23 Feb 2026 • A new brain language foundation model, MANAS-1, trained on 60,000 hours of brainwave recordings from more than 25,000 patients, aims to detect neurological and psychiatric disorders earlier by interpreting electroencephalography (EEG) signals.

Developed under India’s AI Mission by Intellihealth (NeuroDx) and led by neurologist Dr. Puneet Agarwal, the model uses 400 million parameters to analyze electrical brain activity, offering insights that conventional imaging tools like MRI may miss. Unlike traditional AI systems, MANAS-1 is designed as a foundational platform that can support disease-specific diagnostic tools, particularly for conditions such as epilepsy and dementia.

The initiative also addresses a major public health gap: delayed diagnosis of brain disorders in India, driven by limited specialist access, especially in rural regions. Developers say AI tools built on MANAS-1 could assist frontline providers at community health centers and district hospitals in screening patients and enabling earlier referrals, though regulatory approval will be required before clinical use.

With a next-generation version already in development, MANAS-1 signals a growing push to integrate AI into neuroscience, potentially reshaping early detection, expanding access to care, and accelerating efforts to decode the brain’s electrical activity.

Source: The CRS Journal | Read full story