University of Texas at Dallas Researchers Utilize 3D Motion Capture and Machine Learning to Detect Anxiety and Depression Through Gait Analysis

Researchers at UT Dallas use 3D motion capture and AI to detect anxiety and depression with 77% accuracy by analyzing how people walk and stand.

By: AXL Media

Published: Mar 31, 2026, 5:46 AM EDT

Source: Information for this report was sourced from University of Texas at Dallas

Decoding the Biomechanics of Emotional Distress

The relationship between physical movement and internal emotional states has long been observed qualitatively, but new research from the University of Texas at Dallas is translating these observations into quantifiable data. Dr. Gu Eon Kang and his team in the Neuromuscular and Musculoskeletal Biomechanics Lab have successfully used 3D motion capture technology to identify specific "movement signatures" associated with mental health struggles. By focusing on the mechanics of how a person walks or transitions from a seated to a standing position, the study suggests that the body’s physical carriage can reveal neurological undercurrents of anxiety and depression that might otherwise remain hidden during a standard clinical intake.

Machine Learning Accuracy in Mental Health Screening

To validate their findings, the research team employed sophisticated machine-learning models trained on data from thirty young adults. Participants were outfitted in form-fitting suits equipped with 68 reflective markers while a 16-camera system recorded their every joint movement. When tasked with predicting the mental state of individuals whose data was not used during the training phase, the model correctly identified symptoms of depression and anxiety with an accuracy rate of 75% for walking tasks and 77% for sit-to-walk transitions. This objective approach aims to provide a data-driven layer to mental health evaluations, which currently rely heavily on subjective patient questionnaires.

Subtle Physiological Indicators of Anxiety and Depression

The study noted that while it is a common trope to expect a slower pace from a "sad" individual, the actual biomechanical differences are far more nuanced. Doctoral student Angeloh Stout observed that subjects with higher symptom scores exhibited measurable hesitations during physical transitions and distinct variations in how their joints coordinated during a stride. These subtle changes in gait and posture serve as a physiological mirror to a person’s psychological state. According to Kang, while these tools are intended to support rather than replace professional diagnosis, they offer a reliable modality for detecting the physical manifestations of chronic mental stress.