BOLD Signal Changes Associated with Orientation Learning and Transfer at Different Brain Areas

Ding-zhi Hu,Xin-yu Xie,Cong Yu,Peng Zhang
DOI: https://doi.org/10.1167/jov.20.11.1749
2020-01-01
Journal of Vision
Abstract:Perceptual learning research often targets learning specificities as means to understand the underlying mechanisms. However, learning specificities diminish with double training (Xiao_et_al_2008), and learning of basic visual features like orientation is not much location specific (Zhang_et_al_2010). Here we used fMRI to investigate the brain mechanisms underlying orientation learning and transfer. Twenty-three participants practiced orientation discrimination at the lower-left or -right quadrant at 135o or 45o. Five sessions of training improved orientation discrimination by 43.2±2.5% for the trained orientation at the trained quadrant, and 32.5±3.4% for the orthogonal orientation at the untrained quadrant, with 75% learning transfer. Subjects participated fMRI experiments before and after training. After training, BOLD responses in V1-V3 increased only for the trained stimulus. But BOLD responses in frontoparietal areas decreased for both trained and untrained stimulus, especially more for the untrained stimulus. Importantly, the amount of learning was negatively correlated with response changes to the trained stimulus in V1 and frontoparietal cortex, while the amount of learning transfer were positively correlated with a transfer index (trained-untrained) of BOLD response in V1, IFJ, IPS and PFC. Moreover, the task-based functional connectivity between frontoparietal cortex (IFJ/IPS/PFC) and V3 became weaker for the trained stimulus, but the connectivity between frontoparietal cortex and V2 became stronger for the untrained stimulus. The negative correlation between learning and BOLD signal change, and the decrease of functional connectivity between frontoparietal areas and early visual cortex for the trained stimulus, suggest that orientation learning may lead to a more efficient decision making in fronto-parietal areas. But the positive correlations between the transfer index of BOLD changes and perceptual learning transfer, and the increased functional connectivity between frontoparietal areas and early visual cortex for the untrained stimulus, suggest that learning transfer may require more engagement of frontoparietal areas and their top-down modulation of early visual activities.
What problem does this paper attempt to address?