Post-traumatic stress disorder (PTSD) is one of the most difficult psychiatric disorders to diagnose because diagnosis presently relies on a subjective process in which patients share information and clinicians make assessments based on what their patients tell them. Oftentimes, and for various reasons, patients may either exaggerate symptoms or downplay them.
Researchers at NYU School of Medicine, led by Charles R. Marmar, MD, the Lucius N. Littauer Professor of Psychiatry and chair of the and executive director of Ƶ’s Cohen Veterans Center, have created a new algorithm that can analyze pre-recorded interviews with patients and determine, based on voice analysis and with almost 90 percent accuracy, who does and does not have PTSD. Dr. Marmar and his colleagues recently published their findings in the peer-reviewed journal Depression and Anxiety.
“What we found was quite different than what we had hypothesized,” Dr. Marmar says. “We thought the telling features would reflect agitated speech. What we found instead was that voice biomarkers for PTSD identified flatter, more atonal speech.”
Members of the U.S. Armed Forces and others studying PTSD welcome the findings and hope that Dr. Marmar and his team will build on their initial work. At the Department of Veterans Affairs, PTSD claims have tripled over the past few years, raising concerns that the current, subjective diagnosis methods are not accurate enough.
Read more from .
