Using only brain scans and advanced computational models, researchers were able to identify bipolar disorder in a group of patients with 73% accuracy. When they replicated the study in another group, they achieved 72% accuracy in distinguishing people with bipolar disorder from healthy individuals.
Their findings—which, according to researchers, boasted accuracy rates consistent with the accuracy rates of many other diagnostic tests used in medicine—were published in the online June 4 Psychological Medicine.
Because bipolar and other psychiatric disorders are typically diagnosed based on symptoms alone, researchers consider the study a landmark in the diagnosis of mental illness.
Due to its complex presentation, bipolar disorder can easily be misdiagnosed as depression, schizophrenia, or other disorders, explained researcher Sophia Frangou, MD, professor of psychiatry and chief of the psychosis research program at the Icahn School of Medicine at Mount Sinai, New York. Accurate diagnosis can sometimes take up to a decade, causing significant disruption in many people’s lives.
Through their study, Dr. Frangou and fellow researchers sought to discover whether MRI could effectively aid mental health providers in better identifying patients with bipolar disorder. Their results confirmed the efficacy of brain scans, a diagnostic test considered acceptable by most patients.
"This approach does not undermine the importance of rigorous clinical assessment and the importance of building relationships with patients but provides biological justification for the type of diagnosis made," said Dr. Frangou. "However, diagnostic imaging for psychiatry is still under investigation and not ready for widespread use. Nonetheless, our results together with those from other labs are a harbinger of a major shift in the way we approach diagnosis in psychiatry."
1. Rocha-Rego V, Jogia J, Marquant A, Mourao-Mirand J, Simmons A, Frangou S. Examination of the predictive value of structural magnetic resonance scans in bipolar disorder: a pattern classification approach. Psychological Medicine. 2013 June: 1-14.