文摘
MVPA searchlights revealed regional patterns of neural activity associated with compensatory naming processing in aphasia, which was not detected by conventional univariate analyses. The largest and most robust cluster was found in right occipitotemporal cortex, which is mirrored by predominant semantic error rates. This is one of the first attempts of employing machine-learning techniques to predict behavioral naming performance using on-line fMRI data in aphasia studies.