Functional Magnetic Resonance Imaging
Functional Magnetic Resonance Imaging (fMRI) studies brain activity by measuring blood oxygenation levels, aiming to understand brain function and its relation to cognition and behavior. Current research heavily utilizes deep learning, including transformer networks, autoencoders, and diffusion models, to analyze high-dimensional fMRI data, improve spatial and temporal resolution, and decode cognitive states or even reconstruct visual imagery from brain activity. These advancements are improving diagnostic accuracy for neurological disorders like autism and Alzheimer's disease, and enabling novel applications such as personalized brain-computer interfaces and the development of more brain-like artificial intelligence models.
Papers
Bayesian temporal biclustering with applications to multi-subject neuroscience studies
Federica Zoe Ricci, Erik B. Sudderth, Jaylen Lee, Megan A. K. Peters, Marina Vannucci, Michele Guindani
BrainMAE: A Region-aware Self-supervised Learning Framework for Brain Signals
Yifan Yang, Yutong Mao, Xufu Liu, Xiao Liu