Decoding brain activity using a large-scale probabilistic functional-anatomical atlas of human cognition

Timothy N. Rubin, Oluwasanmi Koyejo, Krzysztof J. Gorgolewski, Michael N. Jones, Russell A. Poldrack, Tal Yarkoni, Samuel J. Gershman
2017 PLoS Computational Biology  
A central goal of cognitive neuroscience is to decode human brain activityi.e., to infer mental processes from observed patterns of wholebrain activation. Previous decoding efforts have focused on classifying brain activity into a small set of discrete cognitive states. To attain maximal utility, a decoding framework must be openended, systematic, and contextsensitivei.e., capable of interpreting numerous brain states, presented in arbitrary combinations, in light of prior information. Here we
more » ... ake steps towards this objective by introducing a Bayesian decoding framework based on a novel topic modelGeneralized Correspondence Latent Dirichlet Allocationthat learns latent topics from a database of over 11,000 published fMRI studies. The model produces highly interpretable, spatiallycircumscribed topics that enable flexible decoding of wholebrain images. Importantly, the Bayesian nature of the model allows one to "seed" decoder priors with arbitrary images and textenabling researchers, for the first time, to generative quantitative, contextsensitive interpretations of wholebrain patterns of brain activity.
doi:10.1371/journal.pcbi.1005649 pmid:29059185 pmcid:PMC5683652 fatcat:lb43b2o72jehrahuzifvkuysri