Abstract
Some types of Visual Perceptual Learning (VPL) are highly specific for the location of the stimulus in the visual field. For example, performance enhancement after training on the texture discrimination task (TDT, Karni & Sagi, 1991) is observed only at or close to the location in which a task target was trained. A number of human neuroimaging and monkey unit-recording studies have shown that VPL with such location specificity is associated with changes in the local region of the earliest visual area that retinotopically corresponds to the trained target location. Recently, it has been suggested that some subcortical areas code retinal locations. How does the location coding in the subcortical areas relate to the location specificity in VPL? To address this question, 11 young adults were trained on a TDT for 14 daily training sessions. To counterbalance trained visual fields a target was presented in either the upper right quadrant only (N=6) or the upper left quadrant only (N=5) during training. To measure brain activation during the time course of training, subjects were also asked to perform a TDT during the brain activation measurements at 4 different stages; pre-training, and after the 1st, 6th and 14th training sessions. To examine a location specific effect, a target was presented either at the trained location or an untrained location in each trial. As in Karni & Sagi (1991), performance improved only at the trained location. BOLD signal significantly changed from the pre- training scans to post-training scans only in the ipsilateral part of the caudate to the trained retinal location. When the target was presented at an untrained location, BOLD activation did not significantly change in the caudate. These results indicate that part of the caudate changes in association with VPL and may play a role in the location specificity of VPL.
Meeting abstract presented at VSS 2013