eduzhai > Applied Sciences > Engineering >

Time-Resolved fMRI Shared Response Model using Gaussian Process Factor Analysis

  • Save

... pages left unread,continue reading

Document pages: 15 pages

Abstract: Multi-subject fMRI studies are challenging due to the high variability ofboth brain anatomy and functional brain topographies across participants. Aneffective way of aggregating multi-subject fMRI data is to extract a sharedrepresentation that filters out unwanted variability among subjects. Somerecent work has implemented probabilistic models to extract a sharedrepresentation in task fMRI. In the present work, we improve upon these modelsby incorporating temporal information in the common latent structures. Weintroduce a new model, Shared Gaussian Process Factor Analysis (S-GPFA), thatdiscovers shared latent trajectories and subject-specific functionaltopographies, while modelling temporal correlation in fMRI data. We demonstratethe efficacy of our model in revealing ground truth latent structures usingsimulated data, and replicate experimental performance of time-segment matchingand inter-subject similarity on the publicly available Raider and Sherlockdatasets. We further test the utility of our model by analyzing its learnedmodel parameters in the large multi-site SPINS dataset, on a social cognitiontask from participants with and without schizophrenia.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...