Purchase this article with an account.
Jonathan Folstein, Allen Newton, Ana Beth Van Gulick, Thomas Palmeri, Isabel Gauthier; Category learning causes long-term changes to similarity gradients in the ventral stream: A multivoxel pattern analysis at 7T. Journal of Vision 2012;12(9):1106. doi: 10.1167/12.9.1106.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Category learning causes acquired distinctiveness, selective enhancement of visual discriminability along category-relevant object dimensions. We previously used 3T fMRI adaptation to show that acquired distinctiveness measured behaviorally is accompanied by acquired distinctiveness in object-sensitive regions in the left ventral stream. Here, using high-resolution 7T fMRI, we use multivoxel pattern analysis (MVPA) to show long-term acquired distinctiveness in the ventral stream. Subjects were trained over several days to categorize objects created from a two-dimensional morphspace according to a particular category boundary. Immediate behavioral testing revealed a post-training advantage for discriminating stimulus pairs differing along the relevant dimension compared to the irrelevant dimension; this behavior advantage persisted for well over a week. A day or more after initial category learning, subjects were scanned at 7T while they performed a task requiring them to attend to stimulus location rather than stimulus shape. Voxel pattern similarity was measured in object-sensitive ROIs using a support vector machine (SVM) MVPA. The SVM was trained to classify stimulus-elicited voxel patterns according to the relevant category boundary learned by the subjects or the orthogonal, irrelevant, boundary. Probability of classifying patterns as the SVM’s target category decreased monotonically from the most ideal member to the most ideal non-member. In the left LOC, the similarity gradient was steeper when objects differed along the relevant compared to the irrelevant dimension, showing that category learning can selectively sensitize relevant aspects of shape variation in LOC. This study is the first to assess long-term changes to similarity gradients in the ventral stream along relevant vs. irrelevant object dimensions and the first to use MVPA at 7T to show neural signatures of acquired distinctiveness following category learning.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only