DeepCor: Denoising fMRI Data with Contrastive Autoencoders (IMAGE)
Caption
For three decades, functional neuroimaging (fMRI) has been shaping the understanding of the human brain. A major obstacle for fMRI research is that information about brain responses is mixed with “noise”: distortions in the measurements caused by head movements of the participants, heart rate, and perturbations in fMRI machines. Removing noise more effectively could pave the way to new discoveries about the brain and its disorders. Boston College researchers have developed a new method to remove noise from fMRI data using generative AI. Their study found that the method can improve by 200 percent over previous approaches, offering the opportunity to measure brain responses more accurately in basic science as well as medical research.
Credit
Nature Methods
Usage Restrictions
Credit must be given to the creator. Only noncommercial uses of the work are permitted. No derivatives or adaptations of the work are permitted.
License
CC BY-NC-ND