Wednesday , May 22 2019
Home / australia / Artificial intelligence can diagnose posttraumatic stress disorder by analyzing voices

Artificial intelligence can diagnose posttraumatic stress disorder by analyzing voices

A specially designed computer program can help diagnose PTSD in veterans by analyzing their voices.

Posted online on April 22 in the magazine Depression and anxiety, the study found that an artificial intelligence tool can distinguish – with an accuracy of 89 percent – among the voices of people with or without disorder of disorder of stress disorder disorder.

"Our results suggest that speech-based features can be used to diagnose this disease and, with more refinement and validation, they can be used at the clinic in the near future," explains Charles R. Marmar, MD, Lucius N , principal author of the study. Littauer Professor and chair of the Department of Psychiatry at the NYU School of Medicine.

More than 70 percent of adults around the world experience a traumatic event at some point in their lives, with up to 12 percent of people in some countries experiencing difficulties that suffer from stress disorder disorder. Those who have this condition experience a strong and persistent anxiety when they remember an unleashing event.

The authors of the study say that the diagnosis of post-traumatic stress disorder is most often determined by clinical interviews or a self-evaluation assessment, both inherently prone to bias. This has led to efforts to develop objective, measurable physical markers of PTSD progression, such as laboratory values ​​for medical conditions, but progress has been slow.

Learn to learn

In the current study, the research team used a technique of statistical learning / machine, called random forests, which has the ability to learn in what way individuals can be classified according to examples. These AI programs build "decision" rules and mathematical models that allow decision making with increasing accuracy as the amount of training data increases.

The researchers first recorded standard diagnostic interviews, which lasted for hours, called PTSD scale administered by clinicians or CAPS, of 53 veterans from Iraq and Afghanistan with PTSD related to military service, as well as 78 veterans without the disease . The recordings were fueled by a voice software from SRI International, the institute that also invented Siri, to obtain a total of 40,526 shots based on speech captured in small conversations, that the Ia program on the computer passed for patterns.

The random forest program links patterns of voice characteristics specific to PTSD, including a less clear discourse and a lifeless metallic tone, which for a long time was reported anecdotally that was useful in the diagnosis. Although the current study did not explore the mechanisms of post-traumatic stress disorder, the theory is that traumatic events change the brain circuits that process emotion and muscle tone, which affects the voice of " a person.

In advance, the research team plans to form the AI ​​voice tool with more data, validate it even more in an independent sample and request the approval of the government to use the tool in a clinical manner.

"Speech is an attractive candidate for use in an automated diagnostic system, perhaps as part of a future application of a smartphone PTSD, since it can be measured at a good price, remotely and not intrinsic, "says lead author Adam Brown, PhD, assistant professor at the Department of Psychiatry at the NYU School of Medicine.

"The word analysis technology used in the current study on PTSD detection is in the range of capabilities included in our speech analysis platform called SenSay Analytics ™," says Dimitra Vergyri , director of the Speech Research and Technology Laboratory of SRI International. "The software analyzes words – in combination with the frequency, rhythm, tone and articulating characteristics of speech – to infer the speaker's state, including emotion, feeling, cognition, health, mental health and the quality of communication. A series of industry applications visible in startups such as Oto, Ambit and Decoded Health. "


Along with Marmar and Brown, the authors of the Department of Psychiatry study were Meng Qian, Eugene Laska, Carole Siegel, Meng Li and Duna Abu-Amara. The authors of the study of SRI International were Andreas Tsiartas, Dimitra Vergyri, Colleen Richey, Jennifer Smith and Bruce Knoth. Brown is also an associate professor of psychology at the New School for Social Research.

The study was supported by the United States Army's Acquisition and Medical Acquisition Activity (USAMRAA) and the W81XWH-C-0004 Grant of the Telemedicine and Advanced Technology Research Center (TATRC), as well as by the Steven and Alexandra Cohen Foundation.

Media inquiries:

Jim Mandler

(212) 404-3500

Disclaimer: AAAS and EurekAlert! It is not responsible for the veracity of the news published in EurekAlert. by the institutions that contribute or by the use of any information through the EurekAlert system.

Source link