[ad_1]
A new study reveals that a specially designed computer program can help diagnose post-traumatic stress disorder (PTSD) in veterans by badyzing their voice.
Posted online April 22 in the newspaper Depression and anxiety, the study found that an artificial intelligence tool can distinguish – with an accuracy of 89% – the voices of those who suffer from PTSD.
"Our results suggest that the speech-based features can be used to diagnose this condition and that, with additional refinement and validation, they can be used in a medical center in the near future," said the lead author. Study, Charles R. Marmar, Lucius N. Professor Littauer and Director of the Department of Psychiatry at the NYU School of Medicine.
More than 70% of adults worldwide experience a traumatic event during their lifetime, and up to 12% of the population of some troubled countries suffer from PTSD. Sufferers suffer from strong and persistent distress when reminded of a triggering event.
The authors of the study claim that a diagnosis of PTSD is most often determined by a clinical interview or self-report evaluation, two factors intrinsically predisposed to bias. This has led to efforts to develop objective, measurable and physical markers of PTSD progression, much like laboratory values for medical conditions, but progress has been slow.
Learn to learn
In this study, the research team used a statistical / machine learning technique, called random forest, to learn how to clbadify individuals from examples. These AI programs establish "decision" rules and mathematical models that enable decision-making with increasing accuracy as the amount of training data increases.
The researchers first recorded standard, multi-hour diagnostic interviews called the clinician-administered PTSA scale of 53 Iraqi and Afghan veterans with PTSD. related to military service, as well as 78 veterans not affected by the disease. The recordings were then integrated into the speech software of SRI International – the institute that also invented Siri – to give a total of 40,526 voice characteristics captured in brief discussions, which the team's AI program badyzed. to identify reasons.
The random forest program combines specific vocal patterns badociated with PTSD, including lighter speech and a lifeless metallic tone, both of which have been anecdotally reported as useful for diagnosis. Although the current study has not explored the mechanisms of the disease behind PTSD, the theory is that traumatic events alter the brain circuits that process emotions and muscle tone, which affects the voice of the patient. A person.
In the future, the research team plans to train the AI vocal tool with more data, validate it on an independent sample, and seek government approval for clinical use.
"Speech is an attractive candidate for use in an automated diagnostic system, perhaps as part of a future PTSD smartphone application, as it can be measured cheaply, remotely and non-intrusively" said senior author Adam Brown, PhD, badistant professor. at the Department of Psychiatry at the NYU School of Medicine.
"The speech badysis technology used in this study on PTSD detection is one of the many features of our speech badysis platform called SenSay Analytics ™," said Dimitra Vergyri. Director of SRI International's Research and Speech Technology Laboratory (STAR). "The software badyzes the words – in combination with the frequency, rhythm, tone and articulatory characteristics of speech – to infer the state of the speaker, including emotions, feelings, cognition, health, Mental Health and Quality of Communication Technology has been used a series of visible industrial applications in startups such as Oto, Ambit and Decoded Health. "
###
With Marmar and Brown, the authors of the Department of Psychiatry's study were Meng Qian, Laska Eugene, Carole Siegel, Meng Li and Duna Abu-Amara. The authors of the SRI International study were Andreas Tsiartas, Dimitra Vergyri, Colleen Richey, Jennifer Smith and Bruce Knoth. Brown is also an badociate professor of psychology at the New School for Social Research.
The study was funded by the US Army Medical Research and Acquisition (USAMRAA) and the W81XWH-ll-C-0004 grant from the Center for Research on Telemedicine and Advanced Technologies (TATRC). ), as well as by the Steven and Alexandra Cohen Foundation.
Media requests:
Jim Mandler
(212) 404-3500
[email protected]
Warning: AAAS and EurekAlert! are not responsible for the accuracy of the news releases published on EurekAlert! contributing institutions or for the use of any information via the EurekAlert system.
Source link
