Indian Team Releases MANAS 1 EEG Foundation Model
An Indian team led by Dr Puneet Agarwal unveiled MANAS 1, an open-source 400-million-parameter EEG foundation model trained on 60,000 hours from over 25,000 patients, released during a recent AI summit and hosted on Hugging Face. Backed by the Indian AI Mission, the model aims to decode brain electrical signals for earlier detection of epilepsy, dementia, and psychiatric disorders, and to enable disease-specific tools requiring regulatory approval.
Key Points
- 1Introduces MANAS 1, a 400M-parameter EEG foundation model trained on 60,000 hours from 25,000+ patients
- 2Addresses early detection gaps by decoding EEG signals MRI cannot, potentially improving neurological and psychiatric diagnosis
- 3Enables clinicians and researchers to build disease-specific screening tools for referrals in low-resource healthcare settings
Scoring Rationale
Strong novelty and practical applicability raise score, tempered by limited clinical validation and early-stage reporting.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
