Researchers from the National University of Singapore have developed a new analytical tool offering a new approach to analysing brain activity. 

The Brain-JEPA (Joint-Embedding Predictive Architecture) model identifies brain regions and how they function individually and dynamically with other regions. It also breaks down complex patterns of brain activity into smaller pieces so they can be analysed more efficiently.

HOW IT WORKS

Developed using AI and volumes of brain records, the model has two novel features: brain gradient positioning, which maps how different brain regions function together, and spatiotemporal masking, which focuses on critical patterns in MRI signals.  

According to the NUS Medicine researchers, Brain-JEPA does away from previous approaches that reconstruct raw brain signals. Instead, the model learns the abstract patterns in brain activity and predict demographics, understand personality traits, and diagnose brain disorders across different ethnic groups. 

Dong Zijian, one of the researchers, claims their model is "faster, smarter, and works well across diverse populations." "Unlike previous models, it works well on people from different ethnic backgrounds, which makes it a powerful tool for global research." 

The research was recently presented at the Conference on Neural Information Processing Systems, an annual interdisciplinary academic conference on machine learning. 

WHY IT MATTERS

Explaining its benefits to healthcare, Helen Zhou, associate professor and director of NUS Medicine Centre for Translational Magnetic Resonance Research, said Brain-JEPA "enables earlier and more accurate diagnoses of brain disorders, better disease progression predictions, and personalised treatment plans, while being globally applicable across diverse ethnic groups."

"Brain-JEPA could also help doctors diagnose conditions like Alzheimer’s disease earlier and more accurately, and also offer insights into how the brain works, which could lead to treatments for mental health and neurological disorders," added Dr Li Ruilin, part of the research team. 

Researchers are now working to establish fundamental rules about how such models as Brain-JEPA behave regarding dataset size, model size, and other aspects. They also intend to work on guidelines for developing next-generation brain AI models. 

THE LARGER TREND

Large language models have also been applied to research advancing the diagnosis of neurodegenerative diseases. For example, in South Korea last year, researchers from the government-backed Electronics and Telecommunications Research Institute introduced an LLM-based predictive tool for screening mild cognitive impairment (MCI) and identifying Alzheimer's disease.

Prior to the LLM boom two years ago, Fujifilm had already utilised AI to develop a similar technology to predict patients' progression from MCI to Alzheimer's in two years.