Background and objective: Visual restoration therapy is a home-based treatment program intended to expand visual fields of hemianopic patients through repetitive stimulation of the borderzone adjacent to the blind field. We hypothesized that the training itself would induce visual field location-specific changes in the brain's response to stimuli, a phenomenon demonstrated in animal experiments but never in humans with brain injury.
Methods: Six chronic right hemianopic patients underwent functional magnetic resonance imaging (fMRI)--responding to stimuli in the trained visual borderzone versus the nontrained seeing field before and after 1 month of visual restoration therapy. Spatially normalized fMRI time-series data were analyzed in a fixed-effects group analysis comparing blood oxygen level dependent (BOLD) activity in the borderzone versus seeing location at baseline and at 1 month. Percent BOLD change was measured to determine each condition's contribution to the time-by-condition interaction.
Results: There was a significant time by condition interaction manifested as increased BOLD activity for borderzone detection relative to seeing detection after the first month of therapy, which correlated with a relative improvement in response times in the borderzone location out-of-scanner. The right inferior and lateral temporal, right dorsolateral frontal, bilateral anterior cingulate, and bilateral basal ganglia showed the greatest response.
Conclusion: Visual restoration therapy appears to induce an alteration in brain activity associated with a shift of attention from the nontrained seeing field to the trained borderzone. The effect appears to be mediated by the anterior cingulate and dorsolateral frontal cortex in conjunction with other higher order visual areas in the occipitotemporal and middle temporal regions. Demonstration of a visual field-specific training effect on brain activity provides an important starting point for understanding the potential for visual therapy in hemianopia.