AI-driven tripartite classification for optimizing wearable bioelectronics in depression management

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: 95.2% overall classification accuracy reported for AI agent; subjects treated during pre-disease state showed faster recovery and greater therapeutic response to VNS compared to those treated after disease onset — no quantitative effect size provided

• Primary Outcome: Tripartite classification of depression states (normal, pre-disease, disease) using nine multimodal biomarkers with AI-based classification

• Nnt Or Sensitivity: Not reported — no sensitivity, specificity, positive predictive value, negative predictive value, AUC-ROC, or hazard ratio provided. The 95.2% accuracy figure alone is insufficient to characterise classifier performance, particularly given the absence of class prevalence data and the risk of accuracy paradox in imbalanced datasets

• Confidence Interval: Not reported

Clinical Application

Not currently feasible for clinical translation. Key barriers include: (1) requirement for invasive neural probe implantation incompatible with outpatient or community mental health settings; (2) undefined 'pre-disease' thresholds not mapped to clinically actionable criteria; (3) absence of regulatory approval pathway for the combined diagnostic-therapeutic device; (4) no health economic analysis; (5) continuous nine-biomarker monitoring infrastructure not available in standard care settings. In the Australian context, this research has no immediate clinical applicability. The RACGP guidelines for depression management emphasise validated screening tools (PHQ-9, K10) and stepped-care models. TGA would require substantial clinical trial evidence before any such device-based diagnostic-therapeutic system could be approved. The PBS does not currently list VNS for depression (it is available for epilepsy via specialist access). The Australian mental health system, including headspace, ATAPS, and Better Access initiatives, operates on psychosocial and pharmacological frameworks that are far removed from implantable bioelectronic monitoring. Future relevance may emerge if the technology is miniaturised, non-invasive, and validated in large human cohorts — but this is speculative at present. Currently not applicable to human clinical populations. If the underlying model is validated in humans, potential future applicability may include individuals at high risk for first-episode depression (e.g., those with subsyndromal symptoms, strong family history, or prior subthreshold episodes) who might benefit from early intervention monitoring.

Abstract

Current disease-sensing devices primarily focus on distinguishing between healthy and diseased states, effective for diagnosis but limited in guiding optimal intervention timing for prevention. We developed a tripartite framework identifying pre-disease state in depression, a reversible phase preceding irreversible onset. Using complex systems theory, we analyzed early-warning signals emerging as biological systems approach critical transitions. Continuous monitoring of nine multimodal biomarkers-spanning electrophysiological, behavioral, and biological-enabled classification into normal, pre-disease, and disease states by quantitatively defining critical points. An artificial intelligence agent classified disease states with 95.2% accuracy using multimodal data, enabled by ultrasoft neural probes for stable, low-damage recordings. Therapeutic validation with a skin-attachable wireless vagus nerve stimulator integrating soft three-dimensional electrodes demonstrated superior efficacy during pre-disease states. Subjects treated during pre-disease showed faster recovery and greater therapeutic responses, while those treated after disease onset failed to achieve full recovery. This framework provides evidence-based rationale for early intervention.

References

1. 1.Lee, J., Hong, Y.-M., Kim, E., Seo, H., Chung, W. G., Park, W., Song, H., Kim, S., & Park, J.-U. (2026). AI-driven tripartite classification for optimizing wearable bioelectronics in depression management. Science Advances. https://doi.org/10.6084/m9.figshare.31813771