Diagnosing post-tractic stress disorder in children can be very difficult. Many, especially limited communication skills or emotional awareness, struggle to explain what they are feeling. Researchers at the University of South Florida are working to improve patient results by addressing those intervals and merging their expertise in childhood trauma and artificial intelligence.

Under the leadership of Elison Sallum, Professor at the USF School of Social Work, and Sean Canavan, Belini College of Artificial Intelligence, Cybercity and Associate Professor, Interdisciplinary Team is creating a system that can provide a purpose, cost-affected equipment to help children identify PTSD.

Published in study Pattern recognitionParticipant is the first of its kind to include reference-individual PTSD classification while fully preserving privacy.

Traditionally, diagnosis of PTSD in children depends on the subjective clinical interview and self-reported questionnaire, which can be limited by cognitive development, language skills, avoidance behavior or emotional suppression.

It actually started when I saw how intense the facial expressions of some children became during the trauma interview. Even when they were not saying much, you could see what they were doing on their face. When I talked to Sean whether AI could help in finding it in a structured manner. ,

Professor in Alison Sallum, USF School of Social Work

Canavan, which specializes in facial analysis and emotional recognition, prepared existing equipment in its laboratory to build a new system that prefer the patient privacy. The technology is overcome by identifying details and analyzes only de-identified data, including head pose, eye gaze and facial landmarks such as eyes and mouths.

“This is what makes our view unique,” said Canavan. “We do not use raw videos. We completely get rid of the identity of the subject and only keep data about facial movement, and we are the factors in the fact whether the child was talking to the parents or the doctor.”

The team created a dataset with 18 sessions with children as they shared emotional experiences. Videos of more than 100 minutes per child and each video contain around 185,000 frames, the AI ​​model of Canavanaval extracted a series of subtle facial muscle movements associated with emotional expression.

The findings showed that different patterns are detected in the face movements of children with PTSD. Researchers also found that facial expressions were disclosing more than the parents’ child’s conversation during the doctor-headed interviews. It aligns with existing psychological research that shows children that the physician can be more expressive and avoid sharing crisis with parents due to shame or their cognitive abilities.

“This is where AI can offer a valuable supplement,” Sallum said. “Not the place of physicians, but increasing their equipment. The system can eventually be used during medical sessions to respond to real-time reaction and repeatedly, helps in monitoring progress without probably disturbing interviews.”

The team expects to expand the study to examine the penis, culture, and age, especially any possible prejudice from preschool, especially preschool, where oral communication is limited and the diagnosis almost completely depends on the observation of the parents.

Although the study is still in its early stages, Sallum and Canavan feel that potential applications are far -reaching. Many of the current participants had complex clinical paintings, including co-suffered conditions such as depression, ADHD or anxiety, reflecting real-world matters and promising to the system’s accuracy.

“Such data is incredibly rare to the AI ​​system, and we are proud to do such a morally sound study. It is important when you are working with weak themes,” said Cainavan. “Now we have a promising ability from this software to inform the physician, purposeful insight.”

If validated in large trials, the USF approach can be redefined how PTSD is diagnosed and tracked in children, using everyday equipment such as video and AI to bring mental health care in future.