Visual and neural plasticity: A study of line orientation discrimination

Abstract

Discrimination learning, the ability to better discriminate fine perceptual changes, is an important skill humans possess. In the present set of studies, some of the parameters that govern line orientation discrimination teaming are examined along with theories of the mechanisms underlying such learning. Specifically, three questions are addressed. First, how does training on one part of the orientation dimension affect the rest of that dimension? When subjects were trained with one pair of orientations at one retinal location, learning that was a true change in perception did occur. When subjects were tested with a novel orientation, rotated 90 degrees at the same retinal location, significant interference was found. Next, how training at one part of the orientation dimension affects the retinal location dimension was examined. Retinal location is coded in the same part of the brain as orientation and therefore may also be affected by training. Training with one pair of orientations at one retinal location did have an effect on at least one neighboring retinal location such that performance dropped below baseline, indicating interference. Finally, how training on the entire dimension of orientation affects the process of discrimination learning was examined. Training subjects with four pairs of orientations that span the entire orientation dimension did allow learning to occur. This learning may not have been a true change in perception. When tested with a novel, intermediate pair of orientations, previous learning may have positively transferred, but did not appear to interfere. Testing with a neighboring retinal location showed no sign of interference as did training with one pair of orientations. Based on the neurophysiological studies on monkeys by Recanzone et al. (1992; 1993), it was suggested that increased discrimination ability occurs through the reorganization of cortical sensory maps such that there is more cortical area devoted to the practiced stimuli. The results of the present set of studies suggest that discrimination learning does occur by recruiting neighboring cells to respond to the practiced stimuli when learning is occurring for one part of a dimension. When learning occurs for an entire dimension, a different mechanism appears to be at work.