A study submitted for peer review by NYU Langone researchers suggests that digitally analyzing a patient’s voice could significantly improve screening for post-traumatic stress disorder (PTSD). More than 70 percent of adults worldwide experience a traumatic event at some point in their lives, with up to 12 percent of the population in some countries suffering from PTSD. Yet the methods typically used to diagnose PTSD—clinical interviews or self-report assessments—are inherently subjective and prone to bias.
“Our findings show that speech-based characteristics are promising for screening for PTSD and, with further refinement in the future, may be employed in routine clinical practice,” says first author Charles R. Marmar, MD, the Lucius N. Littauer Professor of Psychiatry and chair of the Department of Psychiatry. The study was supported by a grant from the U.S. Army Medical Research & Acquisition Activity (USAMRAA) and Telemedicine & Advanced Technology Research Center (TATR), as well as by the Steven and Alexandra Cohen Foundation and Cohen Veterans Bioscience Inc. It was co-led by Adam D. Brown, PhD, adjunct assistant professor of psychiatry and associate professor of psychology at the New School for Social Research.
Using Artificial Intelligence to Detect Speech Patterns
The team began by recording standard Clinician-Administered PTSD Scale (CAPS) interviews of 52 Iraq and Afghanistan veterans with military service–related PTSD, as well as 77 veterans without the disorder. The recordings were then fed into voice recognition software from Stanford Research Institute to yield a total of 40,526 speech-based features. Using artificial intelligence (AI) to sift through these fragments for distinctive patterns, the program was able to distinguish PTSD with close to 90 percent accuracy.
“Our findings show that speech-based characteristics are promising for screening for PTSD and, with further refinement in the future, may be employed in routine clinical practice.”—Charles R. Marmar, MD, Chair of the Department of Psychiatry
Moving forward, the researchers plan to train the AI voice tool in new samples to further validate the tool. “In the past, clinicians reported anecdotally that mental health issues correspond with changes in vocal qualities, but with advances in computation and technology, we are now able to detect specific patterns of speech characteristics highly associated with PTSD,” notes Dr. Brown. “We believe speech is an attractive candidate for use in an automated diagnostic system, perhaps in a future PTSD smartphone app, because it can be measured objectively, cheaply, remotely, and nonintrusively.”
Measuring the Impact of Trauma on United Nations Staff
A growing body of research suggests that individuals who bring urgently needed help to global trouble spots—such as humanitarian aid workers, human rights advocates, and peacekeepers—may be at elevated risk for psychiatric disorders due to exposure to direct and secondary traumatic stress in their work. Yet there have been few studies on the rates of psychiatric morbidity among such personnel or on the occupational and organizational factors that might be associated with their vulnerability. This year, NYU Langone researchers completed the largest and most comprehensive study of these issues to date.
Led by Dr. Brown, the team—including Dr. Marmar; Danny Horesh, PhD, adjunct professor of psychiatry; Carole Siegel, PhD, research professor of psychiatry; and Francis G. Mas, MD, clinical professor of psychiatry—surveyed 17,363 personnel from 11 United Nations (UN) agencies involved with aid, human rights, and peacekeeping activities. Responders completed online measures of PTSD, generalized anxiety disorder, major depressive disorder, trauma exposure, mental healthcare usage, and sociodemographic information.
Although the results of the study remain embargoed, they indicate that participants experienced traumatic events, as well as associated psychiatric disorders, at a significantly higher rate than the general population. Only a small proportion of those affected, however, sought mental health treatment in the past year. “Our findings suggest that the implementation of a comprehensive mental health strategy is needed throughout the UN,” says Dr. Brown. “We recommend that such measures also be considered by other organizations providing aid to the world’s most vulnerable communities.”