http://schizophreniajourney.com/pathogenesis/researchers-identify-74-schizophrenia-cases-via-brain-scan-data/

Researchers identify 74% of schizophrenia cases via brain scan data

An intense search for a reliable biomarker takes a new turn.

An intense search for a reliable biomarker takes a new turn.

In the race for a reliable schizophrenia biomarker, researchers say they’ve made significant progress toward developing a way to use brain scans to confirm that patients have the condition and to detect early warning signs.

So far, the researchers report, they’ve been able to correctly identify 74% of schizophrenia cases by analyzing patterns in brain scan data gathered via functional MRI (fMRI) (npj Schizophrenia. 2017 May 16;3[22]:1-12).

“This research demonstration represents a key advance about what machine learning and data mining can do in terms of understanding brain activity,”

psychologist Andrew J. Greenshaw, PhD, a coauthor of a new study and professor of psychiatry at the University of Alberta, Edmonton, said in an interview.

Dr. Greenshaw (far left) and colleagues discuss the need for integrated data capture for early identification of mental illness. Courtesy University of Alberta, Edmonton

Biomarkers are a hot topic in mental illness in general and schizophrenia in particular. Researchers have identified potential schizophrenia markers in the blood and in the brain, where they’re revealed via imaging and EEG.

“Failures are far too common for devastating, disabling conditions like schizophrenia.

But we do not use any direct, reliable assays of brain function to inform diagnoses,” Gregory Light, PhD, professor and deputy vice chair of the psychiatry education and training school at the University of California, San Diego, said in an interview.

Instead, psychiatric diagnoses “are based on a clinician’s inference about a person’s inner experience,” Dr. Light said.

The team of researchers, led by University of Alberta postdoctoral, fellow Mina Gheiratmand, PhD, focused on the results of whole-brain fMRI scans of 95 subjects – 46 patients with schizophrenia or schizoaffective disorder, and 49 healthy controls matched by gender and age.

All subjects had IQs of more than 75, no major illness other than schizophrenia, and no substance or alcohol dependence.

Study coauthor Guillermo A. Cecchi, PhD, said in an interview that the study focuses on how schizophrenia “is a disruption of how the brain is talking to itself.” He is principal research staff member at IBM’s Thomas J. Watson Research Center in Yorktown Heights, N.Y.

Guillermo A. Cecchi, PhD

Researchers focused on brain scan data that reveals the connections between tens of thousands of nodes in the brain that represent the brain’s estimated 100 billion neurons. This allowed the researchers to develop maps of connections between nodes in the brains of healthy people and the brains of people with schizophrenia, Dr. Cecchi said.

He likened the process to analyzing a database of 2.5 million websites that link to a single site. By analyzing the linking websites, ranking their importance, and counting the number of connections, he said, you could detect a pattern that tells you about the target site. The researchers were able to use brain data to correctly identify the brains of patients with schizophrenia 74% of the time. They also developed ways to predict with statistically significant accuracy the severity of symptoms on five scales – inattentiveness, bizarre behavior, positive formal thought disorder, avolition/apathy, and alogia.

“We are able to infer the clinical scales based on how the different parts of the brain are talking to each other or related to each other,”

Dr. Cecchi said. The researchers also noted that the brains of patients with schizophrenia showed a higher level of “hyperconnectivity” than that of controls. Other researchers have been studying hyperconnectivity in schizophrenia patients.

There’s also the matter of the required resources. “The technology is not widely accessible, and it’s not practical in terms of screening patients on a day-to-day basis. It remains a research tool at the moment,” said Dr. Greenshaw, the professor of psychiatry and study coauthor. MRI scans alone cost hundreds of dollars, he said, and require visits to imaging centers.

“But the promise is there,” he said, emphasizing that the technology could help psychiatrists move closer to the illusive goal of cracking the mystery of prodromal schizophrenia, the period when the condition is in its infancy and may be especially vulnerable to treatment.

“If we understand the features in those who are developing symptoms,” he said, “we might be able to catch them before they fall.”

What’s next? “Ultimately, there will be a biomarker purgatory, a vetting of biomarkers to determine which ones will survive, and which ones have clinical utility and can actually be delivered in real-world clinical settings,” said Dr. Light, who’s working on using EEGs to predict and monitor responses to treatments for mental illness.

For example, he asked, can MRI scans be performed quickly and efficiently, even in rural settings, without specialized tools, equipment, and staff members?

“We still have a very long way to go,” he said. Still, considering the tremendous burden of schizophrenia on patients, family members, and society, he said, “we need to do better for this group of people.”