Purchase this article with an account.
Sebastian Frank, Mark Greenlee, Peter Tse; Feature conjunction learning is an enduring form of visual learning. Journal of Vision 2016;16(12):544. doi: 10.1167/16.12.544.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We report evidence that feature conjunction learning represents an enduring form of visual learning. Participants were trained on a visual search task for feature conjunctions consisting of moving trajectories (Frank et al. 2015 Cerebral Cortex). The first and last training sessions were carried out during functional magnetic resonance imaging. After three weeks of training, participants improved remarkably in performance, indicative of learning. This improvement was differentially pronounced across trained retinotopic locations, revealing a unique perceptual learning pattern for each participant. As a result of training, activity in motion-sensitive area MT+ (V5) increased. Having demonstrated learning, we were interested in the long-term stability of these learning effects. Therefore, participants were retested in the learning task three months, one year, and two years after the end of training. Acquired behavioral improvements were remarkably stable over time and still present even two years after the end of original training. This was also true for each participant's retinotopic learning pattern, such that better performance during retests was achieved at retinotopic locations where participants had learned best. Activity changes during training also predicted later retest performance: the higher the change of activity in area MT+ during training, the better participants performed two years later. Our results show that feature-conjunction learning, once established, remains stable over years and represents an enduring form of visual learning.
Meeting abstract presented at VSS 2016
This PDF is available to Subscribers Only