Anxiety detection using neural and physiological signals and artificial intelligence: A comprehensive review.

Clinical Snapshot

Key Findings

• P Value: Not applicable; no statistical synthesis conducted

• Effect Size: Not reported (narrative synthesis only); the abstract references 'impressive accuracies' without specifying values

• Primary Outcome: Classification accuracy of ML/DL models for anxiety detection using EEG, ECG, PPG, and EDA signals; no pooled quantitative estimate provided

• Nnt Or Sensitivity: Not reported; individual study sensitivity/specificity data not synthesised. The review notes hybrid models (CNN-LSTM, Transformers) consistently deliver 'state-of-the-art results' in multimodal scenarios, but no quantitative benchmarks are provided in the abstract

• Confidence Interval: Not reported; no meta-analytic pooling performed

Clinical Application

Currently low for routine clinical implementation. Key barriers include: absence of clinically validated, large-scale datasets; lack of regulatory-grade model validation; interpretability deficits in complex deep learning architectures; cost and infrastructure requirements for multimodal physiological monitoring; and absence of standardised clinical workflows for integrating AI-derived anxiety indices into diagnostic or treatment decisions. In Australia, anxiety disorders affect approximately 1 in 6 adults and represent the most prevalent mental health condition (AIHW). Current RACGP guidelines for anxiety management rely on validated self-report tools (GAD-7, PHQ-9) and clinical interview — subjective instruments that this technology aims to supplement. No AI-based anxiety detection device is currently listed on the Australian Register of Therapeutic Goods (ARTG) or reimbursed under Medicare or the PBS. The TGA's Digital Health regulatory framework (Software as a Medical Device, SaMD) would apply to any clinical deployment. The Australian Digital Health Agency's national digital health strategy provides a relevant policy context. Equity considerations are paramount: wearable biosensor costs and digital literacy requirements may exacerbate existing mental health access disparities in rural, remote, and First Nations communities. Future research should include Australian population cohorts to ensure model generalisability across diverse demographic groups. Patients with anxiety disorders across the spectrum (GAD, panic disorder, social anxiety disorder, PTSD) in whom objective, continuous physiological monitoring may complement or augment subjective clinical assessment. Most applicable initially to specialist mental health settings with access to wearable biosensor technology.

Abstract

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.

References

1. 1.Ghonchi, H., Foulsham, T., & Ferdowsi, S. (2026). Anxiety detection using neural and physiological signals and artificial intelligence: A comprehensive review. Neuroscience and Biobehavioral Reviews, 106669. https://doi.org/10.1016/j.neubiorev.2026.106669