Brain Scans Let Computer 'Read' Visual Activity

Mar. 24 -- WEDNESDAY, March 5 (HealthDay News) -- Using brain scans and computer models, researchers report they have found a way to "read" visual activity in the brain.

The process relies on functional MRI to scan the brain for activity information related to the viewing of a chosen set of images. A computer database of brain activity-image links is then created, so that future viewings can be deduced based solely on an analysis of fresh fMRI patterns.

"We're not mind-reading," explained study co-author Jack L. Gallant, an associate professor in the department of psychology at the Helen Wills Neuroscience Institute at the University of California, Berkeley. "We're not reconstructing images of what people see or think. We can't do that yet, although it should be possible in principle."

"But already," he added, "this research makes clear that there's a huge amount of information -- way more than we have expected -- that we can dig out of fMRI signals to get a better understanding of brain function. And that is very important, both in terms of pure science and in terms of how this information might eventually lead to all kinds of applications in the future."

Gallant and his colleagues reported their findings in the March 5 online issue of Nature.

To probe the possibilities of brain imaging, Gallant enlisted two of his study co-authors -- Kendrick N. Kay and Thomas Naselaris -- to serve as healthy volunteers with good eyesight.

As a first step, both were shown 1,750 photographic images of animals, buildings, food, indoor scenes, outdoor scenes and people, during which fMRIs recorded activity in the primary visual cortex region of their brains.

The authors noted that fMRIs measure blood flow related to neural activity in the brain, and that the particular region observed is the brain's largest processing module.

The brain activity was then put into a computer program. During a second round, Kay and Naselaris were then shown 120 different photos. The computer model sifted through its previous store of brain activity-to-image patterns to "decode" the second round of fMRI data and find a correct match.