Perceptual relearning of complex visual motion after V1 damage in humans

Abstract

Damage to the adult, primary visual cortex (V1) causes severe visual impairment that was previously thought to be permanent, yet several visual pathways survive V1 damage, mediating residual, often unconscious functions known as "blindsight." Because some of these pathways normally mediate complex visual motion perception, we asked whether specific training in the blind field could improve not just simple but also complex visual motion discriminations in humans with long-standing V1 damage. Global direction discrimination training was administered to the blind field of five adults with unilateral cortical blindness. Training returned direction integration thresholds to normal at the trained locations. Although retinotopically localized to trained locations, training effects transferred to multiple stimulus and task conditions, improving the detection of luminance increments, contrast sensitivity for drifting gratings, and the extraction of motion signal from noise. Thus, perceptual relearning of complex visual motion processing is possible without an intact V1 but only when specific training is administered in the blind field. These findings indicate a much greater capacity for adult visual plasticity after V1 damage than previously thought. Most likely, basic mechanisms of visual learning must operate quite effectively in extrastriate visual cortex, providing new hope and direction for the development of principled rehabilitation strategies to treat visual deficits resulting from permanent visual cortical damage.

Figure 1.

Training-induced improvements in direction range thresholds in the blind fields of VC1–VC5. A, Size and location of random-dot stimuli on which subjects were first trained in their blind field indicated on composite Humphrey visual field maps (for grayscale, see supplemental Fig. 1, available at www.jneurosci.org as supplemental material). Circles illustrate the sizes and locations of random-dot stimuli used to measure performance at equivalent locations in the intact and blind hemifields. B, Scatter plots of direction range (DR) threshold (thresh.) performance as a function of the number (#) of training sessions administered at blind field locations in A. Black dots represent data collected by the subjects during in-home training. Red dots represent data collected from the same subjects in the laboratory, where central fixation was controlled using an integrated ISCAN system. Note the close correlation between the subjects' in-home performance and that recorded in the laboratory. Note also the variability in the number of training sessions, each consisting of 300 trials, required for different subjects' direction range thresholds to reach and stabilize within the normal range. This normal range was represented by the mean (gray horizontal line) and SD of the mean (gray shaded area around the gray line) direction range thresholds obtained at the “control” locations in the intact hemifield of each subject (black circles in A). Deg., Degree.

Figure 2.

Global motion discrimination training improves contrast sensitivity and the extraction of motion signal from noise. A, Histogram illustrating how global direction discrimination training returns direction range (DR) thresholds to normal (relative to those in the intact hemifield; p = 0.127, paired Student's t test) at the retrained blind field locations in VC1–VC5. B, Sample psychometric function for VC2 performing a left–right, global direction discrimination task in her blind field (see location circled in the Humphrey visual field map inset). Before training (open circles), percentage correct performance hovered just above chance at the lowest direction range levels. After training (filled circles), performance improved dramatically and a normal psychometric function was attained for this task. C, Histogram illustrating the positive effect of global direction discrimination training on motion signal (MS) thresholds in the blind field of VC1–VC5. Training on random-dot stimuli with a large range of dot directions (not motion signal) recovered the subjects' ability to extract motion signal from random directional noise at the trained locations. After training, motion signal thresholds were not significantly different (p = 0.759, paired Student's t test) from those in the intact hemified (white column). D, Sample psychometric function for VC2 discriminating the left–right direction of motion of random-dot stimuli in her blind field (see location circled in the Humphrey visual field map inset in B). The percentage of noise dots was varied throughout the session, but, before training (open circles), only stimuli with 100% signal dots could be discriminated at a 75% correct level. After training on direction range thresholds, however, performance and the psychometric function returned to normal (filled circles). E, Histogram illustrating the effect global direction discrimination training using random dots (of set contrast) on contrast sensitivity for left–right direction discrimination of luminance-modulated, drifting sine-wave gratings (spatial frequency, 0.5 or 1 cycle°; temporal frequency, 10 Hz). Training with random dots significantly improved contrast sensitivity at the retrained, blind field locations (p = 0.405, paired Student's t test, relative to equivalent locations in the intact hemifields). F, Plot of contrast sensitivity versus spatial frequency for VC2 before (open circles) and after (filled circles) training with random-dot stimuli at the location indicated in B. Temporal frequency was held at 10 Hz. Global direction discrimination training significantly improved contrast sensitivity at the trained, blind field location, with the greatest improvements occurring between 0.5 and 2 cycles°. All data in the histograms are expressed as means and SEM. *p < 0.05, paired Student's t tests relative to intact hemifield values. Deg., Degree; freq., frequency.

Figure 3.

Retinotopic specificity of training-induced improvements in direction range thresholds. A, Visual field maps for VC2 and VC3, illustrating the locations where visual training was performed. The shade of gray of circles in the visual fields match the shading of data points in B and C. B, Plots of direction range (DR) threshold versus the number (#) of training session at blind field locations in A. Once recovery of DR thresholds was attained at a given blind field location, moving the stimulus to a different location within the blind field, even one that was only 2° away, caused DR thresholds to fall to 0°. The process of retraining then had to be restarted anew. C, This contrasts with learning rates in the intact hemifields of the same subjects, in which DR thresholds ranged between 240 and 305° initially and stabilized to 330–340° with just a few days of training. However, this small improvement appeared to transfer very effectively when the stimulus was moved 2° deeper into the intact hemifield. Solid lines in the DR threshold graphs indicate moving averages with periods of four training sessions. Deg., Degree.

Figure 4.

Global motion discrimination training decreases the size of Humphrey visual field defects. A, Composite Humphrey visual maps obtained in VC2 and VC3 after global direction discrimination training at locations denoted by gray circles and, in VC7, more than 1 year after his initial examination. The solid red lines outline the 5 dB contour line after training or during visit 2 for VC7. Note that for VC2 and VC3 (but not for VC7), this border had receded significantly (shaded red area) from the 5 dB pretraining border (dotted red line), although the posttraining Humphrey maps were obtained 30 or more months after stroke. Color conventions and scaling for the composite maps are as in Figure 1 and supplemental Figure S1 (available at www.jneurosci.org as supplemental material). Enlargements of the visible field always overlapped with retrained locations. B, Cluster analysis of Humphrey visual fields for VC2, VC3, and VC7, illustrating the change in mean pattern deviation for different field locations over time. The left graph in each pair color codes the visual field locations assigned to each cluster. The right graph then illustrates the mean pattern deviation (in decibels) for each cluster across time. Light gray (Cluster 1) represents visual field locations within the normal range of visual sensitivity; black (Cluster 2) represents areas of lost function in which luminance sensitivity did not change significantly over time. Red (Cluster 3) represents a third pattern of change across time for each participant. It can be seen that, in the case of VC2 and VC3, this third pattern was an improvement from pretraining to posttraining time points. In the case of patient VC7, this third pattern represents a significantly smaller change between the two visual field tests administered than was observed in patients who underwent training (such as VC2 and VC3). Deg., Degree.