Maryam Shanechi and her team at USC have created a new AI algorithm that can separate brain patterns linked to specific behaviors. This work, published in Nature Neuroscience, could improve brain-computer interfaces and reveal new brain patterns.
Our brains encode multiple behaviors simultaneously, like moving an arm or feeling hungry, making it hard to identify patterns linked to just one behavior.
This dissociation is important for creating brain-computer interfaces to help paralyzed patients move again. These devices decode movement thoughts from brain activity and use them to control things like robotic arms. Prof. Maryam Shanechi and her team, including Omid Sani, developed a new AI algorithm called DPAD to solve this challenge.
Shanechi said, “Our AI algorithm, DPAD, separates brain patterns linked to specific behaviors, like arm movement, from other brain activities. This improves movement decoding for brain-computer interfaces and helps discover new brain patterns.”
Sani added that the algorithm first learns the behavior-related patterns and then learns the rest, preventing confusion. Neural networks provide flexibility in detecting different brain patterns.
This algorithm can also be used in the future to decode mental states like pain or depression, helping treat mental health conditions. It could track patients’ symptoms and adjust treatments to fit their needs, leading to brain-computer interfaces for movement disorders and mental health conditions.
Journal reference :
- Sani, O.G., Pesaran, B. & Shanechi, M.M. Dissociative and prioritized modeling of behaviorally relevant neural dynamics using recurrent neural networks. Nature Neuroscience. DOI: 10.1038/s41593-024-01731-2