Abstract
Perceptual learning (PL) can transfer to a new location/hemisphere completely when the new location is additionally trained with an irrelevant task (Xiao et al, 2008). This double-training result, along with additional findings of learning transfer to untrained orientations/directions (Zhang et al., 2010; Zhang & Yang, 2014) suggests that PL is a high-level learning process occurring beyond the retinotopic visual cortex. However, it is unclear why PL is location specific in the first place, and whether bottom-up stimulation of the untrained location during irrelevant training is sufficient to enable learning transfer. In a failed attempt to solve these puzzles, we found that practicing Vernier at one retinal location, with simultaneous passive exposure to additional Vernier stimuli at a diagonal quadrant location, leads to no learning transfer (Wang et al., 2013). However, the impact of passive exposure stimulation might have been suppressed by demanding Vernier practice. Here we used a continuous flashing suppression paradigm to re-investigate the role of location training in double-training enabled learning transfer. Vernier was trained at one quadrant location, and a randomly oriented Gabor was simultaneously shown at a diagonal location. The perception of the Gabor was suppressed by flashing white noise shown in the other eye, and the observers were unaware of the Gabor’s presence. This bottom-up stimulation of untrained location resulted in complete transfer of Vernier learning from the trained location. In a control experiment in which the randomly oriented Gabor was absent while other conditions were unchanged, no learning transfer was evident. These results suggest that bottom-up stimulation of an untrained retinal location can enable transfer of Vernier learning from a trained location. Moreover, location specificity, at least with Vernier learning, may result from under-stimulation of untrained locations. Whether this under-stimulation results from no stimulation or even suppression during training requires further investigation.
Meeting abstract presented at VSS 2015