESULTS: The model achieved mean absolute errors (MAE) of 4.78 mmHg for systolic BP (SBP) and 2.98 mmHg for diastolic BP (DBP) on UCI, and 8.15 mmHg for SBP and 4.59 mmHg for DBP on VitalDB. CONCLUSIONS: The ResNet-BiGRU model demonstrates accurate, robust, and generalizable cuffless BP estimation, showing strong potential for wearable health monitoring applications.
28 Apr 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Neural reservoir control of a bio-hybrid soft arm
A long-standing engineering problem, the control of soft robots is difficult because of their highly nonlinear, heterogeneous, anisotropic, and distributed nature. Here, bridging engineering and biology, neural reservoirs are employed for the dynamic control of a bio-hybrid model arm made of multiple muscle-tendon groups enveloping an elastic spine. We show how the use of reservoirs facilitates simultaneous control and self-modeling across challenging tasks, outperforming classic neural network approaches. Further, through the use of spiking reservoirs on neuromorphic hardware, energy efficiency gains of up to 75 and 45 times are obtained relative to standard and high-efficiency CPUs, with implications for the on-board control of untethered, small-scale systems.
28 Apr 2026
Read appraisal →Biomedizinische Technik. Biomedical engineering
Evaluation of an adaptive system for breathing control in closed-loop diaphragm pacing
OBJECTIVES: Existing diaphragm pacing (DP) system use an open-loop control method with a fixed stimulation mode to control breathing. It requires doctors to manually adjust stimulation parameters to meet the patient's ventilation needs. METHODS: A neural network adaptive controller was tested to control breathing in DP system. For the diaphragm motion model, the respiratory airflow of healthy adults and rabbits was collected and compared with the simulation calculation results to verify the accuracy of the model. The performance of the adaptive controller was evaluated comparatively with that of the PID controller. Adaptive controller consists of a neural network and adaptively adjusts stimulation parameters to produce the desired respiratory volume waveform. Superiority of the output performance of the adaptive controller was further studied by setting various diaphragm model parameters. We further verified the feasibility of the adaptive controller through animal testing. RESULTS: The adaptive controller is better than the PID controller in maintaining the stability of the desired breath volume. Application of the adaptive controller can reduce the root mean square (RMS) error between the desired breath volume and the actual value to less than 6 %. CONCLUSIONS: This study demonstrates the potential application of adaptive controllers in closed-loop DP systems.
27 Apr 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
Machine learning multiscale collective cell dynamics: From single-cell characterization to multicellular monolayer modeling
Collective cell dynamics are fundamental to numerous physiological processes, including embryo development, tissue morphogenesis, immune response, and disease progression. Accurately modeling these dynamics across scales remains challenging, as traditional physics-based models usually rely on many unmeasurable parameters and unclear active physics, limiting their ability to capture both single-cell features and emergent multicellular behaviors. Here, we introduce a scale-adaptive hybrid machine learning (ML) framework, in which complementary physics-guided and physics-agnostic method each resolves an above key issue at single-cell and multicellular scales, respectively. The physics-guided ML method integrates physical models with experimental data to infer previously unmeasurable parameters, enabling more accurate characterization of single-cell shape and velocity features, but unknown active forces and inherent physical assumptions still constrain its ability to predict multicellular scale dynamics. In contrast, the physics-agnostic ML method bypasses explicit physical assumptions to directly model multicellular behaviors from historical state sequences. This method provides robust long-term predictions of coarse-grained density oscillations and waves observed experimentally, yet it suffers from high stochasticity when applied to single-cell scale modeling. By assigning each approach to the scale where it is most effective, our framework leverages their respective strengths while mitigating their limitations. This complementary strategy bridges the gap between theoretical modeling and experimental observations, offering a versatile computational paradigm for multiscale collective cell dynamics with broad potential in diverse physiological and pathological contexts.
27 Apr 2026
Read appraisal →ACS sensors
A Temperature-Pressure Dual-Mode Flexible Patch Sensor with Temperature Compensation and Pressure Decoupling Functions for Pulse Detection and Recognition
Accurate and continuous detection of human pulse waves is critical for long-term health monitoring and physiological assessment during physical activity. However, most conventional flexible pulse sensors are single-mode, making them susceptible to interference from both ambient and physiological temperature fluctuations. Developing integrated dual-mode sensors remains challenging due to the difficulty in effectively decoupling pressure and temperature signals. Herein, we report a dual-mode flexible patch sensor featuring a unique "compass-like" structural design. The sensor demonstrates a broad pressure detection range (50 Pa-200 kPa) and rapid response/recovery times (60/96 ms), while its integrated temperature-sensing unit achieves a high resolution of 0.1 °C. Through unique structural design, the sensor achieves robust pressure decoupling of the temperature signal, with resistance fluctuations of only 0.102% under pressures up to 30 kPa. This independent temperature signal is subsequently utilized as a reference for the real-time temperature compensation of the pressure signal. The compensated pressure signal exhibits minimal temperature drift (41 × 10-6 °C-1 across 5-45 °C), enabling stable, interference-resistant monitoring of radial artery pulse waves under dynamic thermal conditions. Furthermore, when integrated with a ResNet-based deep learning model, the sensor successfully classifies diverse pulse signals with an accuracy of 97.6%. This research provides an effective solution for interference-resistant wearable electronics, offering significant potential for intelligent disease diagnosis and sports health management.
27 Apr 2026
Read appraisal →ACS sensors
Nanobiosensors and Artificial Intelligence Strategies for Glycan Profiling in Cancer Progression: A Critical Review
Glycans are complex carbohydrates that are integral to cancer progression and tumor biology. Conventional methods face challenges due to the structural diversity and low abundance of glycans, limiting the detection of cancer-associated glycan signatures. This bottleneck hampers the accurate monitoring of cancer progression and the development of effective diagnostic tools. Nanobiosensor-enabled glycan profiling overcomes conventional limitations by providing highly sensitive detection of low-abundance glycans and precise discrimination of similar forms through nanoscale electrochemical or optical signal amplification and selective biorecognition. This review evaluates emerging nanobiosensors, including optical modalities such as surface plasmon resonance, localized surface plasmon resonance, surface-enhanced Raman scattering, fluorescence, and colorimetric nanobiosensors. It also covers electrochemical systems, including impedimetric, voltammetric, and amperometric sensors that use nanomaterials to enhance the sensitivity, specificity, and reliability of glycan detection. After a critical evaluation of nanobiosensors, this review highlights their applications for glycan analysis in cancer progression and potential of miniaturized point-of-care platforms. In parallel, this review highlights emerging artificial intelligence (AI) and machine learning (ML) approaches that are being explored to analyse complex biosensor outputs. By enabling multivariate pattern recognition, dimensionality reduction, and nonlinear modeling of biosensor responses, these computational frameworks may facilitate the extraction of diagnostically relevant signatures from high-dimensional optical and electrochemical datasets. The review identifies challenges in standardization and clinical scalability of nanobiosensors and explores future directions, including AI integration to advance early cancer detection and personalized patient management.
27 Apr 2026
Read appraisal →ACS sensors
Multifunctional Wearable OLED-OPD System with Wavelength-Tunable Emission for Phototheranostics
The growing demand for noninvasive wearable healthcare technologies calls for compact platforms capable of integrating diagnostic sensing and therapeutic light delivery within a single device. Conventional systems typically rely on spatially separated light sources or rigid inorganic LEDs, leading to inefficient area usage and limited mechanical compatibility with soft biological tissues. Here, we present a multifunctional wearable optoelectronic platform based on wavelength-tunable organic light-emitting diodes (WTOLEDs) integrated with organic photodiodes (OPDs). The vertically stacked red and blue OLED structure enables polarity-dependent wavelength switching under AC driving, generating alternating emissions at 624 nm and 464 nm from a single compact device without lateral source separation. The surrounding geometry, in which the light-emitting region encircles the light-receiving element, provides an area-efficient optical interface that enhances optical coupling and signal sensitivity within a flexible form factor. Using this platform, we demonstrate dual-wavelength physiological monitoring via photoplethysmography (PPG) and therapeutic applications, including red-light photobiomodulation and blue-light-mediated bilirubin photodegradation. The device further exhibited stable thermal and mechanical performance under wearable operating conditions. These results establish a compact phototheranostic architecture enabling wavelength-selective sensing and therapy within a single wearable platform.
26 Apr 2026
Read appraisal →ACS sensors
Wearable Flexible Acoustic Alarm Device Based on MXene/Laser-Induced Graphene@Polyimide Thermoacoustic Film
Flexible acoustic devices have advantages of high flexibility and large sound output area, capable of adhering to the surface of the skin or clothing, which shows great potential in the field of wearable electronics. However, the design of acoustic material structures and their functional applications are still difficult. Here, we fabricated a porous laser-induced graphene (LIG) microarray layer on a polyimide (PI) film by laser direct writing. Subsequently, we uniformly coated the few-layer MXene nanosheets onto the LIG surface to obtain an MXene/LIG@PI film featuring low specific heat capacity per unit area, large interlayer spacing, excellent electrical conductivity, and high thermal conductivity. Meanwhile, the MXene/LIG@PI-based acoustic device composed of flexible copper electrodes exhibits excellent acoustic performance with a high sound pressure level of 77.6 dB and maintained a stable acoustic spectrum as the frequency increased from 20.0 Hz to 20.0 kHz. Finally, the portable acoustic alarm device composed of the piezoelectric module and the MXene/LIG@PI-based acoustic module was designed, which can be worn on various parts of the human body clothing, to provide audible alarm reminders in case of accidental falls or drops during outdoor activities. This study provides a novel approach for designing new flexible acoustic materials and devices.
26 Apr 2026
Read appraisal →ACS sensors
Wearable Multimodal Detection System for Real-Time In Situ Analysis of Exhaled Breath Condensate
Exhaled breath condensate (EBC) analysis, a promising noninvasive respiratory monitoring method, has emerged as a pivotal technique for assessing the health status of patients with respiratory disorders and is widely used in clinical research and daily health management. However, conventional analytical methods face challenges in real-time in situ detection of physicochemical indicators and active EBC collection in a power-free way. Herein, a wearable multimodal detection system (WMDS) with efficient collection and real-time analysis of EBC is developed. Specifically, the WMDS consists of a bio-inspired collector, an electrochemical sensor (EBC analysis), a respiratory sensor (humidity and respiratory rate), a temperature sensor, and a flexible printed circuit board. The miniature-sized collector with a cactus spine-like structure can actively harvest 4.1 μL of EBC within 1 min without power consumption. Leveraging self-developed sensors and wireless data transmission circuitry, the WMDS enables real-time in situ monitoring of multimodal EBC analytes (hydrogen peroxide, nitrite, urea) and respiratory parameters (temperature, humidity, and rate). Remarkably, the WMDS exhibits dual-range detection capability covering both physiological and pathological conditions: the low-concentration range of 0-500 μmol/L is applicable for routine health monitoring and early disease screening, with detection limits (LODs) of 0.209, 0.155, and 0.573 μmol/L and sensitivities of 2.7 × 10-2, 3.4 × 10-2, and 9.0 × 10-3 μA/(μmol/L) for EBC analytes; the high-concentration range exceeding 500 μmol/L is designed for severe pathological condition detection, where LODs and sensitivities are 302.2, 278.7, 325.2 μmol/L and 1.9 × 10-2, 1.3 × 10-2, 3.9 × 10-3nA/(μmol/L) respectively. As a proof-of-concept, the WMDS is applied to on-body respiratory monitoring, validating its potential application in real-time in situ health monitoring.
26 Apr 2026
Read appraisal →ACS sensors
Self-Correction of pH-Induced Signal Variations in Methylene Blue-Labeled Aptamer Electrochemical Biosensors: Wearable Cortisol Detection in Sweat
Common electrochemical aptamer-based (E-AB) biosensors employing methylene blue (MB) redox reporters suffer from significant pH-induced signal variations. Consequently, dynamic pH fluctuations in biofluids, such as in sweat, can greatly distort the measured signal and lead to inaccurate target quantification. Here, we introduce a self-correction strategy that enables accurate analyte quantification by compensating for pH-induced signal variations in MB-based aptasensors using the Nernstian shift of the MB peak potential. The highly reproducible reversible pH dependence of the MB peak potential enables real-time monitoring of the sample pH that provides a continuously synchronized built-in compensation of the pH interference, thereby eliminating the need for external pH sensors or additional measurement steps. This precise yet simple and effective self-correction of pH effects ensures that the Square Wave Voltammetry (SWV) signal reliably reflects real-time analyte variations, as demonstrated using a cortisol MB-labeled E-AB biosensor in both in vitro and on-body settings. In vitro measurements in artificial sweat across pH 5.5-7.5 showed excellent correlation with real-time pH and cortisol changes, confirming the reliability of the peak potential-based correction. On-body measurements using an epidermal wearable patch showed cortisol changes in sweat that would have been missed without this signal-correction method. This approach is broadly applicable to other MB-labeled E-AB biosensors and biofluids, providing a robust strategy for continuous on-body monitoring.
26 Apr 2026
Read appraisal →ACS sensors
Vertically Oriented Assembly of a Few-Layer Ti3C2Tx Grid by Electric Field Inducing and Freeze Drying for Wearable Levodopa Sensing in Sweat
Compared with the parallel assembly, the vertical assembly of MXene nanosheets on the substrate exhibits open three-dimensional porous structure, which is beneficial to the construction of high-performance electrodes. Herein, the vertically oriented assembly of Ti3C2Tx nanosheets is proposed by a novel preparation process of electric field inducing and freeze drying, resulting in the Ti3C2Tx grid with thin-wall and open porous structure. First, the electric field induces the formation of Ti3C2Tx hydrogel layer on the substrate. Second, freeze drying causes the growth of ice crystals for squeezing few-layer Ti3C2Tx nanosheets to stand up. More importantly, such a Ti3C2Tx grid with large surface area and intrinsic conductivity can effectively load electrocatalysts, resulting in the Au-Ti3C2Tx grid as a high-performance working electrode for the electrochemical detection of levodopa. The main parameters include 1-30 μM and 30-160 μM two-stage linear ranges and 37 nM detection limit. The structural advantages of Ti3C2Tx grid-based electrode materials are their three-dimensional conductive network and interconnected porous structure, which contribute to the penetration of the electrolyte, the diffusion of the analyte, and the electrocatalytic reaction. Furthermore, the self-adapting shape of the Ti3C2Tx grid leads to the fan-shaped Au-Ti3C2Tx grid collaborated with a self-designed leaf-like microfluidic. Ultimately, a wearable levodopa sensor composed of Au-Ti3C2Tx grid, microfluidic, flexible electronic system, and mobile phone client is established for detecting levodopa in human sweat, which is expected to help Parkinson's patients control the drug dosage. Our work proposes the vertically assembled Ti3C2Tx nanosheet loaded noble metal electrocatalysts to develop high-performance electrodes for electrochemical sensing.
26 Apr 2026
Read appraisal →ACS sensors
All-in-One Wearable Hydrogel Patch for Multimodal Visualized and Quantitative Heatstroke Monitoring
Heat-related illnesses, particularly heatstroke, pose serious risks to outdoor workers, athletes, and individuals with impaired thermoregulation. Existing monitoring strategies are limited by invasive measurements, reliance on external power sources, and lack of real-time multimodal feedback. Here, this study presents an all-in-one wearable patch that integrates a dual-functional self-powered thermogalvanic hydrogel with a selective electrochemical interface for intuitive visual feedback, dynamic temperature tracking, and durable sweat analysis. The hydrogel enables real-time thermochromic visualization and active thermoelectric sensing of body temperature, while an integrated electrochemical module allows continuous and selective detection of Na+ and K+ ions in sweat. The multiplexed physical-chemical health-monitoring patch provides a multimodal dual-stage evaluation of body temperature and sweat ion level toward improving metabolic status control and alerting individual to heatstroke risk, holding promise for developing personalized healthcare and a scalable route to proactive heatstroke prevention.
26 Apr 2026
Read appraisal →ACS sensors
Laser-Engineered MXene Heterostructure for Wearable Ammonia Sensors
Exhaled ammonia is a vital biomarker for clinical diagnosis of hepatorenal dysfunction and metabolic disorders. However, conventional sensors are often limited by complex fabrication processes, high operating temperatures, and particularly, compromised selectivity under high-humidity conditions. To overcome these challenges, we developed a laser-assisted heterostructure engineering strategy that enables in situ construction of Schottky barrier in MXene/polyacrylonitrile (PANF) nanofiber membranes, termed as a laser-engineered MXene (LEM) sensor. The laser micropatterning technique provides three key innovations: (1) kinetically controlled deposition of MXene (Ti3C2Tx) with minimal oxidation, (2) self-aligned formation of Schottky barriers through interfacial TiO2/MXene heterojunctions, and (3) creation of hierarchical gas transport channels within the 3D nanofibrous matrix. The optimized LEM sensor exhibits outstanding sensitivity (2.5% ppm-1), a low detection limit (0.2 ppm) at room temperature (25 °C), and humidity operation stability (> 10% response at 90% relative humidity). This performance surpasses that of conventional MXene-based sensors by fivefold in sensitivity without the need for thermal activation. Mechanism studies reveal that the laser-induced heterointerface facilitates charge transfer-dominated ammonia adsorption and desorption kinetics, while the nanofibrous architecture ensures molecular accessibility to active sites. This ambient-processable and scalable fabrication method paves the way for advanced wearable breath analyzers by combining manufacturing versatility, humidity-resistant excellent selectivity, and energy-efficient operation.
26 Apr 2026
Read appraisal →Lab on a chip
MaGIC-OT: an AI-guided optical tweezers platform for autonomous single-cell isolation in microfluidic devices
Automating the isolation of rare cells such as circulating tumour cells (CTCs) within crowded microfluidic environments remains a bottleneck in liquid biopsy workflows. Optical tweezers offer contact-free, selective manipulation but traditionally rely on expert operators. We present MaGIC-OT (machine-guided isolation of cells using optical tweezers), a platform that integrates classical path planning and deep reinforcement learning (DRL) to automate single-cell manipulation inside a microfluidic chip. We built a high-fidelity simulation to train and benchmark control policies and show that cooperative, human-in-the-loop training improves DRL performance. Trained agents outperform expert users in speed and isolation success in silico, and we demonstrate proof-of-concept isolation of a cancer cell from a spiked blood sample on-chip. MaGIC-OT provides a flexible framework for intelligent optical manipulation, aligning microfluidic device design with autonomous control strategies and offering a pathway toward high-purity, label-free single-cell workflows.
26 Apr 2026
Read appraisal →PloS one
RMETNet: A cross-subject motor imagery EEG signal classification model based on TSLANet and riemannian geometry features
Motor imagery electroencephalogram (MI-EEG) analysis is essential for natural interaction and autonomous control in brain-computer interfaces (BCIs). However, deep learning models often struggle with inter-subject variability, which limits their ability to generalize across subjects. This study proposes RMETNet, a novel framework that integrates TSLANet, a spatio-temporal convolution module, and a multi-scale Riemannian geometry feature module. TSLANet suppresses noise and captures complex temporal patterns for preliminary signal decoding, while the spatio-temporal convolution module extracts higher-order representations. The Riemannian branch learns geometry-based distribution features across subjects, and the fused features are used for classification. To address inter-subject distribution shifts, RMETNet incorporates Maximum Mean Discrepancy (MMD) loss for domain adaptation, aligning feature distributions between source and target domains. Experiments show that on the four-class BCI Competition IV 2a (BCICIV2a) dataset, RMETNet achieved accuracies of 71.39% in the cross-subject setting and 80.71% in the subject-dependent setting; on the two-class BCI Competition IV 2b (BCICIV2b) dataset, it achieved 80.93% and 86.76%, respectively. The model consistently outperformed baseline algorithms. Ablation and visualization analyses further validated its effectiveness in reducing inter-subject feature distribution disparities and enhancing MI-EEG decoding. The code is available at: https://github.com/rokanfeermecer486/RMETNet.
26 Apr 2026
Read appraisal →Biomedical physics & engineering express
A graph deep learning method for diagnosis of Parkinson's disease using brain functional connectivity features
Early and precise identification of Parkinson's disease (PD) is crucial for clinical intervention. Resting-state functional magnetic resonance imaging (rs-fMRI) provides a valuable approach for revealing PD-related differences in brain functional connectivity (FC). However, existing methods often focus solely on characterizing the spatial topology of FC while neglecting its time-varying dynamic fluctuations. Furthermore, they frequently exhibit limited generalization capability when dealing with small sample sizes, and their decision-making mechanisms lack interpretability. To address these limitations, this study proposes an interpretable graph convolutional network framework. This framework integrates both static and dynamic FC information to capture both the stable topological structure and the dynamic temporal characteristics of brain networks. Simultaneously, it models population relationships by constructing an inter-subject similarity graph to enhance the model's representational capacity. Additionally, this study incorporates interpretability analysis techniques to deeply dissect the model's decision-making mechanism and identify key brain regions critical for classification. Results demonstrate that the proposed model achieves superior performance in PD classification tasks and exhibits good generalization ability. More importantly, by interpreting the model's decisions, key brain regions associated with PD discrimination were successfully identified. This study provides an effective computational framework for PD identification and offers new insights into understanding its pathological mechanisms.
25 Apr 2026
Read appraisal →Science advances
Wearable thumb sleeves enabled by self-supervised learning with few stretchable sensors and few-shot data for switchable finger tasks
Human fingers exhibiting remarkable dexterity are ideal for natural human-machine interaction. Traditional methods require at least one device per finger and extensive labeled data, often limiting models to a single user and task. Here, we propose a wearable thumb sleeve integrated with self-supervised learning, which exhibits user independence and data efficiency, enabling recognition of various finger-related tasks. The thumb sleeve is equipped with only two stretchable sensors at the thumb joints and learns latent features from unlabeled random thumb movement data. By using fine-tuning with five-shot labeled data, it can rapidly adapt to new users and tasks, including eight directional commands and 10 knuckle key inputs. It allows free switching between tasks without the need to reconstruct or retrain the model. The proposed approach demonstrates strong potential for real-world applications, serving as a substitute for a mouse and keyboard to enable tasks such as online shopping.
25 Apr 2026
Read appraisal →Lab on a chip
Machine learning-driven single-cell phenotyping in size-controlled microenvironments via parallel deterministic droplet microfluidics
Understanding how individual cells respond to distinct physical microenvironments is critical for mechanobiology, cell therapy, and tissue engineering. Current single-cell encapsulation methods are often limited by Poisson loading and fixed droplet sizes, preventing parallel generation of multiple, size-specific microenvironments and constraining high-resolution phenotypic analyses. Here, we present a droplet microfluidic platform that enables deterministic single-cell encapsulation within microgels of multiple sizes from a single precursor stream, achieved through parallelized flow-focusing combined with cell-selective gelation. This system produces distinct microgel size regimes simultaneously, minimizing empty compartments and enabling direct, side-by-side comparisons of cellular behavior under controlled yet variable confinement. Using machine learning to analyze 3D morphological and cytoskeletal features, we reveal heterogeneous, size-dependent phenotypic responses and demonstrate that cellular phenotypes alone can predict microgel confinement across time. Together, these results establish a data-driven framework for mapping single-cell responses across engineered microenvironments and provide a scalable platform for predictive studies of mechanosensitive behavior in heterogeneous niches.
24 Apr 2026
Read appraisal →The Analyst
Recent progress in wearable electrochemical sensors based on MXene-conductive hydrogels
MXene-based conductive hydrogels (MCHs) have emerged as highly promising materials for next-generation wearable electrochemical sensors, owing to their exceptional electrical conductivity, mechanical flexibility, and biocompatibility. This review provides a comprehensive and up-to-date overview of recent advances in the development of wearable sensors that incorporate Ti3C2Tx MXene hydrogels for the detection of a wide range of analytes in various biofluids. Diverse hydrogel matrices have been integrated with MXenes to fabricate highly sensitive platforms capable of monitoring key biomarkers such as glucose, dopamine, uric acid, lactate, norepinephrine, sodium, creatinine, and pH. These sensors have been successfully deployed at various locations in the body, including the forearm, chest, wrist, and head, using flexible formats such as skin patches, microfluidic devices, pantyliners, wearable caps, and attachable body accessories. Notably, several configurations demonstrate ultralow detection limits, reaching the nanomolar level, enabling real-time, noninvasive analysis of sweat, urine, and other physiological fluids. The wide range of functional additives and customizable design approaches underscores the modularity and tunability of MCH-based systems for specific applications. This review critically evaluates the design principles, sensing mechanisms, performance metrics, and practical limitations of current MCH-based wearable platforms, providing insights that can guide future innovations in smart wearable healthcare technologies.
23 Apr 2026
Read appraisal →Journal of neural engineering
Physiologically inspired modeling of cortical dynamics through spiking neural networks
Objective. The characterization of neural activity underlying neurophysiological function presents a major challenge in computational neuroscience. Several methods have been proposed to investigate cortical network dynamics by reconstructing underlying neural activity from electroencephalography (EEG) signals. However, these methods generally pose significant mathematical challenges.Approach. This study introduces a novel framework to model the underlying brain activity network from a functional and physiologically-inspired perspective, combining spiking neural networks with EEG signal analysis. The dynamics of single neurons are described by the well-known Izhikevich model, and distinct populations of cortical inhibitory and excitatory neurons are employed to model experimental EEG recordings. Functional interactions among distinct populations are mathematically formalized through connective probabilities.Main results. The proposed framework is validated by testing it on synthetic data, as well as on two experimental datasets comprising data from 30 healthy subjects undergoing a cold-pressure test (CPT), and 36 subjects undergoing a mental arithmetic stressor. Experimental results suggest that the proposed framework provides novel and complementary insights into characterizing neuronal changes in comparison to standard EEG power analysis.Significance. The proposed framework constitutes a promising tool for functionally characterizing the underlying cortical dynamics under pathophysiological conditions.
20 Apr 2026
Read appraisal →PLoS computational biology
'Backpropagation and the brain' realized in cortical error neuron microcircuits
Neural responses to mismatches between expected and actual stimuli have been widely reported across different species. How does the brain use such error signals for learning? While global error signals can be useful, their ability to learn complex computation at the scale observed in the brain is lacking. In comparison, more local, neuron-specific error signals enable superior performance, but their computation and propagation remain unclear. Motivated by the breakthrough of deep learning, this has inspired the 'backpropagation and the brain' hypothesis, i.e., that the brain implements a form of the error backpropagation algorithm. In this work, we introduce a biologically motivated, multi-area cortical microcircuit model, implementing error backpropagation under consideration of recent physiological evidence. We model populations of cortical pyramidal cells acting as representation and error neurons, with bio-plausible local and inter-area connectivity, guided by experimental observations of connectivity of the primate visual cortex. In our model, all information transfer is biologically motivated, inference and learning occur without phases, and network dynamics demonstrably approximate those of error backpropagation. We show the capabilities of our model on a wide range of benchmarks, and compare to other models, such as dendritic hierarchical predictive coding. In particular, our model addresses shortcomings of other theories in terms of scalability to many cortical areas. Finally, we make concrete predictions, which differentiate it from other theories, and which can be tested experimentally.
19 Apr 2026
Read appraisal →PloS one
Statistically valid explainable black-box machine learning: applications in sex classification across species using brain imaging
Sex classification using neuroimaging data has the potential to revolutionize personalized diagnostics by revealing subtle structural brain differences that underlie sex-specific disease risks. Despite the promise of machine learning, traditional methods often fall short in providing both high classification accuracy and interpretable, statistically validated feature importance scores for high-dimensional imaging data. This gap is particularly evident when conventional techniques such as random forests, LIME, and SHAP are applied, as they struggle with complex feature interactions and managing noise in large datasets. We address this challenge by developing an integrated framework that combines Oblique Random Forests (ORFs) with a novel, permutation-based feature importance testing algorithm. ORFs extend traditional random forests by employing oblique decision boundaries through linear combinations of features, thereby capturing intricate interactions inherent in neuroimaging data. Our feature importance testing method, NEOFIT, rigorously quantifies the significance of each feature by generating null distributions and corrected p-values. We first validate our approach using simulated datasets, establishing its robustness and scalability under controlled conditions. We then apply our method to classify sex from both voxel-wise structural MRI and cortical thickness data in humans and macaques, facilitating direct cross-species comparisons. ORFs achieves AUC > 0.80 on human data, and >0.70 on macaque data, while NEOFIT identifies statistically significant features aligned with sex-dimorphic neuroanatomy. Our results demonstrate that the proposed framework not only enhances classification performance but also provides clear, interpretable insights into the neuroanatomical features that distinguish sexes. These methodological advancements pave the way for improved diagnostic tools and contribute to a deeper understanding of the evolutionary basis of sex differences in brain structure.
18 Apr 2026
Read appraisal →Carbohydrate polymers
Multifunctional conductive hydrogel based on carboxymethyl cellulose/oxidized sodium alginate for machine learning-guided sports training
Multifunctional hydrogels have shown great promise in advanced wearable sensing technologies; however, their performance under real-world sweaty and dynamic conditions remains a significant challenge. In this study, a highly conductive and adhesive hydrogel was developed using poly(AM-co-AA), carboxymethyl cellulose (CMC), oxidized sodium alginate (OSA), and polydopamine-modified carbon nanotubes (PDA-CNT). Featuring a dual dynamic cross-linking system based on reversible Schiff base bonds and strong hydrogen bonding, the hydrogel demonstrates high mechanical toughness (2416 kJ/m3), excellent stretchability (1829%), and stable electrical conductivity (1.5 S/m). It exhibits outstanding sweat resistance, maintaining strong adhesion and reliable sensing performance even during intense physical activities, along with high fatigue resistance over 10,000 deformation cycles, enabling long-term reliable monitoring. As a highly sensitive strain sensor, it offers multi-range gauge factors (up to 2.53) and accurately captures human motion and physiological signals without interference from sweat or motion artifacts. When integrated with machine learning algorithms, the sensor achieves a validation accuracy of 98.75%, providing real-time feedback on movement techniques and supporting personalized training optimization. Furthermore, the hydrogel demonstrates significant antibacterial properties and good biocompatibility, making it an ideal candidate for next-generation wearable electronic devices suited for long-term athletic training and high-sweat environments.
17 Apr 2026
Read appraisal →Nanoscale
Bifunctional PVDF/MXene membranes with high piezoelectric sensitivity and near infrared photothermal antibacterial efficacy toward wearable human activity monitoring
In the realm of wearable health monitoring, establishing a reliable interface that ensures both high-fidelity sensing and long-term biological safety remains a critical challenge, particularly due to the risk of bacterial infection caused by sweat accumulation at the device-skin contact. Herein, we address this issue by developing a scenario-driven, bifunctional PVDF/MXene nanofibrous membrane that integrates robust piezoelectric performance with on-demand photothermal antibacterial capability. In this design, MXene nanosheets function simultaneously as nucleating agents to boost the electroactive β-phase content of PVDF from 80.3% to 93.3% and as efficient near-infrared (NIR) photothermal agents. The optimized membrane exhibits a superior piezoelectric sensitivity of 0.502 V/N, enabling the precise detection of diverse human activities ranging from subtle pulse waveforms to vigorous jumping motions, and motion-triggered Morse code communication. Crucially, addressing the hygiene concerns of reusable wearables, the membrane achieves rapid, non-contact sterilization during off-wear intervals. Under safe NIR irradiation (0.3 W cm-2), the surface temperature rises to ∼60 °C within 60 s, achieving an antibacterial efficacy of >99% against S. aureus and E. coli. Combined with excellent in vitro cytocompatibility, this work provides a "one-structure, dual-function" solution for constructing hygienic, high-performance electronic skins.
17 Apr 2026
Read appraisal →Physics in medicine and biology
A surface-conforming flexible collimator for improving electron radiotherapy.
Objective.We aim to simplify the traditional method of patient-specific skin collimation and to improve the precision of electron radiotherapy treatment with a new surface-conforming collimation system.Method.A surface-conforming collimation device for electron radiotherapy, consisting of parallel steel cables, was created. The cables move freely relative to each other and can be fixed and reshaped with a clamp system along the length of the device. The cables do not completely fill the space; thus, their packing fraction was determined to characterize dosimetric effects. The cable collimator was tested with a 10 × 10 cm cone with four electron energies (6, 9, 12, and 15 MeV) at three SSDs (100 cm, 105 cm, and 110 cm). Radiochromic film was placed at the surface and at the maximum dose depth (dmax) for each energy. Penumbral width, defined as the distance between 20% and 80% of the maximum dose, was measured for each energy and depth.Results.The largest improvement with respect to traditional collimation was seen for 6 MeV with 110 cm SSD at the surface. The penumbra at the surface for the conventional collimator in this case was 18.7 mm, while the cable collimator penumbra was 1.4 mm. The smallest improvement was seen for 15 MeV with 100 cm SSD atdmax(2.7 cm), where the penumbra for the traditional collimator was 11.1 mm, while the cable collimator penumbra was 9.8 mm. In all cases, the cable collimator reduced penumbral width. Surface dose inhomogeneities due to scatter were most pronounced at higher energies. The packing fraction was measured to be 0.551 ± 0.016.Significance.The prototype shows significant improvement in penumbra compared to conventional collimation. The system aims to serve as a replacement for current skin collimation techniques and could be used with Cerrobend applicators for additional surface blocking.
17 Apr 2026
Read appraisal →Theory in biosciences = Theorie in den Biowissenschaften
Analytical and AINN-based investigation of surface wave propagation in dry long bones with initial stress, magnetic field, and rotation effects
This study presents a comprehensive analytical and artificial intelligence neural network (AINN)-based investigation of wave propagation in dry long bones, modeled as an orthotropic hollow cylindrical structure. The proposed model incorporates the combined effects of initial stress, magnetic field, and rotational motion to capture the complex behavior of bone-like media under coupled physical influences. The governing equations are formulated within a continuum mechanics framework and solved analytically using displacement potential methods, yielding solutions in terms of Bessel functions that satisfy the cylindrical geometry and boundary conditions. Two distinct cases, corresponding to the absence and presence of rotation, are examined to assess the influence of rotational effects on wave dynamics. A detailed parametric analysis is carried out to evaluate the variation of phase velocity and frequency with respect to wave number, initial stress, magnetic parameter, density, and geometric ratios. The analytical results indicate that initial stress enhances wave propagation, while magnetic effects introduce damping and rotation significantly modifies dispersion characteristics. To improve computational efficiency and predictive capability, an AINN model is developed and trained using analytically generated data. The AINN predictions show excellent agreement with the analytical results, as confirmed through parity plots, error analysis, residual distribution, and loss convergence behavior. The novelty of the study lies in the unified analytical-AINN framework that integrates mechanical, electromagnetic, and rotational effects within a single model. The findings provide important insights for wave-based characterization of bone structures and have potential applications in non-destructive evaluation, ultrasonic diagnostics, and advanced biomedical sensing systems.
17 Apr 2026
Read appraisal →Computers in biology and medicine
Empirical evaluation of variability and multi-institutional generalizability of deep learning survival models: application to renal cancer CT scans
Deep learning survival (DLS) modeling has shown significant promise across multiple disease conditions for predicting patient outcomes via biomedical data, but remain relatively under-explored in conjunction with medical imaging. This poses the question of what methodological choices promote model generalization, especially when predicting continuous survival outcomes in external validation cohorts. In this work, we systematically examined the impact of data partitioning, data order, model initialization, and augmentation strategies on the generalization of CT-based DLS survival models in renal cancers. A multi-institutional CT cohort was assembled, leveraging public repositories such as TCGA-KIRC, TCGA-KICH, TCGA-KIRP, and KITS19; totaling 525 patients from across 9 institutions. A 3D ResNet-18 model was trained to predict overall patient survival with a Cox loss function and evaluated across three controlled experiments: (i) random versus covariate-balanced data partitioning, (ii) randomizing vs fixing data ordering and initial model weights, and (iii) varying intensity, spatial, and noise based augmentations; evaluated in terms of concordance index (c-index) and hazard ratio (HR). Intelligent covariate-balanced data partitioning demonstrated optimal performance on an external validation set (c-index 0.74, HR 5.08) compared to random partitioning (c-index 0.56, HR 0.29). Different model initializations significantly impacted the average performance and variance across replicate trainings, while data ordering had little to no effect. Combining multiple augmentations produced the best performance on external validation (c-index +4.76%, HR +44.39%) compared to no augmentations, while most impactful individual augmentation type was additive gaussian noise (c-index +3.17%, HR +23.47%). Our results demonstrate that careful selection of strategies for data partitioning and ordering, weight initialization, and image augmentation are critical for developing robust and generalizable DL models for continuous survival prediction using medical imaging data.
17 Apr 2026
Read appraisal →Science advances
Enhancing biomedical optical volumetric imaging via self-supervised orthogonal learning
Optical volumetric imaging grapples with inherent noise problems arising from photon budget constraints, light scattering, and space-bandwidth product bottlenecks, all of which degrade structural fidelity and become even more pronounced than in planar imaging. While deep learning presents potential for denoising, supervised methods are hindered by the impracticality of paired datasets, and existing self-supervised approaches fail to fully exploit the intrinsic volumetric structural redundancy inherent to optical imaging. Here, we present a self-supervised volumetric biomedical imaging denoiser (VALID) that leverages intrinsic three-dimensional spatial coherence for highly efficient volumetric denoising through a self-supervised orthogonal learning framework. VALID demonstrates robust denoising performance across diverse imaging modalities, including two- and three-photon microscopy, light-field microscopy, and optical coherence tomography, substantially enhancing structural fidelity in deep-tissue, multimodal, and dynamic imaging scenarios. By combining computational efficiency with zero-shot adaptability, VALID establishes a transformative approach to volumetric image enhancement with structure-aware precision.
16 Apr 2026
Read appraisal →Science advances
Drawn-on-skin electronic tattoo as a closed-loop sensing-stimulation system for the muscles
Simultaneous electromyography (EMG) sensing and closed-loop electrical stimulation (ES) could enable interactive muscle training, motor rehabilitation, and kinesthetic feedback. However, current systems often suffer from nonconformal device/skin interfaces, poor wearability, and anatomical variability among users, limiting signal fidelity and stimulation precision. Here, we present a fully wireless, skin-integrated electronic tattoo platform that records high-quality EMG and delivers closed-loop ES through custom drawn-on-skin, conformal electrodes with strong adhesion and high spatial accuracy. The system supports closed-loop muscle interactions across body regions and even between individuals. Machine learning models classify EMG patterns from different hand gestures with >90% accuracy and adapt ES parameters to stimulate specific muscle groups. In a heavy object holding task, users reached the required grip behavior substantially faster than unassisted controls, indicating improved neuromuscular efficiency. This closed-loop framework not only supports personalized muscle control but also supports coordinated activation across multiple sites or users, unlocking possibilities in interactive motor training, remote rehabilitation, and virtual reality environments.
16 Apr 2026
Read appraisal →Physics in medicine and biology
A deep learning-based framework for patient-specific radiation dose prediction in beta-emitting radionuclide therapies
Objective.Conventional image-based models for radionuclide therapy dosimetry are typically radionuclide-specific and rely on nuclear medicine (NM) images for training. We developed a deep learning (DL) model that predicts doses for not only training radionuclides but also radionuclides not included in the training data without patient NM images.Approach.The DL model was trained to predict voxel dose kernels (VDKs) for beta-emitters using CT-derived density maps and energy spectra. The DL model-derived VDKs were convolved with time-integrated activity to generate patient-specific dose maps. Zero-shot prediction was tested on Ho-166 and Lu-177. The performance of DL model was validated at kernel, phantom, and patient levels. The proposed DL model, conventional DL model trained on Y-90 NM/CT images of 22 patients, voxelS-value (VSV), and organ-level dosimetry (OLINDA/EXM) were benchmarked against Monte Carlo simulation (MC).Main results.Training utilized a single CT image, circumventing large patient cohort requirements. The DL model-based VDKs showed strong correlation with MC-based VDKs (R2: 0.99 and 0.91 for test Ho-166 and Lu-177) and, at the phantom level, achieved overall lower errors than VSV in heterogeneous region. At the patient level, the proposed DL model predicted organ doses with the smallest differences from MC while reducing computation time from 4.6 h to 2.6 min. For the test Lu-177, the proposed model exhibited good agreement with MC (2.1 Gy, 11.6%), while the conventional model trained using Y-90 showed large discrepancies (129.4 Gy, 657.4%). For the mean tumor dose averaged across the patient cohort for test Ho-166, the proposed DL model yielded the most accurate results with a difference of only 1.1 Gy (2.9%), outperforming multiple VSV (3.8 Gy, 7.2%), single VSV (3.8 Gy, 7.3%), and OLINDA/EXM (1.4 Gy, 4.5%).Significance.The proposed DL model enables dose prediction for radionuclides not included in the training data without NM images.
16 Apr 2026
Read appraisal →Progress in biomedical engineering (Bristol, England)
Navigation paradigms for non-invasive BCI-controlled wheelchairs: a systematic review
Brain-controlled powered wheelchairs represent a promising advancement for individuals with neurological conditions that significantly impair motor function. Despite substantial progress, brain-controlled wheelchairs have not been adapted for real-world settings. This article systematically reviews recent trends in brain-computer interface (BCI) technology for wheelchair navigation and control, highlighting the contributions and limitations of various navigation paradigms. The review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines, sourcing studies from four databases (PubMed, Scopus, IEEE Xplore, Google Scholar) published between 2000 and April 2025. It focused on non-invasive BCI paradigms and real-world navigation experiments. The results were narratively synthesized and classified into two primary categories: BCI-based navigation paradigms and wheelchair-based navigation paradigms, along with intersecting concepts such as single-variant BCI, hybrid BCI, control switches, and proportional control. Of the 149 full-text articles reviewed, 47 were included and categorized by navigation paradigm, comprising 20 BCI-based and 27 wheelchair-based studies, with 6 involving participants with motor disabilities. Quality assessment scores ranged from 40% to 95%, with approximately 40% of the studies demonstrating a low risk of bias. The findings indicate that low-level navigation control was predominant in BCI wheelchair studies, with 31 studies employing minimal or no obstacle avoidance. Most studies (57%) integrated sensors for obstacle avoidance, localization, mapping, and autonomous navigation. Twenty-two studies utilized control switches, and five incorporated proportional control for wheelchair navigation. Additionally, motor imagery and steady-state visually evoked potential paradigms have emerged as the most common approaches for generating control commands, highlighting their potential for effective navigation. Given the potential societal impact on a large number of individuals, future research should prioritize enhancing the reliability and adaptability of BCI wheelchair systems in real-world environments.
16 Apr 2026
Read appraisal →Biomedical physics & engineering express
EEG-CMT: spatial-temporal representation of EEG for emotion recognition using convolutional neural networks and vision transformers
Background. Recent researches on electroencephalogram (EEG) based emotion recognition face challenges in effectively mapping the spatial positional relationships of EEG acquisition electrodes. Additionally, conventional models struggled to simultaneously capture both fine-grained temporal-spatial features and long-range dependencies in EEG signals.New method. To address these limitations, we propose a novel EEG data processing method that incorporates spatial relative position encoding and a hybrid neural architecture integrating convolutional neural networks (CNNs) with self-attention mechanisms. This approach systematically encodes the spatial topology of electrodes to enhance the representation of temporal-spatial information. CNNs are employed to extract localized temporal-spatial micro-patterns, while self-attention modules model global contextual dependencies across extended sequences, thereby enhancing model's representational capacity.Results. The experimental results and feature visualizations demonstrate that our method achieves state-of-the-art (SOTA) subject-dependent performance on two benchmark emotion recognition datasets, reaching an average accuracy of 97.51% on the SEED dataset and 96.13% on the SEED-IV dataset. Moreover, the learned spatial features align well with known neuroscientific patterns of emotional processing. Comprehensive ablation studies further validate the necessity and effectiveness of both the spatial-encoded data processing strategy and the hybrid architecture design.Comparison with Existing Methods. Compared to other hybrid neural network models, our proposed method (EEG-CMT) achieves the highest classification accuracy. Specifically, it outperforms baseline algorithms by margins ranging from 0.86% to 11.43% on the SEED dataset, and from 9.49% to 39.52% on the SEED-IV dataset.Conclusions. The proposed method effectively addresses key limitations in existing EEG-based emotion recognition models by jointly leveraging spatial topology and hybrid modeling techniques. These innovations significantly improve the model's ability to recognize emotions from EEG data and provide neural interpretable insights, offering a promising direction for future research in affective brain-computer interfaces.
15 Apr 2026
Read appraisal →Biomacromolecules
A Phase-Separated Cellulose-Ionogel via Sunlight-Initiated Green Fabrication for Self-Powered Wearable Sensors
Ionogels, as soft ionic conductors, face synthesis challenges including toxicity, complexity, and high energy consumption. Herein, we present a green one-pot strategy that effectively dissolves cellulose and undergoes sunlight-induced photopolymerization to form ionogels without the need for cross-linkers or initiators. 1-Butyl-3-methylimidazolium chloride ([BMIM]Cl), as the solvent, enables the disruption of the extensive hydrogen-bond network of cellulose, resulting in rapid and complete dissolution. Subsequent one-step photopolymerization, which proceeds solely under sunlight, simultaneously drives in situ cross-linking and a controlled phase separation process, yielding high-performance ionogels. Importantly, the resulting cellulose ionogel exhibits superior fracture strength (2.75 MPa), high toughness (18.4 MJ m-3), and strong adhesion (6.6 MPa), ameliorating the traditional trade-off between mechanical strength and adhesion capabilities. This work develops an integrated ionogel platform as a soft TENG electrode for human motion monitoring, informing the design of sustainable self-powered electronics.
14 Apr 2026
Read appraisal →Biomacromolecules
Highly Flexible, Adhesive, Antimicrobial, and Self-Powered Hydrogel Strain Sensor for Human Motion Monitoring
Piezoelectric hydrogel sensors are a recent innovation that combines the flexibility of hydrogels with the self-powering of piezoelectric materials. Here, we present a self-powered hydrogel strain sensor by incorporating piezoelectric Na0.5Bi0.5TiO3 nanoparticles into a dual-network hydrogel composed of poly(vinyl alcohol), acrylamide, and 2-acrylamido-2-methyl-1-propanesulfonic acid. The hydrogel exhibits excellent mechanical properties, achieving a tensile strength of 1.28 MPa, an elongation at break of 1127%, and strong tissue adhesion (55.97 kPa on porcine skin). The sensor delivers a maximum output voltage of 237.4 mV, a sensitivity of 17.99 mV under small compressive strain, and a broad tensile strain detection range of 30 to 400%. It reliably monitors both subtle physiological activities (facial expression, swallowing, pulse) and large joint movements. Moreover, it exhibits notable antibacterial performance via ultrasound-driven piezocatalysis, reducing bacterial viability to below 1.5%. This tough, adhesive, and antibacterial piezoelectric hydrogel sensor holds significant promise for wearable electronics and human-machine interfaces.
14 Apr 2026
Read appraisal →PloS one
Unsupervised industrial image defect detection based on autoencoder and GANs
In the process of intelligentization in modern manufacturing, especially in industrial fields such as automobile manufacturing, semiconductor production, and electronic product assembly, product quality control is crucial. Traditional defect detection methods face problems such as supervised learning methods relying on a large amount of labeled data, weak generalization ability, and high cost. In the process of intelligent manufacturing, industrial product quality control is a key link to ensure production safety and product consistency. Especially in typical industrial scenarios such as automobile manufacturing, semiconductor production, and electronic product assembly, traditional defect detection methods are difficult to meet the needs of actual production lines for high-precision and high-efficiency detection due to their reliance on a large amount of labeled data, weak generalization ability, and high cost. To solve the problem of defect detection in small and zero sample scenarios in these industries, improve detection sensitivity and localization accuracy, and enhance the model's generalization ability to unknown defects, a unsupervised industrial image defect detection method based on autoencoder and Generative Adversarial Networks (GANs) is proposed. This study constructs an Multi-level Deep feature Adaptive fusion AutoEncoder (MDAAE) module, extracts multi-scale features through Pre-Trained Convolutional Neural Backbone Networks (PTCNBN), introduces Attention Mechanism (AM) to dynamically calculate feature weights, and achieves feature fusion and reconstruction. Meanwhile, the self-AM is fused to improve the GANs. The self-attention module efficiently captures long-range dependencies and generates an adversarial network training objective function to optimize the generator's unsaturated loss. The results showed that the Area Under the Curve (AUC) of the Recall-Precision curve of the research method reached 93.6 ± 0.5%, and its F1-Score value exceeded 0.890 ± 0.003 in defect types such as scratches and dents. In practical applications, the inference time of the research method remained stable at 3.0 ± 0.2 GB of central processor memory, and the Fréchet Inception Distance (FID) value fully converged at the 2230th iteration, with a stable FID value of 3. The false alarm rate was only 8.6 ± 0.7% under strong light conditions. The proposed UIIDD method based on autoencoder and GANs had good robustness, generalization ability, reliability, and efficiency. This study effectively solves the problems of poor robustness, weak generalization ability, and high cost of traditional defect detection methods.
12 Apr 2026
Read appraisal →Journal of speech, language, and hearing research : JSLHR
A Speech-Segregation Algorithm for Spatial Hearing Aids to Operate With Multiple Sound Sources
PURPOSE: Hearing-aid users often face challenges in noisy environments due to time, level, and spectral cues being compromised by current generation hearing aids. This study explored a physiologically based speech-segregation algorithm that selectively removes or attenuates unwanted sound sources. METHOD: In our previously developed localization algorithm, the time-frequency responses after a unique normalization approach always reside inside the unit circle of the model space. Given the "sparseness" property of daily sound, the model forms distinct clusters that correspond to the source locations. In the present study, the localization model was adapted to segregate speech by using a binary mask to remove the cluster of unwanted sound. The speech target was one of 200 intelligible sentences. The interfering sound was time-reversed sentences in a random sequence spoken by the same speaker. Automatic speech recognition transcribed the sound mixture before and after the segregation algorithm. RESULTS: When both target and noise were located at the front, applying a hard mask (i.e., 1 or 0) almost perfectly removed the energy of noise. When the sound sources moved to the side or back with smaller angular separations, clusters were less distinguishable, leading to worse intelligibility performance. Applying a soft mask (i.e., 1 or 0.2) instead showed slightly lower performance for the front but improved performance for the back and side. CONCLUSION: Our algorithm performs localization and segregation in a combined and straightforward manner, potentially for spatial hearing aids to function better in challenging listening environments.
12 Apr 2026
Read appraisal →PloS one
Topology-aware design of spiking neural networks via modular graph architectures
Spiking Neural Networks (SNNs) offer a biologically plausible and energy-efficient alternative to traditional artificial neural networks (ANNs), yet their design remains constrained by limited architectural flexibility and slow training dynamics. In this work, we introduce a novel SNN framework that leverages modular graph-based topologies and explicit synaptic delays to significantly enhance both training efficiency and classification performance. Our architecture, TANet-Tiny, incorporates structured graph stages with up to 32 nodes and diverse community-driven connectivity patterns derived from KMeans clustering, Louvain modularity, and Watts-Strogatz small-world models. We integrate these topologies into a topology-aware search space and explore them via a Spatio-Temporal Topology Sampling (STTS) approach, enabling the discovery of high-performing networks without exhaustive search. Experimental results on MNIST, CIFAR-10, and CIFAR-100 demonstrate that our modular designs achieve state-of-the-art accuracy while requiring 6-10 × fewer training epochs, with top-1 accuracy reaching 99.57% on MNIST and over 92% on CIFAR-10, all with reduced parameter counts. We introduce an accuracy-per-epoch metric to quantify training efficiency and show that modularity, rather than network size, is the critical driver of performance. This work lays the groundwork for scalable, interpretable, and low-latency SNN architectures suitable for deployment in neuromorphic and edge computing environments.
12 Apr 2026
Read appraisal →International journal of pharmaceutics
Optimization of drug diffusion in drug-eluting stents for coronary artery based on deep reinforcement learning
Drug-eluting stents (DES) are extensively used to treat coronary artery disease, and improving their therapeutic efficacy remains a long-standing research objective. Existing optimization strategies for DES drug delivery lack adaptive, learning-based frameworks capable of simultaneously enhancing drug concentration and uniformity in the therapeutic domain while regulating levels in the non-treatment domain, and show limited capacity to address complex, nonlinear optimization problems. This study applies a deep reinforcement learning (DRL)-based framework to optimize drug diffusion in DES. By integrating neural networks with the proximal policy optimization (PPO) algorithm, the framework enables closed-loop, feedback-driven regulation of drug diffusion in real time and demonstrates robustness to stochastic disturbances and parametric variations. The method was systematically evaluated across three stent embedding configurations: half-embedded, fully embedded, and non-embedded. Results show the DRL agent successfully balances the three therapeutic objectives and adopts distinct control strategies for different configurations. All configurations exhibited improved overall performance. For the half-embedded configuration, the DRL policy increased therapeutic domain concentration by 6.99%, accompanied by a slight rise in the non-treatment domain and a minor decrease in uniformity within the therapeutic domain. Under the fully embedded configuration, concentrations in both domains remained essentially unchanged, whereas uniformity in the therapeutic domain improved by 8.14%. For the non-embedded configuration, both therapeutic domain concentration and uniformity increased (by 2.04% and 6.30%, respectively), with negligible change in the non-treatment domain. These results demonstrate the feasibility of DRL for adaptive DES optimization and highlight its potential for patient-specific applications and guidance in clinical selection of drug-embedding configurations.
12 Apr 2026
Read appraisal →PloS one
Adaptive Feature Fusion Gate and Gated Channel-Spatial Attention in CNN-Transformer Models for Music Genre Classification
With the rapid growth of music data, automatic music genre classification has become a critical task in music information retrieval. Traditional methods based on handcrafted features are increasingly inadequate when handling large-scale analysis. This paper proposes the Convolutional Neural Network-Gated Transformer Network (CT-GateNet), a hybrid architecture that integrates a gated channel-spatial attention mechanism with an adaptive feature fusion gating mechanism to achieve discriminative feature learning and efficient feature integration. To mitigate data scarcity, a data augmentation strategy based on a denoising diffusion probabilistic model is introduced. Experiments are conducted on three public music genre datasets: GTZAN, FMA-SMALL and FMA-Medium. The method achieves classification accuracies of 98.72%, 89.42%, and 69.07% on GTZAN, FMA-SMALL and FMA-Medium, respectively, demonstrating outstanding performance and robust generalization capabilities. These results validate CT-GateNet's effectiveness in music genre classification and provide valuable insights for audio classification research.
11 Apr 2026
Read appraisal →Journal of controlled release : official journal of the Controlled Release Society
Redox-active microneedles for disease management.
Reactive oxygen and nitrogen species are essential to several biological functions, from maintaining cellular homeostasis to driving disease progression and aging. As such, the precise modulation of cellular redox states has emerged as a promising medical strategy. In parallel, microneedle technology has rapidly advanced over the past two decades, offering minimally invasive, targeted, and painless delivery of therapeutics. This review presents a comprehensive analysis of the emerging field of redox-active microneedles (RAMs), a novel class of wearable medical devices with broad potential for treating infected wounds, cancer, psoriasis, diabetes, arthritis, hair loss, skin aging, and potentially many other conditions. By integrating redox biology with advanced microneedle-based drug delivery systems, RAMs are poised to transform personalized and precision medicine, an evolution supported by a growing number of patents, clinical trials, and rapid advancements in catalytic, photodynamic, and nanomedicine research.
11 Apr 2026
Read appraisal →PloS one
MSCNet: Efficient and accurate semantic segmentation of LiDAR data using Multi-scale Convolution
In autonomous driving and intelligent robotics, the semantic information of LiDAR (Light Detection and Ranging) sensor data is crucial for understanding the surrounding environment. However, directly operating on point clouds is computationally expensive. To address this, some researchers have projected three-dimensional LiDAR data onto a two-dimensional spherical range view and used two-dimensional convolutional neural networks to segment the projected images. While the results are promising, many of these models are structurally complex, with high spatiotemporal complexity, which makes them unsuitable for real-time applications. To solve these issues, this paper proposes a multi-scale LiDAR data semantic segmentation method, MSCNet, with fewer parameters and higher segmentation accuracy. In the encoding phase, a single-channel multi-scale feature fusion block is introduced to alleviate the distribution differences between input channels. To obtain more stable local features, multi-scale dilated convolution residual blocks are designed to encode information from different receptive fields. To quickly capture global features, a pyramid pooling module is introduced. Experimental results on the SemanticKITTI, SemanticPOSS, and Pandaset datasets show that MSCNet achieves a good balance between parameter, accuracy, and running time. Particularly on the SemanticPOSS and Pandaset datasets, MSCNet achieves the best performance. Under the same parameter conditions, this method outperforms existing point cloud-based and projection-based methods.
10 Apr 2026
Read appraisal →ACS applied materials & interfaces
Physically Cross-Linked Biopolymer-Based Strong and Stretchable Organohydrogel with Self-Healing and Biocompatibility for Environmentally Tolerant Strain Sensors
The development of wearable strain sensors increasingly focuses on multifunctional materials that integrate mechanical robustness, self-repairing capacity, temperature-resilient conductivity, and biocompatibility. Herein, we developed a physically cross-linked biopolymer-based organohydrogel by incorporating gelatin (Gel) and β-cyclodextrin-grafted chitosan (CDCS), designated as Gel/CDCS, reinforced through hydrogen bonding, ionic interactions, and host-guest complexation. The optimized Gel/CDCS1 organohydrogel with 1% of CDCS exhibited outstanding stretchability, with a tensile strain at break of 410.9% and a tensile strength of 92.2 kPa, while maintaining 85% of its mechanical properties after self-healing at 40 °C. The introduction of kosmotropic salts (Na3Cit and NaCl) and a glycerol-water binary solvent system conferred remarkable environmental adaptability, allowing stable performance from -20 to 37 °C. Notably, it achieved a high ionic conductivity of 0.946 S m-1 at room temperature, alongside 0.606 S m-1 at -20 °C and 0.330 S m-1 at 37 °C, ensuring operation under extreme conditions. The strain sensor displayed high sensitivity with a gauge factor (GF) increasing from 1.06 to 1.34 over a broad strain range and maintained a stable electrical response over 600 consecutive stretching cycles. Additionally, it enabled reliable detection of diverse human motions, including facial expressions, joint bending, and breathing, demonstrating excellent conformability and signal stability during on-body monitoring. Additionally, Gel/CDCS1 exhibited excellent biodegradability, degrading by over 80% in soil within 4 days, and showed high biocompatibility with fibroblast viability of around 80%. These attributes establish Gel/CDCS1 as a promising eco-friendly candidate for next-generation flexible electronics, wearable strain sensors, and biomedical applications.
10 Apr 2026
Read appraisal →ACS applied materials & interfaces
Bioinspired Janus Electronic Textiles with Moisture and Thermal Management
Wearable electronics have attracted great attention in recent decades. However, excessive sweat accumulation on skin often has a significant effect on the signal stability, comfort, and safety of electronic skin in practical applications. Herein, a skin-comfortable Janus electronic textile (e-textile) that integrates unidirectional liquid transport and passive radiative cooling properties is developed. This e-textile is fabricated through laser ablation and single-side plasma treatment of carbon nanotubes (CNTs) doped with polydimethylsiloxane (PDMS). The resulting e-textile features a conical pore structure and asymmetric wettability, enabling unidirectional sweat transport from the skin surface to a superhydrophilic silk layer, thereby maintaining a dry and comfortable skin microenvironment. The Janus e-textile achieves a solar reflectivity of 69.2% and a mid-infrared emissivity of 98.3%, reducing the temperature by approximately 5.0 °C compared with a cotton fabric under solar intensity, demonstrating a superior radiative cooling performance. Moreover, the Janus e-textile exhibits stable resistance changes under cyclic stretching and bending, accurately detecting subtle movements across various body parts while reliably acquiring electromyography (EMG) signals. This work positions the Janus e-textile as a pioneering solution for next-generation wearable technologies, integrating thermal and moisture management with advanced resistive strain sensing and EMG capabilities for applications in sports, rehabilitation, and health monitoring.
10 Apr 2026
Read appraisal →ACS applied materials & interfaces
Textile-Based Wearable Electrothermal Pad Using Woven AgNW/EG-PEDOT:PSS Heaters and PS/PEDOT:PSS Thermistor Yarns
Wearable thermal management devices are gaining increasing attention for healthcare, rehabilitation, and personal comfort applications. Here, we present a textile-based wearable electrothermal pad that integrates a woven heater with a yarn temperature sensor on a flexible, breathable fabric platform. The heater element was fabricated using cotton yarn coated with silver nanowires (AgNW) and poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) doped with ethylene glycol (EG-PEDOT:PSS), enabling uniform heat distribution, washability, and low-voltage operation. The AgNW/EG-PEDOT:PSS yarn heater rapidly reached 45 °C within 60 s at an applied voltage of 1.5 V and demonstrated good thermal stability and repeatability over 2,000 stretching-releasing cycles. The temperature sensor, based on polystyrene-coated PEDOT:PSS (PS/PEDOT:PSS) yarn, exhibited thermistor behavior with resistance decreasing as the temperature increases. It showed a negative temperature coefficient of resistance of 0.25%°C-1 and a response time of less than 130 s. The yarn temperature sensor also maintained resistance stability under bending and stretching-releasing conditions. To demonstrate a wearable electrotherapy platform, the PS/PEDOT:PSS yarn temperature sensor was integrated with a woven AgNW/EG-PEDOT:PSS heater to enable closed-loop feedback for thermal regulation. Programmable, on-demand thermotherapy and real-time thermal management could be wirelessly controlled via a smartphone. This custom-designed electrothermal pad offers a promising solution for wearable thermotherapy and smart clothing applications.
10 Apr 2026
Read appraisal →ACS applied materials & interfaces
Evolving Beyond Nature: AI-Driven Rapid Advancements of Bio-Inspired Electronic Skin
Electronic skin (e-skin) is a rapidly developing technology at the intersection of humans and machines with tremendous applications in robotics, healthcare, sports performance, and other fields of wearable electronics. With the recent popularization of artificial intelligence (AI), data-driven approaches to materials design and geometric functional optimization have accelerated e-skin development, enabling optimized microstructures, novel functionalities, and improved usability. Recent progress in bioinspired materials design and functionality enhancements for e-skin is paving the way for future-facing machine learning-powered tools to further accelerate their development. In this review, we discuss the traditional properties of e-skins, their sensing mechanisms, and suitable applications. Then, we investigate and categorize bioinspired structures from material, structural, and functional perspectives. Next, we outline the fundamentals of AI for data-driven optimization and how it can be utilized toward bioinspired e-skin design. This blended e-skin design framework culminates in novel applications, from customizable layouts for personalized therapeutics to dynamic environment recognition for enhanced robotics. Finally, we will explore the challenges of e-skin human-machine interfaces, including their hardware limitations, sensitivity, and longevity issues, and offer insights into future-proof research.
10 Apr 2026
Read appraisal →ACS applied materials & interfaces
NeuroTex: Functionalized Hybrid Polymer-Metal Textile for Soft, Robust, and Dry Auricular Neural Interfaces
Wearable bioelectronics would benefit from dry, easy-to-apply electrodes for sensing physiological signals. Bioelectronic interfaces must maintain high conductivity, low electrode-skin impedance, and reliable coupling between the electrode and skin under sweat, motion, and prolonged use. Metallic textiles offer good conductivity and stable connections but are hydrophobic, prone to cracking and oxidation, and dependent on conductive gels to achieve the low impedance required for low-amplitude electrophysiological signals. Here, we introduce NeuroTex, a commercially available silver nanowire textile that is PEDOT:PSS-functionalized to improve the electrode-skin-electronics interface. The PEDOT:PSS coating reduced the area-normalized resistance of the textile by almost 50%, suppressed silver oxidation in phosphate-buffered saline, and lowered the contact impedance by more than 85%. Mechanical evaluations under torsion, stretching, bending, and low-force fatigue confirmed the preserved fiber integrity and intact electronic connectivity. Furthermore, the soft and hydrophilic properties of the coating enhanced the conformal skin interfacing. Integrated into an in-ear monitoring platform, NeuroTex enabled multimodal biosignal recordings of in-ear electroencephalography (ieEEG), electro-oculography (ieEOG), electromyography (ieEMG), and electrocardiography (ieECG). Reliable acquisition of the technically demanding ieEEG signal was achieved, demonstrating alpha-modulation responses comparable to those of gel-based electrodes. ieEEG recordings recorded over several hours captured distinct sleep patterns to further confirm stable performance during natural sleep. These results establish NeuroTex as a gel-free, mechanically resilient, and electrochemically stable interface suitable for wearable physiological monitoring.
10 Apr 2026
Read appraisal →Science advances
Filament sensing tension: A bionic artificial tendon for self-force-regulated artificial muscle-driven wearable robotics
Artificial muscles have been in development for decades and play a crucial role in the field of wearable robotics. However, a universal device for both sensing and regulating forces in artificial muscles is still absent, rendering scenario-required force control unfeasible. Inspired by Golgi tendon organs, this work proposes bionic ExoTendon for force sensing and regulating in artificial muscles as a general solution. Leveraging the principles of triboelectrification and electrostatic induction, the ExoTendon offers high linearity, minimal hysteresis, and adjustable sensitivity and range, making it suitable as the tendon of artificial muscles. Specifically, the ExoTendon effectively enable adjustment to the pretension of the artificial muscles, yielding various assistive forces under the same input. With the optimized pretension level and closed-loop force control, the artificial muscle-based exosuit achieves self-force regulation, substantially improving walking balance and speed in stroke subjects under low input, demonstrating its promise in wearable robotics and rehabilitation medicine.
9 Apr 2026
Read appraisal →Physiological measurement
Continuous blood pressure monitoring via hemodynamic parameter and pulse transit time derived from capacitive sensing pads
Objective.Continuous blood pressure (BP) monitoring is crucial for detecting nocturnal hypertension and acute hemodynamic changes. Conventional cuff-based methods disrupt sleep and miss transient BP fluctuations from sleep-related events or instability. Cuffless methods, such as pulse transit time (PTT), offer potential but often struggle to reliably track acute BP fluctuations due to complex and nonlinear hemodynamics.Approach.We developed a capacitive sensing pad system incorporating PTT with additional hemodynamic features for unobtrusive, continuous BP monitoring in supine subjects. The pad contains ultra-sensitive single-electrode capacitive (SEC) sensors made from carbon nanotube composites. In passive contact, SEC sensor on the back of the chest captures intrathoracic blood volume changes through deep tissue permittivity and ballistocardiography, while another under the leg detects arterial pulse-induced tissue vibrations. PTT is derived from the temporal delay between chest and leg signals. A neural network model incorporates PTT and intrathoracic blood volume features to improve BP estimation.Main results.In human trials (N= 30) with arm-cuff reference, the system showed strong correlation (r⩾ 0.94), with a mean error (ME) ⩽ 0.1 mmHg and standard deviation (SD) ⩽ 5.6 mmHg. In a subset (N= 8) with continuous finger-cuff reference, it maintained strong correlation (r⩾ 0.89), with a ME ⩽ 0.2 mmHg and SD ⩽ 7.8 mmHg.Significance.These results suggest the possibility of bed-based unobtrusive and continuous BP monitoring in the supine position.
9 Apr 2026
Read appraisal →Biomedical physics & engineering express
Bluetooth low energy thermohygrometric sensor for continuous respiratory rate monitoring: 48 h breathing simulator and human proof-of-concept
Continuous respiratory rate monitoring is essential for early detection of clinical deterioration, but conventional methods like capnography are often impractical outside high-acuity settings. We developed a low-cost, Bluetooth Low Energy (BLE) respiratory rate monitor incorporating a MEMS temperature-humidity sensor (AHT21) and Nordic nRF52840 microcontroller, designed to attach externally to a standard oxygen mask. The device detects breaths using rapid, cyclic changes in exhaled temperature and humidity via a slope-based time-domain algorithm. Bench validation over 48 h using a breathing simulator (5-60 breaths min-1, tidal volume 300 ml, 34 °C, 95%-100% RH) demonstrated 96.4% overall breath detection accuracy and strong correlation with programmed rates (r = 0.996, P < 0.0001). Sensor saturation occurred at high respiratory rates (60 breaths min-1), and high ambient humidity levels (90% RH) both of which reduced measurement accuracy. Human testing in three healthy adults showed strong agreement with capnography (mean bias 0.44 breaths min-1; 95% limits of agreement -1.86 to 2.75), with minor under-detection at the highest rate (50 breaths min-1). The device operated reliably across oxygen flow rates of 4-15 l min-1, with reduced performance at 0-2 l min-1due to saturation of the sensor. These results demonstrate that BLE-enabled MEMS thermohygrometric sensors can accurately monitor respiratory rate under controlled conditions while highlighting operational limits due to sensor saturation. This study extends prior work on humidity-based respiratory monitoring and illustrates the potential for low-cost, portable respiratory rate devices suitable for low-acuity or resource-limited clinical settings.
9 Apr 2026
Read appraisal →Physiological measurement
Towards fair and trustworthy heart rate estimation from wrist-worn photoplethysmography: a multi-wavelength dataset and uncertainty-aware deep learning approach evaluated across skin tones, sexes, and motion conditions
Objective.To improve fairness (reduced disparities across skin tones and sexes) and trust (well-calibrated uncertainty metrics that indicate unreliable predictions) in wrist-worn photoplethysmography (PPG) heart rate estimation. While PPG offers a convenient, low-cost method for continuous heart rate monitoring, accuracy is often degraded by motion artefacts and demographic variation.Approach.We collected a multi-wavelength (blue, green, red, IR) wrist-worn PPG dataset covering diverse activities, skin tones, and sexes, and developed a multi-branch PPG-accelerometer fusion convolutional neural network quantifying both aleatoric and epistemic uncertainty (the latter via Monte Carlo dropout, concrete dropout, and deep ensembles). We compared wavelength sets, quantified demographic effects, and applied uncertainty-aware post-processing to filter out high-uncertainty predictions. Evaluation used leave-one-subject-out cross validation across both the collected multi-wavelength dataset and five commonly used single-wavelength datasets.Main results.Multi-wavelength fusion improved accuracy under motion, with blue-green-red-IR achieving the lowest mean absolute error (MAE). Baseline error was higher for participants with darker skin tones and for females; uncertainty-guided rejection reduced MAE and effectively eliminated these gaps (reducing parity gaps from 1.2 BPM to 0.0 BPM for skin tone). Concrete dropout provided the best calibration-accuracy-efficiency trade-off, achieving the lowest average miscalibration area (MA=0.12) while maintaining competitive MAE.Significance.Combining calibrated uncertainty with group-aware evaluation yields a more fair and trustworthy wrist-worn PPG heart rate estimation pipeline, supporting responsible deployment in consumer and sports contexts and informing bias-aware clinical development.
9 Apr 2026
Read appraisal →Physics in medicine and biology
SSI-Net: a hybrid physics-constrained deep learning framework for quantitative ultrasound speed-of-sound reconstruction
Objective.Quantitative ultrasound tomography faces challenges in reconstructing speed-of-sound (SoS) distributions due to the ill-posed nature of the inverse problem and the computational complexity of full-waveform inversion. This study aims to develop an efficient and physically consistent deep learning framework for accurate SoS mapping from ultrasound signals.Approach. We propose speed of sound inversion network (SSI-Net), a dual data-physics-driven framework that integrates a bidirectional gated recurrent unit encoder for temporal feature extraction, a U-Net decoder for high-resolution spatial mapping, and a physics-constraint module based on an exact, differentiable finite-difference time-domain solver of the nonlinear Westervelt equation. The network model is trained through a joint optimization strategy that minimizes both data-fidelity loss and physics-informed residual loss, with a two-phase training schedule to ensure stable convergence.Main results.SSI-Net was validated across simulated, tissue-mimicking phantom, andin vivomouse datasets. It achieved superior reconstruction accuracy, with peak signal-to-noise ratio improvements of 7.4%-18.9% over state-of-the-art methods, while reducing the physics-based residual by 25.6%-57.5%. The framework demonstrated strong generalization to different transducer geometries (ring and linear arrays) and maintained stable performance across a ten-fold frequency range. Inference was completed within 109.4 ms per sample.Significance.By embedding an exact wave solver into a trainable architecture, SSI-Net combines the representational capacity of deep learning with strict physical consistency. It provides a robust, efficient, and clinically translatable solution for quantitative SoS imaging, offering a promising tool for tissue characterization in diagnostic ultrasound.
9 Apr 2026
Read appraisal →Physiological measurement
NeuroSleep: neuromorphic event-driven single-channel EEG sleep staging for edge-efficient sensing
Objective. Reliable, continuous neural sensing on wearable edge platforms is fundamental to long-term health monitoring; however, for electroencephalography (EEG)-based sleep monitoring, dense high-frequency processing is often computationally prohibitive under tight energy budgets.Approach. To address this bottleneck, this paper proposes NeuroSleep, an integrated event-driven sensing and inference system for energy-efficient sleep staging. NeuroSleep first converts raw EEG into complementary multi-scale bipolar event streams using Residual Adaptive Multi-Scale Delta Modulation, enabling an explicit fidelity-sparsity trade-off at the sensing front end. Furthermore, NeuroSleep adopts a hierarchical inference architecture that comprises an Event-based Adaptive Multi-scale Response module for local feature extraction, a Local Temporal-Attention Module for context aggregation, and an epoch-leaky integrate-and-fire module to capture long-term state persistence.Main results. Experimental results using subject-independent 5-fold cross-validation on the sleep-EDF expanded sleep-cassette subset (78 subjects, 153 overnight recordings) with single-channel EEG demonstrate that NeuroSleep achieves a mean accuracy of 74.2% with only 0.932 M parameters while reducing sparsity-adjusted effective operations by approximately 53.6% relative to dense processing. Compared to the representative dense Transformer baseline, NeuroSleep improves accuracy by 7.5% with a 45.8% reduction in computational load.Significance. By coupling neuromorphic event encoding with state-aware context modeling, NeuroSleep offers a deployment-oriented framework for single-channel sleep staging that reduces redundant high-rate processing and improves energy scalability for wearable and edge platforms.
8 Apr 2026
Read appraisal →Journal of neural engineering
Large-scale training data enhances silent speech decoding with around-ear EEG.
Objective. Silent speech decoding (SSD) offers a potential communication alternative for individuals with impaired vocalization. However, conventional multi-electrode electroencephalography (EEG) or facial electromyography (EMG) systems require cumbersome preparation and are unsuitable for daily use. This study evaluates the practicality of SSD using a wearable around-ear EEG device, focusing on data scaling, cross-subject transfer, vocabulary extensibility, and online decoding performance.Approach. We collected 72 h of around-ear EEG from 24 healthy participants and one individual with incomplete locked-in syndrome (LIS) during silent, vocalized, and attempted speech, and integrated these around-ear EEG recordings with prior EMG + high-density EEG datasets, yielding 282.4 total h of training data. Using a 64-word classification task as the evaluation metric, we assessed: (1) whether larger datasets improve around-ear EEG-based SSD, (2) whether healthy-participant data supplement limited LIS-participant data despite articulatory differences, (3) transferability to unseen vocabulary, and (4) online user-interface performance.Main results. Large-scale EEG/EMG data improved SSD accuracy in both healthy participants and the LIS participant. Training on the heterogeneous dataset achieved 56.6% accuracy for healthy users and 47.3% for the LIS participant. Fine-tuning this decoder for new vocabulary increased the accuracy by 22 percentage points relative to training from scratch. Regression analysis showed that, for decoding in the LIS participant, data from the LIS participant contributed approximately four times the weight of healthy-participant data, quantifying data strategies for SSD. Online experiments achieved top-1/top-5 accuracies of 47.2%/76.0% for healthy users and 26.5%/49.1% for the LIS participant.Significance. The results indicate that lightweight, commercially feasible around-ear EEG can enable practical SSD when combined with large-scale healthy-participant data, supporting online operation. Moreover, models trained on a 64-word vocabulary facilitate decoding of a new vocabulary, providing a path toward SSD systems requiring minimal LIS-participant data. This study advances non-invasive SSD systems suitable for everyday communication.
8 Apr 2026
Read appraisal →Journal of neural engineering
EDAPT: towards calibration-free BCIs with continual online adaptation
Objective.Brain-computer interfaces (BCIs) suffer from accuracy degradation as neural signals drift over time and vary across users, requiring frequent recalibration that limits practical deployment. Our goal is to develop a framework that eliminates the need for separate calibration phases by enabling continual, real-time model adaptation to new users and changing signal characteristics.Approach.We propose EDAPT, a task- and model-agnostic framework for continual online learning. EDAPT first establishes a robust baseline decoder through population-level pretraining (PRE) on data from multiple users. It then personalizes this model during deployment using supervised continual finetuning on a trial-by-trial basis. Due to its modular design, EDAPT can be composed with unsupervised domain adaptation (UDA) techniques to further address distribution shifts.Main results.We validate EDAPT across nine datasets, three BCI paradigms, and four deep learning architectures. EDAPT consistently improves decoding accuracy over static models for nearly all subjects and datasets, raising mean balanced accuracy from 0.80 to 0.87 on representative datasets (table3). Ablation studies confirm that the combination of PRE and online finetuning is the primary driver of this performance gain, with further improvements on some datasets when using UDA techniques. We demonstrate real-time feasibility of the framework, with adaptation latencies under 200 ms on consumer-grade hardware. Our scaling analysis further reveals that decoding accuracy is primarily determined by the total pretraining data budget, rather than its specific allocation between subjects and trials.Significance.These findings demonstrate that continual online learning is a practical and effective strategy for creating high-performance, user-adaptive BCIs. By systematically addressing the bottleneck of model recalibration, EDAPT reduces a major barrier to the widespread adoption of BCI technology and helps advance neurotechnology toward robust, user-friendly, real-world applications.
8 Apr 2026
Read appraisal →PloS one
Advanced microfluidic and 3D cell culture platforms for modeling vascularization in diabetic foot ulcers: A systematic review of translational challenges and perspectives
The healing process of diabetic foot ulcers (DFUs) presents a slow pattern with an increased risk of infections, ischemia, and thrombosis correlated with high levels of reactive oxygen species production. Vascular injury is one of the factors contributing to the difficulty of wound healing in diabetic patients. Although the understanding of the pathophysiology of DFUs has significantly increased in recent years, associated treatments still have a high level of failure, leading to high morbidity rates, mortality, and amputations. Three-dimensional (3D) cell culture platforms offer a new approach to investigating and treating these wounds, as they can reproduce one or more physiological systems in a relevant microenvironment. This systematic review describes the advancements, challenges, and future implications of advanced 3D culture models in vascularization, encompassing pathophysiological understanding, treatment, and prognostic perspectives. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. This work was registered under PROSPERO protocol number CRD42022336473. The eligibility criteria addressed studies related to chronic DFUs that analyzed vascularization during the healing process. The selected study designs involved 3D cultures and organ-on-a- chip (OoC) platforms, utilizing either primary or secondary human cell lines. Studies published more than 10 years ago or using only animal cells in 2D culture environments were excluded. The search was conducted in PubMed, LILACS, Embase, MEDLINE, IEEE, the BSV regional portal, ScienceDirect, Scopus, CINAHL, EBSCO, and Web of Science. A total of 2,539 relevant studies were identified as of June 1, 2025. After screening, only seven met the inclusion criteria. These studies reflect a growing interest in using hydrogel scaffolds and microfluidic systems to replicate diabetic skin environments; however, the field remains in an early stage of development. OoC platforms, in particular, stand out for their ability to recreate dynamic, tissue-like conditions and vascular function. While some promising attempts have been made to combine hydrogels with these technologies, the evidence is still limited and inconclusive for chronic wound modeling. This review underscores both the potential and the urgent need for more robust and translational research in this area-especially toward building personalized, clinically relevant models to support future drug testing and therapeutic innovation.
7 Apr 2026
Read appraisal →PloS one
A robust deep learning approach for impulse noise filtering using hybrid auto-encoder with fuzzy median filter
De-noising convolutional neural networks (DnCNNs), are a powerful nonlinear mapping models in image processing for impulse noise removal. During training and validation, a set of 12 standard testing images is used to evaluate model performance. DnCNNs demonstrate strong capability in classification of impulse noise with excellent results. To evaluate de-noising performance, a suitable noise ratio should be added so that most appropriate DnCNN model can be used for impulse noise detection. This research proposes an effective image restoration technique that integrates DnCNN and an autoencoder with a fuzzy median filter to detect and eliminate high-density impulse noise. The proposed deep learning de-noising technique used to classify noisy and clean pixels, and result are presented in different metrics such as accuracy, FPR, FNR and f1 score. Further, to remove impulse noise an auto-encoder with fuzzy median filter are used that then reconstructs the clean image based with parametric values. Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), are used to assess our methodology, it is compared to conventional impulse noise filtering techniques, experimental results indicate a significant improvement in image quality. Based on the final de-noised images, this research contributes to developing deep learning-based, de-noising techniques that enhance image restoration quality while preserving image details and essential features.
5 Apr 2026
Read appraisal →Nanoscale
A machine learning-assisted Ag-TiO2 SERS platform for intraoperative osteomyelitis diagnosis
Osteomyelitis is a progressive bone infection with high disability and recurrence rates. However, current diagnostic approaches, including bacterial culture, imaging, and serological assays, suffer from poor sensitivity, low specificity, and delayed turnaround, making intraoperative evaluation particularly challenging. Here, we present an Ag-TiO2-based surface-enhanced Raman scattering (SERS) platform for rapid, label-free detection of osteomyelitis using minimal volumes of wound saline irrigation fluid (WSIF). Benefiting from the synergistic interplay between electromagnetic field amplification and enhanced interfacial charge-transfer interactions, the Ag-TiO2 substrate provides ultrasensitive detection with high spectral reproducibility, enabling reliable identification of subtle biochemical changes in complex clinical samples. Quantitative analysis indicates that strengthened charge-transfer processes significantly enhance molecular Raman responses, while plasmonic electromagnetic effects remain essential for overall signal amplification. Distinct Raman fingerprints of infected, non-infected, and recovered patients were successfully classified using artificial intelligence (AI) models, revealing metabolic alterations in amino acids, purines, proteins, and lipids associated with bacterial infections and immune response. This integration of SERS with AI not only elucidates the mechanistic basis of spectral enhancement but also demonstrates clinical feasibility for intraoperative osteomyelitis diagnosis, offering a promising strategy to improve surgical precision and reduce recurrence risk.
5 Apr 2026
Read appraisal →Biofabrication
High-fidelity 'top-down' DLP bioprinting of multi-material soft tissue constructs enabled by computer vision-based layer control
Bioprinting continues to redefine the frontiers of regenerative medicine by enabling the fabrication of complex, three-dimensional (3D) tissue constructs that emulate native biological and mechanical functions. However, despite significant progress, critical challenges remain, particularly in achieving precise multi-material integration, high-resolution patterning, and structural fidelity necessary for functional tissue engineering. A major limitation in 'top-down' vat photopolymerization bioprinting, especially digital light processing (DLP)-based approaches, lies in the precise control of layer thickness, a parameter that directly affects mechanical integrity, biological activity, and spatial resolution. This study presents a novel, automated platform designed to overcome one of the most persistent bottlenecks in multi-material top-down DLP bioprinting: the real-time, accurate measurement of the dynamic gap between the cured layer and the bioink surface. Through a comparative assessment of classical computer vision and deep learning (convolutional neural network-based) techniques, we demonstrate a system capable of achieving sub-0.1 mm precision (0.092 mm) with strong correlation to mechanical measurements (R= 0.994). This vision-based system adapts to a wide range of bioinks with varying viscosities, opacities, and photopolymerization kinetics, eliminating the need for manual recalibration during material switching. As a demonstration of its capabilities, we successfully printed a multimaterial vascular-like tissue structure with high spatial fidelity across heterogeneous biomaterials. Further, we bioprinted a multimaterial skin tissue model, featuring compartmentalized dermal/bone analogs, to enablein vitrofunctional evaluation. These case studies highlight the platform's potential to advance biofabrication workflows by improving reproducibility, material adaptability, and structural precision, paving the way toward clinically scalable tissue manufacturing systems.
4 Apr 2026
Read appraisal →Journal of biophotonics
Cell-MICS: Detecting Immune Cells With Label-Free Two-Photon Autofluorescence and Deep Learning
Multiphoton imaging has been widely used for deep-tissue imaging. Although its label-free, metabolic contrast is ideal for investigating inflammation, the label-free two-photon induced autofluorescence is often regarded as less specific compared to conventional antibody markers. In this work, we investigate the potential for multiphoton imaging with computational specificity (MICS) by training a convolutional neural network on images of different immune cells. A low-complexity squeezeNet architecture was able to achieve reliable immune cell classification results (0.89 ROC-AUC, 0.95 PR-AUC for binary classification between T cells and neutrophils; 0.689 F1 score, 0.697 precision, 0.748 recall for multi-class classification between six isolated cell types). Perturbation tests confirmed that the model was not confused by the extracellular environment and that 2P-AF from NADH and FAD is equally important for the classification. In the future, deep learning could provide computational specificity for specific immune cells in unstained tissues, with great potential for label-free in vivo endomicroscopy.
4 Apr 2026
Read appraisal →Physiological measurement
Deep learning approach for stroke volume variation estimation: retrospective modeling and prospective deployment in real-world practice
Objective. We developed stroke volume variation (SVV) Net, a deep learning-based model for estimating SVV, and validated its performance and clinical applicability through both large-scale retrospective modeling and prospective real-time deployment within intraoperative settings.Approach. A SVVNet was trained on arterial blood pressure (ABP) waveforms from 2849 patients undergoing general anesthesia at Asan Medical Center (AMC). For retrospective validation, we utilized AMC cohorts and a publicly accessible database to assess agreement, trending ability, noise robustness (quantifies via signal quality index), and inter-individual deviation against ABP-based pulse wave analysis (ABP-PWA) and other DL models. For prospective validation, the optimized model was deployed onto a tablet and prospectively evaluated in real-time during 67 surgeries.Main results. In the retrospective external validation, SVVNet achieved higher agreement mean absolute error (MAE = 0.70, Pearsonρ= 0.92), compared to ABP-PWA algorithm (MAE = 1.30, Pearsonρ= 0.80) and lower inter-individual deviation. It was more robust to signal noise (MAE = 4.33 vs 30.87). In the prospective real-time deployment, our deployed model achieved an MAE of 1.33 and a Pearson correlation of 0.87 against the ground truth.Significance. This study emphasizes the advantages of the DL-based model compared to ABP-PWA-based algorithms and presents prospective results that support the clinical applicability of SVVNet in the operating room. Our model has the potential to replace pre-existing monitoring systems and to enhance the complex surgical environment.
4 Apr 2026
Read appraisal →Physiological measurement
StructEIT: realistic 3D EIT model generation from CT scans for deep learning applications
Objective.Artificial intelligence has significantly improved image reconstruction quality across various medical imaging modalities. However, its application in electrical impedance tomography (EIT) reconstruction remains limited, mainly due to the absence of comprehensivein vivodatasets that incorporate realistic anatomical geometries and conductivity distributions. This limitation constrains the development of supervised and data-driven reconstruction methods.Approach.To address this bottleneck, we developed StructEIT, an integrated EIT modeling framework for generating anatomically and biophysically realistic EIT simulation models. The framework incorporates three key components: (1) a structure extraction module, which automatically processes human CT scans to extract body contours and organ boundaries, thereby providing high-fidelity spatial geometry for three-dimensional finite element modeling; (2) a surface electrode attachment module, which enables flexible and accurate placement of electrodes on irregular body surfaces, supporting diverse configurations and ensuring precise definition of the electrode-tissue interface; and (3) a tissue property assignment module, which establishes frequency-dependent conductivity models for multiple organs, enabling physiologically realistic conductivity values across tissues.Main results and Significance.By bridging the gap between CT imaging and EIT, StructEIT facilitates flexible, realistic, and scalable generation of high-resolution EIT datasets. Using this framework, we constructed Chest-EIT, a thoracic EIT simulation dataset comprising over 1400 publicly available CT cases, with multiple electrode configurations provided for each case.
4 Apr 2026
Read appraisal →Computer methods in biomechanics and biomedical engineering
Reconstruction of blood flow velocity with deep learning information fusion from spectral ct projections and vessel geometry
In this work, we investigate a new deep learning reconstruction method of blood flow velocity within deformed vessels from contrast enhanced X-ray projections and vessel geometry. The principle of the method is to perform linear or nonlinear dimension reductions on the Radon projections and on the mesh of the vessel. These low dimensional projections are then fused to obtain the velocity field in the vessel. The accuracy of the reconstruction method is proved using various neural network architectures with realistic unsteady blood flows. The approach leverages the vessel geometry information and outperforms the simple PCA-net.
4 Apr 2026
Read appraisal →Advanced science (Weinheim, Baden-Wurttemberg, Germany)
A Wireless, Battery-Free Artificial Throat Patch with Deep Learning for Emotional Speech Recognition
Technologies for sensing and translating throat signals into speech have advanced the communication for people with vocal impairments. However, the practical deployment of these technologies is currently hindered by tethered system, bulky components and the lack of emotional recognition capability. Here, we demonstrate a wireless, battery-free artificial throat patch system (ATPS) coupling with deep learning method to enable simultaneous recognition of speech and emotion. The sensing module of the ATPS integrates a carbon nanotube-based thin-film strain sensor and a miniaturized flexible printed circuit board (FPCB). Real-time sensing and transmission of throat signals is facilitated by a smartphone-linked system incorporating an FPCB with an embedded near-field communication antenna and low-power electronic components. Furthermore, the information-rich sensor readouts are further extracted and successfully demonstrated to be sufficient to emotional recognition of speech by implementing a hybrid deep learning architecture. This proposed system expands the arsenal of tools available for mute patients in speech recognition and communication and further realizes personalized clinical implementations.
4 Apr 2026
Read appraisal →Expert review of medical devices
Modalities and algorithms for generalized motor seizure detection and prediction: a scoping review
INTRODUCTION: Generalized tonic - clonic seizures (GTCS) is a major cause of sudden unexpected death in epilepsy (SUDEP); thus, continuous monitoring is essential. Electroencephalography (EEG) is the diagnostic gold standard; however, non-EEG wearables have emerged as promising, user-friendly alternatives enhancing comfort, mobility, and social acceptance. AREAS COVERED: This scoping review compared sensing modalities and computational methods used for seizure detection and prediction, identified effective sensor combinations, and highlighted current research gaps. Following PRISMA-ScR guidelines, Scopus, IEEE Xplore, and PubMed databases were searched up to 22 April 2025. Twenty-nine studies met inclusion criteria. For seizure detection, combining electrodermal activity (EDA) with motion sensors - accelerometers (ACC) and gyroscopes (GYR) - and surface electromyography (sEMG) achieved high reliability, 98.6% precision, while using ACC and EDA alone yielded superior sensitivities up to 97.2% and a lower false alarm rate (FAR) of 0.53/24 h. Regarding seizure prediction, combining EDA, blood volume pulse (BVP), ACC, and temperature showed the highest sensitivity of 75.6%. EXPERT OPINION: Wearable multimodal non-EEG seizure detection and prediction systems offer personalized care but face validation, hardware, and privacy hurdles. Future success depends on efficient AI, IoT integration, patient-centric design, and clear regulations ensuring accessible and trustworthy clinical tools. People with epilepsy experience sudden episodes of abnormal brain activity called seizures, or fits. One serious type, the generalized tonic – clonic seizure (GTCS), causes loss of consciousness, muscle stiffening, and body jerking, which can lead to injury or even sudden unexpected death in epilepsy (SUDEP).Traditionally, seizures are monitored in hospitals using electroencephalography (EEG), which records brain signals. However, EEG systems are uncomfortable for daily use and limit continuous monitoring. Recent innovations in wearable sensor technology – such as motion sensors (accelerometers), skin conductance sensors (electrodermal activity), and heart or blood flow sensors (photoplethysmography) – offer a more practical, noninvasive alternative. These devices can be worn as smartwatches or bracelets, fitting naturally into everyday life.Artificial intelligence (AI) and machine learning play an increasingly important role in these systems. By recognizing complex patterns in motion and physiological signals, AI algorithms can automatically detect seizures and, in early research, may even predict when a seizure is likely to occur. While results so far are promising, more clinical studies and larger datasets are needed to confirm accuracy and reliability across different patient groups.Advances in wearable sensors and artificial intelligence could revolutionize epilepsy management by making continuous seizure monitoring and prediction possible in daily life. Although more evidence is required before widespread clinical adoption, these technologies hold strong potential to improve safety, autonomy, and quality of life for people living with epilepsy.
2 Apr 2026
Read appraisal →