Heart2Mind: Human-Centered Contestable Psychiatric Disorder Diagnosis System using Wearable ECG Monitors

Wearable ECG monitors can objectively measure psychiatric symptoms: The system uses heart rate variability (HRV) and R-R intervals from wearable devices as objective biomarkers of autonomic dysfunction in psychiatric conditions, potentially addressing the current reliance on subjective clinical assessments that can delay diagnosis.

The system achieved 91.7% accuracy in identifying psychiatric disorders, demonstrating that combining cardiac monitoring with advanced AI analysis could provide a reliable diagnostic tool, though clinical validation would be needed before real-world deployment.

Explainability and contestability are built into the diagnostic process: Rather than presenting a "black box" diagnosis, the system uses Self-Adversarial Explanations to flag inconsistencies in AI predictions and allows clinicians to validate or challenge the AI's conclusions, putting human judgment at the center of decision-making.

Clinicians retain authority to override AI recommendations: The Contestable Diagnosis Interface enables doctors to contest erroneous predictions and validate correct ones using Large Language Models, ensuring that AI serves as a decision-support tool rather than an autonomous decision-maker in psychiatric care.

This approach could improve accessibility to psychiatric diagnosis: By enabling remote monitoring through wearable devices, the system has potential to help address diagnostic delays and accessibility barriers that currently affect millions with psychiatric disorders globally.