Abstract
Visual perceptual training can decrease the size of visual field defects measured with Humphrey perimetry in cortically blind (CB) subjects (Huxlin, et al., 2009; Das et al., 2014). Here, we investigated the spatial relationship between where subjects trained to discriminate motion direction or orientation and perimetrically-measured visual field changes. Sixteen CB subjects underwent monocular visual field mapping using 24-2 and 10-2 Humphrey perimetry with controlled fixation. Ten subjects trained to identify orientation of static gabors and/or global motion direction of random dot stimuli. Following recovery of trained abilities at least at one blind field location, Humphrey fields were re-measured. Six control CB subjects waited one to seven months without training before a second set of visual fields was collected. Composite fields were created by convolving luminance sensitivity between 24-2 and 10-2 fields in each eye, interpolating between measured locations and averaging the interpolated fields between the two eyes. Subtraction analysis of composite fields was used to create difference maps for changes in luminance sensitivity between visual fields. Blind field training significantly improved discrimination performance and luminance detection at trained locations. However, improvements in luminance detection as large as 20-30dB also occurred at untrained locations along the blind field border. 93.5% of recovery occurred within 5 degrees of the blind field border. Significant improvements in Humphrey fields were not attained in untrained subjects. Thus, visual discrimination training in CB fields causes improvements that generalize to untrained locations along the blind field border, suggesting a greater degree of plasticity in this region.