EEG-Based AI System for Early Detection and Non-Invasive In-tervention of Panic Attacks

Panic attacks are sudden, intense episodes of anxiety marked by physical and psychological symptoms such as rapid heartbeat, confusion, and shortness of breath. Although these episodes may appear abrupt, research shows they are often preceded by measurable neurological and physiological changes. This study presents a wearable EEG-based system that integrates AI-driven analysis for early detection of pre-attack patterns and delivers real-time, non-invasive intervention through transcranial direct current stimulation (tDCS). Preliminary validation using publicly available biosignal datasets demonstrates that the proposed detection pipeline can reliably identify early signatures associated with panic episodes, providing sufficient lead time for intervention. By combining continuous brain monitoring with automated neuromodulation in a closed-loop framework, this approach offers a practical path toward proactive and non-invasive support for individuals with anxiety disorders, aiming to improve emotional regulation and overall qual-ity of life.