Abstract
The neural underpinning of discrimination training remains highly controversial. While physiological studies found little, if any, evidence for the training effect in sensory cortex (Law & Gold, 2008), such effects were pronounced in fMRI studies. The apparent difference in results between the two approaches could be due to the sluggish nature of the fMRI signals, which confounded perception with other cognitive processes involved in the discrimination task, such as stimulus comparison. Here, we adopted a long-delay orientation discrimination task to decouple the perceptual encoding of stimulus from the process of stimulus comparison. We trained participants with a short-delay (ISI, 0.6 s) orientation discrimination task for six days (peripheral, 6.5°, at the top left / lower right corner; orientation, 45° / 155° ), whereas testing the training effect with the long-delay orientation discrimination task (ISI, 11.8 s, along both the trained and untrained orientation) in the scanner before, after, and ~10 days after the training. Participants' performance improved gradually over the time courses of training. The training effect preserved when measured with the long-delay discrimination task both immediately and 10 days after the training, and was accompanied by an enhanced brain activity specific to the trained orientation in the primary visual cortex. Importantly, the enhanced brain activation was only evident during the processing of the second stimulus in the trial sequence and disappeared ten days after the training. These results suggest that the pronounced changes of brain activity in the primary visual cortex may be related to the process of stimulus comparison, rather than the perceptual encoding of the stimulus.
Meeting abstract presented at VSS 2017