Jonathan Cohen and his colleagues at the University of Pittsburgh and Carnegie Mellon University are wor king on it. Their research is helping to shape a newly evolving discipline called cognitive neuroscience. In recent work, they have exploited technology developed during the 1980s that many scientists believe will revolutionize study of the brain. Imaging technology, such as magnetic resonance imaging (MRI) and other techniques, combined with computing power makes it possible, in effect, to peel away the bone and membrane surrounding the brain. Without even touching their human subject, researchers can se e what happens inside a living, thinking brain, and they can identify what parts of this intricate, complexly folded, interconnected mass of tissue "light up" during mental activities.
Cohen is codirector of the Laboratory for Clinical Cognitive Neuroscience, a joint venture between Pitt and CMU. Last year, with colleagues Douglas Noll and Walter Schneider, he used a technique known as functional MRI to record a view of the functioning brain that is among the most detailed ever reported. While other brain-mapping techniques give what resembles a satellite view of the world, in which cities can be seen and identified, the Pitt/CMU researchers can see streets. With functional MRI, they ca n map the sites of brain activity to a resolution as fine as one millimeter, comparable to mapping a football field in six-inch units.
Using resources at the Pittsburgh Supercomputing Center (PSC), Cohen and colleagues have reduced image-processing time for a si ngle experiment from a day to an hour. In the future, the researchers plan to further expedite their work by transferring more of their data processing to PSC, and they expect that, ultimately, supercomputing will make it possible to use functional MRI as a real-time clinical tool.
The functional MRI experiments conducted by Cohen investigate a concept known in cognitive psychology as working memory. Each subject's brain is scanned while they perform a working memory task and a control task. In the control task, the subject sees a r andom sequence of letters one at a time on a visual display. They are instructed to press a button whenever the letter "X" shows on the display. In the memory task, subjects see a similar sequence of letters, but they are instructed to press the button on ly when a letter repeats after exactly one intervening letter. For example, A-F-A should prompt a response, but not A-A or A-Q-G-A.
Both tasks, explains Cohen, require subjects to visually monitor sequences of letters presented one at a time, to evaluate their identity and respond by pressing a button. The memory task, however, requires in addition that the subject keep in mind both t he identity and order of the two previous letters and continuously update this mental record as the sequence progresses.
The MRI machine records data from six slice locations in the prefrontal cortex of each subject (panel A). A set of activation images for one subject (panel B) shows the brain areas significantly activated during the memory task and not during the control task. Results to date from these studies, says Cohen, "support the idea that the prefrontal cortex becomes engaged when recently presented information must be represented and actively maintained to perform a task."
Researchers: Jonathan D. Cohen, Carnegie Mellon University & University of Pittsburgh, Walter Schneider, University of Pittsburgh.
Hardware: Alpha Cluster
Software: automatic image registration (AIR)
Keywords: brain, cognitive neuroscience, functional MRI (Magnetic Resonance Imaging), imaging technology, brain-mapping, working memory, neurons, AIR (Automatic Image Registration) software, brain disturbances.
Related Material on the Web:
Projects in Scientific Computing, PSC's annual research report.
Live Demo at SC '96.
References, Acknowledgements & Credits