Innovational Brain Language Model (BrainLM) Uses Generative AI to Unravel Brain Behaviour and Neurological Diseases
The new AI system BrainLM, developed by researchers in neuroscience, interprets and maps brain activity and speeds up its analysis
A team of researchers from Baylor College of Medicine and Yale University joined forces to create a revolutionary AI system called BrainLM.
The system integrates the power of generative artificial intelligence (AI) to develop the foundational model. BrainLM uses technology with similar ideas harnessed in large language models (LLMS), such as ChatGPT and DALL-E. However, the new model is trained to learn from functional brain images rather than from text and images. It is created to model the brain in silico (simulator of brain activity responses to perturbations) to understand brain behaviour and neurological diseases.
BrainLM is more effective than other tools in predicting neurological illnesses such as PTSD and depression and reducing the expense of clinical trials through the identification of candidates for whom new treatments may be more beneficial.
This foundational model is trained on 6,700 hours of functional magnetic resonance imaging (fMRI) from 40,000 participants. According to research that was published as a conference paper at ICLR 2024, the model is able to identify “functional networks and generates interpretable latent representations of neural activity”.
BrainLM is innovative as it can automate fMRI data which is cost-effective and allows researchers to carry out more extensive studies incorporating data from several experiments.
Dr. Chadi Abdallah, associate professor in the Menninger Department of Psychiatry and Behavioral Sciences at Baylor and co-corresponding author of the paper stated how it is commonly known that brain activity is connected to a person’s behaviour and several conditions such as Parkinson’s and seizures.
Therefore: “Functional brain imaging or functional MRIs allow us to look at brain activity throughout the brain, but we previously couldn’t fully capture the dynamic of these activities in time and space using traditional data analytical tools.
“More recently, people started using machine learning to capture the brain complexity and how it relates to specific illnesses, but that turned out to require enrolling and fully examining thousands of patients with a particular behavior or illness, a very expensive process.”
It merely requires brain activity to teach the computer and AI model the evolution of brain activity over time.
According to Dr. Abdallah: “traditional analytical tools have often fallen short in capturing the full spectrum of brain activity in both its temporal and spatial dimensions.”
BrainLM allows researchers to fine-tune a task and pose queries in further studies.
Dr. Abdallah explains how: “If you want to do a clinical trial to develop a medication for depression, for example, it could cost hundreds of millions of dollars because you need to enroll a large number of patients and treat them for a long time.
“With the power of BrainLM, we can potentially cut this cost in half by enrolling only half the subjects using the power of BrainLM to select the individuals most like to benefit from a treatment. So, BrainLM can apply the knowledge learned from the 80,000 scans to apply it to those specific study subjects”.
The future use of BrainLM will involve making predictions relating to neurological diseases.
The research was funded by Yale’s Wu Tsai Institute and Baylor’s Beth K. and Stuart Yudofsky Chair in the Neuropsychiatry of Military Post Traumatic Stress Syndrome.
The researchers involved in the groundbreaking project include: Josue Ortega Caro, Antonio Henrique de Oliveira Fonseca, Syed A Rizvi, Matteo Rosati, Christopher Averill, James L Cross, Prateek Mittal, Emanuele Zappala, Rahul Madhav Dhodapkar and David van Dijk who are affiliated with Baylor College of Medicine, Yale University, University of Southern California and Idaho State University.
