November 08, 2005

MRI for Schizophrenia Diagnosis?

This press release out of UPenn is interesting - because it contradicts a story we covered just last month on how MRI and brain imaging has been promising for a long time but still hasn't reached the point where it is truly valuable in diagnosis of schizophrenia.

This new press release suggests that perhaps things finally at least moving in a positive direction towards the point where they may soon be more valuable in a clinical setting. The study suggests that this new research approach is about 91% accurate in schizophrenia diagnosis (getting better - but still not quite where it needs to be). Still, however, this is good ammunition for those mental health advocates who are trying to gain equal coverage for schizophrenia in US mental health insurance plans (where it is frequently excluded or held to very low coverage levels).

MRI for early diagnosis of schizophrenia

New way of using MRI may show us what the naked eye cannot see
(Philadelphia, PA) - Researchers may have discovered a new way that may ultimately assist in the early diagnosis of schizophrenia - by utilizing MRI to study the patient's brain. 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 tumors, 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%. Recently, Davatzikos and his group announced that further analysis of this data with even more sophisticated classification methods achieved a 91% predictive accuracy for diagnosis of schizophrenia via MRI (MICCAI 2005 meeting, Palm Springs, CA).

"This is the first time this level of predictive power of MRI for classification of schizophrenia is demonstrated in a study of this magnitude," adds Davatzikos. "This tells us there are unique patterns we can use and explore when we want to diagnose patients with schizophrenia. However, the biggest value for this new diagnostic tool will be for early detection before clinical manifestation of the disease. For this, we will need to examine teenagers at risk."

Schizophrenia commonly presents in late adolescence or early adulthood thereby disrupting normal development and attainment of education and achieving independence. "If the disease can be detected early, intervention can ameliorate its potential effects. For example, brain systems implicated in schizophrenia include those required for learning and memory. Knowing that these systems have reduced volume in an individual could justify cognitive remediation efforts that will palliate the deficits and allow better adaptation," said Raquel Gur, MD, PhD, Director of the Schizophrenia Center with the Department of Psychiatry at UPHS, who performed the studies supported by NIMH.

Davatzikos further explains, "If you can diagnose schizophrenia early, utilizing MRI along with other tools like genetic disposition, behavioral profiles and functional imaging -- before a patient actually develops the disease -- we can try to delay the onset of the disease and hopefully have a better outcome for the rest of their life."

"Despite the high accuracy with the MRI classified patients and healthy controls, the diagnosis of schizophrenia is based on the clinical presentation," says Gur. "However, it is time for mental health professionals to think of neuroimaging as an important diagnostic tool that merits further research."

The results of this study are published in the November 2005 issue of the "JAMA - Archives of General Psychiatry." You will be able to access it on-line at: (The article is titled "Whole-Brain Morphometric Study of Schizophrenia Revealing a Spatially Complex Set of Focal Abnormalities").


"Davatzikos further explains, If you can diagnose schizophrenia early, ......before a patient actually develops the disease -- we can try to delay the onset of the disease and hopefully have a better outcome for the rest of their life."

What I want to know is, are they going to give every kid in the nation an MRI to see if s/he is sz?

Posted by: Chris at November 9, 2005 05:44 PM

Probably not - but wouldn't it be great as a part of a validation of any diagnosis that was based on a psychiatrist's assesment of the symptoms, or perhaps done at an early age (e.g. 13 years of age) for people who are of high risk (e.g. have a parent who has schizophrenia, or a number of other family members).

Early diagnosis has been shown to result in better outcomes, and so this could help in that early detection and treatment. Its one more "arrow in the quiver" as they say. One more tool to work with. And - it gives the insurance companies and antipsychiatrists like Tom Cruise an even more tenuous hold on their outrageous positions that they use to with-hold treatment and support for the people who have schizophrenia.

Posted by: szadmin at November 11, 2005 12:01 AM

can schizophrenia develope at the age of 79?

Posted by: Dean Frey at December 4, 2005 07:35 PM

Dean - yes, it can, but its rare. See the "Facts and Statistics" page on the home page of - and you'll see a graph that shows the distribution of when during their lives people get it.


Posted by: szadmin at December 4, 2005 09:04 PM

My son's doctor thinks he is in the prodromal stage of schizophrenia. He just had an MRI done last night. Is there a way we could have this type of analysis done on his MRI if it comes back normal? How relieved should we be if it comes back normal? According to this article it sounds like we shouldn't be relieved at all.

Posted by: elizabeth at January 18, 2006 02:23 PM

one of my relatives has recently diagnosed as a schizo ..he has been started on meds.. no episodes of psychosis on 2 episodes of fear (fright) .should MRi be done for him if is then any particular protocol.. do let me know my email is

Posted by: Debasis Sahoo at January 25, 2006 01:16 PM

i have schizophrenia and i had an MRI. the MRI triggered my anxiety. i imagine that MRI's would be very difficult for lots of schizophrenic who have anxiety triggered by noise and/or vibrations or claustrophobia. still, i hope it's a useful tool.

Posted by: schizo male 33 yrs old at March 26, 2006 02:52 PM

My daughter's doctor thinks she has schizophrenia or severe depression with psychosis. She just had an MRI done. Is there a way we could have this type of analysis done on her MRI if it comes back normal? How relieved should we be if it comes back normal? According to this article it sounds like we shouldn't be relieved at all.

Posted by: Zach at April 27, 2006 10:25 AM

Post a comment

Please enter this code to enable your comment -
Remember Me?
(you may use HTML tags for style)
* indicates required