Researchers at the University of Pennsylvania Health System (UPHS) looked for subtle brain abnormalities that cannot be seen by the human eye. A study examined the entire brain, looking at distributed patterns of abnormalities rather than differences in specific regions of the brain.

"In this study, we used high-dimensional shape transformations in which we compared a brain image with a template of a normal brain. Through this comparison, we then determined where and how the patient's brain differed from healthy controls," explained Christos Davatzikos, PhD, Director of the Section of Biomedical Image Analysis in the Department of Radiology at Penn. "These methods are able to identify abnormalities that could not be detected by human inspection of the images created via MRI And, up until now, structural MRI has typically been used to diagnose physical anomalies like stroke or tumours, but it has not been helpful for diagnosis of psychiatric diseases."

Davatzikos says, "MRI produces images which are traditionally read mostly by radiologists. Now, we can do a quantitative reading of these images - bringing out information that is not obvious to the eye; one can think of computer readings as computational scanners. It's a second level that says 'analyze this image and produce another image that highlights subtle abnormalities in the brain.' This is fundamentally new information now that we can use for a larger spectrum of diseases and look for early diagnosis and prevention - such as the teen at risk for developing schizophrenia."

The results of the study demonstrate that sophisticated computational analysis methods can find unique structural brain characteristics in schizophrenia patients, with a predictive accuracy of more than 83%.; Source: University of Pennsylvania School of Medicine