In a notable development in the field of neuroscience, an Indian research team has introduced MANAS 1, an innovative brain language foundation model designed to facilitate early detection of neurological and psychiatric conditions. The model leverages 60,000 hours of brainwave recordings collected from over 25,000 patients. Its debut was made at an AI summit, and it has been made available as open source on Hugging Face. This initiative is part of the Indian AI Mission, supported by the Ministry of Electronics and Information Technology.

Development and Features

MANAS 1 has been developed by Intellihealth (NeuroDx) under the direction of Dr. Puneet Agarwal, a neurologist and former professor at the All India Institutes of Medical Sciences. This AI model is distinct from traditional artificial intelligence systems as it is specifically trained to interpret electroencephalography (EEG) signals, which represent the electrical activity of the brain. Comprising 400 million parameters, MANAS 1 serves as a foundational platform for creating specialized AI tools aimed at diagnosing various brain disorders.

Understanding Brain Signals

Dr. Agarwal explained that MANAS 1 is engineered to “comprehend the fundamental language of the brain.” This foundational model is akin to other models like ChatGPT in that it learns from extensive EEG data, enabling interpretation of brain signals in ways that existing diagnostic tools, such as MRI scans, may not fully accomplish. By offering a framework for future developments, the model aims to facilitate the creation of targeted AI tools for conditions like epilepsy and dementia, and to assist in exploring facets of brain functionality that remain ambiguous.

Implications for Public Health

Addressing public health concerns, the developers of MANAS 1 highlight the need for early diagnosis in a country where a shortage of neurologists and psychiatrists is prevalent, particularly in rural areas. Oftentimes, brain disorders are diagnosed in later stages, resulting in increased disability and prolonged treatment costs. The tools developed using MANAS 1 could support healthcare professionals at community health centers, Ayushman Arogya Mandirs, and district hospitals, aiding in initial screenings and timely referrals. Any disease-specific AI tools created from this foundational model would need to undergo regulatory evaluation before they can be implemented in clinical settings.

Future Developments

If scaled successfully, MANAS 1 has the potential to bridge the gap between the onset of symptoms and the confirmation of diagnoses, especially in critical conditions such as epilepsy and dementia. Looking ahead, a next-generation version, dubbed MANAS 2, is expected to be released within weeks. As advancements in artificial intelligence penetrate the realm of neuroscience, the introduction of MANAS 1 marks a significant shift towards decoding the brain’s electrical language, offering considerable promise for research, diagnosis, and healthcare accessibility.

Long or Short, get news the way you like. No ads. No redirections. Download Newspin and Stay Alert, The CSR Journal Mobile app, for fast, crisp, clean updates!

App Store –  https://apps.apple.com/in/app/newspin/id6746449540 

Google Play Store – https://play.google.com/store/apps/details?id=com.inventifweb.newspin&pcampaignid=web_share