Post-traumatic stress disorder (PTSD) is a mental health condition that scientists and psychologists have struggled to diagnose for decades.
Initially, PTSD in children was diagnosed through interviews, questionnaires, and discussions. However, PTSD has a high chance of going undiagnosed because children have limited communication skills, emotional awareness, or awareness of their surroundings. Hence, they can struggle to accurately verbalise and convey their emotions.
Now, researchers at the University of South Florida (USF) in Tampa, Florida, US, have successfully developed an AI system to address this problem. The research team headed by Alison Salloum and Shaun Canavan has harnessed facial recognition technology to identify PTSD in children by studying their facial expressions.
Salloum, who is a professor in USF’s School of Social Work, is a licensed clinical social worker, and specialises in diagnosing PTSD and trauma. She has also authored many books on the subject.
In the course of her work, Alison Salloum noticed that the facial expressions of the children she was interviewing intensified during trauma interviews. She subsequently reached out to Canavan, an associate professor in the Bellini College for AI, Cybersecurity and Computing, to know whether an AI-powered facial recognition system would be able to detect these expressions.
Canavan then developed an AI system that prioritises patient privacy by blurring identity details and only analysing physical data such as a child’s head pose, gaze and other facial gestures as well as their eyes and mouth. The underlying AI model was trained on over 100 minutes of video per child containing over 1,80,000 frames, and the system was later able to detect subtle facial muscle movements that were linked to emotional expression.
The researchers claimed that this is the first study to preserve doctor-patient confidentiality alongside context-specific PTSD classification. Salloum also emphasised that the AI system was not a replacement for clinicians but could serve as a valuable supplement.
Story continues below this ad
“The system could eventually be used to give practitioners real-time feedback during therapy sessions and help monitor progress without repeated, potentially distressing interviews.” Salloum was quoted as saying in a blog post on the USF website.
“Data like this is incredibly rare for AI systems, and we’re proud to have conducted such an ethically sound study. That’s crucial when you’re working with vulnerable subjects. Now we have promising potential from this software to give informed, objective insights to the clinician,” Canavan said.
The study also observed clinician-child conversations and interviews, which brought out more detailed facial expressions than a parent-child conversation. As per the researchers, this could be linked to either the child showing reluctance or shame to discuss certain issues with their parents.
The researchers are currently working on removing gender-, culture-, or age-specific biases within the AI system. This is particularly important when studying pre-schoolers. The AI system could also be used to diagnose other mental health conditions in children such as depression, anxiety, and ADHD.
Story continues below this ad
(This article has been curated by Purv Ashar, who is an intern with The Indian Express)