$1.3 million DoD investment advances machine learning research on PTSD and military family health at U of I

MOSCOW, Idaho — Two major multi-institutional grants from the U.S. Department of Defense (DoD) will support machine learning research at University of Idaho to advance the diagnosis of post-traumatic stress disorder (PTSD) and strengthen support systems for military families experiencing stress throughout the deployment cycle. Together, the two multi-site projects total more than $6 million, with $1.33 million supporting research at U of I.

Colin Xu, an assistant professor in the Department of Psychology and Communication in the College of Letters, Arts, and Social Sciences, will lead the development of machine learning models to improve early detection of adverse health outcomes for military servicemembers and their families.

“These projects will examine how machine learning techniques can be applied to psychiatric and public health data in order to strengthen military health,” Xu said. “Through these machine learning models, we hope to improve early identification and prediction of PTSD risk for servicemembers and improve prediction and prevention of adverse health outcomes for servicemember families.”

The first award — a $4.2 million collaborative DoD-funded project, with $974,000 going to U of I — will examine how smart wearable devices, biochemical markers and biophysical signals can be integrated to improve PTSD screening and diagnosis.

This multi-institutional effort includes co-principal investigators Manish Bhomia from Uniformed Services University, Dr. Christina La Croix from Walter Reed National Military Medical Center, Sameer Sonkusale from Tufts University, Dr. Steven Cohen from Northwestern University, and CJ Brush from Auburn University. Xu’s team at U of I will develop the machine learning models to identify how biomarker and biophysical data collected from wearable technologies can be leveraged to improve diagnostic accuracy for PTSD.

“PTSD is a complex multifaceted disorder,” Xu said. “By applying machine learning models to this multimodal biosensor data, we can better understand the biological signatures of PTSD and help clinicians improve diagnosis and early detection.”

The second award — totaling $1.9 million, with $361,000 supporting U of I research — focuses on understanding how deployment-related stress affects military families and increases risk for harmful behaviors.

Xu’s team will partner with co-principal investigators Elizabeth Hisle-Gorman, Christin Ogle and Dr. Stephen Cozza at Uniformed Services University to identify predictors of family violence, substance misuse, suicidality and injury in military families.  

By applying machine learning models to large-scale, longitudinal military family health care records, the project seeks to identify risk factors of harmful behaviors within military families and understand how these risk factors unfold across pre-deployment, deployment and post-deployment phases.

Xu’s machine learning models will help identify subgroups of military families at elevated risk — as well as the timing and interaction of risk factors — to support earlier and more targeted interventions.

“Military families navigate unique stressors throughout the deployment cycle,” Xu said. “Our work aims to give clinicians better insight into who may be at elevated risk, and when, so that support can be provided proactively.”

The PTSD and wearable technologies project will span four years, and the military family health project will run for three years, beginning in 2025.

Xu is recruiting three funded graduate students and two postdoctoral researchers for these projects. Anyone who is interested should contact him at colinxu@uidaho.edu.

Media contact 

Colin Xu  
Assistant professor 
College of Letters, Arts and Social Sciences 
colinxu@uidaho.edu