Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 2 appraisals
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 →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 →