Israeli AI Startup Develops 15-Minute EEG Test to Diagnose Brain Disorders

The Israeli NeuroAI company Hemispheric is developing an artificial intelligence system called Descartes that analyzes brain activity through a brief EEG test lasting about 15 minutes. The non-invasive test involves wearing a sensor-equipped cap while performing tasks on a screen, with the AI model interpreting the electrical brain signals to assist in diagnosing neurological and psychiatric conditions.

Hemispheric aims for the system to help identify and assess disorders such as depression, anxiety, PTSD, ADHD, OCD, sleep disorders, Parkinson's, Alzheimer's, traumatic brain injury, and early cognitive decline. The AI model behind Descartes contains approximately six billion parameters and was trained on a large dataset of brain activity from around 100,000 individuals, totaling about 250,000 hours of recordings.

One pilot study involved 122 Israeli soldiers who served in Gaza, testing the system's ability to evaluate PTSD. The company plans to submit clinical trial results to the U.S. Food and Drug Administration (FDA) next year. Although the technology shows promise, it has not yet been clinically validated or approved for diagnostic use, and experts caution that large training datasets alone do not guarantee clinical reliability.

If successful in clinical and regulatory evaluations, the system could provide doctors with objective, quantitative brain activity data to complement current symptom-based diagnoses. The key challenge remains whether AI can accurately translate complex brain electrical signals into dependable medical information for clinical decision-making.