Restoring Vision to Stroke Victims
Story from the Democrat and Chronicle

Patients who are partially blind after a stroke may be able to retrain their brain after doing months of eye exercises on a computer, according to a small study by University of Rochester researchers.

The study, published in the April 1 issue of the Journal of Neuroscience, found that people who did vigorous exercises every day for several months could coax the healthy regions of their brains to take in visual information, making up for other parts of the brain damaged by strokes.

Neuroscientist Dr. Krystel Huxlin, who led the study at the University of Rochester Eye Institute, said she was surprised by the research results because scientists had long believed that patients could not recover from the type of visual damage done by strokes. Many people are one-quarter to one-half blind after strokes, and though rehabilitation is common for relearning speech and movements, patients rarely receive visual training.

Huxlin studied four women and three men in their 30s to 80s, who had a stroke eight to 40 months before the experiment. All were partially blind and had substantial damage in the primary visual cortex of their brains, meaning their eyes could take in visual information that they could not consciously see.

During the study, they were told to "watch" a group of about 100 small dots move right or left across a computer screen. Though the patients were not aware they could see the dots, they would guess which direction the dots were moving and could improve their success rates over time when told when they've made a correct choice. Eventually, they became aware of the dots and their movements, as their brains relearned how to take in the visual information, the study found.

When they improved, researchers moved the dots deeper into their blind areas and could coax the brain to see more areas.

The five patients who completed the training had significantly improved vision, and several could drive again, shop and exercise after the experiment.

The University of Rochester has filed a patent on the technology.