Purchase this article with an account.
Ke Jia, Elisa Zamboni, Nuno Reis Goncalves, Catarina Rua, Valentin Kemper, Guy Williams, Chris Rodgers, Zoe Kourtzi; Ultra-high field imaging of perceptual learning in the human visual cortex. Journal of Vision 2019;19(10):186b. doi: https://doi.org/10.1167/19.10.186b.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Training is known to improve perceptual judgments; yet the neural underpinnings of learning-dependent plasticity remain controversial. Previous physiological studies found little evidence of learning-dependent plasticity in the visual cortex, whereas fMRI studies have reported more pronounced changes in functional brain activity due to training. Recent advances in brain imaging technology (i.e. ultra-high field 7T imaging) afford us with higher resolution to examine learning-dependent brain plasticity at the finer scale of laminar layers in the human visual cortex. Here, we tested whether training results in brain activity changes in input compared to deeper layers, consistent with learning-dependent changes in local vs. feedback processing, respectively. We trained participants on an orientation discrimination task over five consecutive days. Both before and after the training phase, we measured participants’ behavioral performance along three different orientations (55°, 125°, or 0°) at two different locations (left or right to the fixation) and their functional brain activity across visual cortex layers. The trained orientation and location remained the same across training sessions and were counterbalanced across participants. Participants’ behavioral performance improved significantly over the course of training. Analysis of the fMRI data did not show significant differences in overall BOLD response before vs. after training. However, multivoxel pattern analysis (MVPA) showed enhanced decoding accuracy for the trained orientation after than before training. These learning-dependent changes were specific to middle rather than deeper or superficial layers in primary visual cortex. Further, using forward encoding modeling (FEM), we found enhanced channel response selective to the trained orientation after training. These learning-dependent changes in both behavioral improvement and multivoxel activity patterns (MVPA and FEM) were specific to the trained orientation and location. These results suggest that training to discriminate fine orientation differences relates to enhanced local information processing in the primary visual cortex.
This PDF is available to Subscribers Only