Innovative Walking Test Enhances Concussion Assessment Accuracy

Detecting concussions following head injuries has traditionally relied heavily on subjective symptom reporting, such as dizziness or headaches, which can often be inaccurate or incomplete. Recognizing the need for a more objective and accessible solution, researchers at the University of Missouri have developed a portable, machine learning-enabled system designed to assess signs of concussion more effectively. Led by Associate Professor Trent Guess and doctoral student Jacob Thomas, this innovative tool combines a force plate, a depth camera, and an interface board to analyze body motion, balance, and reaction times in real-time.

The system was tested on 40 college athletes—20 with recent concussion diagnoses—and demonstrated remarkable accuracy in distinguishing between healthy individuals and those with head injuries. It identifies subtle motor impairments, such as slower walking speed, delayed reaction times, and difficulty maintaining balance, which become especially evident during tasks with eyes closed or while performing cognitive challenges like counting backward.

One of the key advantages of this system is its portability and affordability, making it suitable for a variety of settings outside specialized labs, including clinics, sports facilities, and field environments. The machine learning component allows the system to establish a baseline for each individual when they are healthy. Future assessments can then compare current performance to this baseline, providing a personalized evaluation of recovery or ongoing impairment after potential concussions.

Professor Guess emphasizes the system’s potential to improve concussion management by providing objective data that can assist in diagnosis, track recovery, and prevent premature return to physical activity or work. While currently in research phases, such portable assessment tools could also benefit first responders, military personnel, and others in high-risk professions. The ultimate goal is to produce scalable, commercially available devices to enhance concussion assessment protocols across multiple disciplines.

This approach, detailed in the journal Medical Engineering and Physics, exemplifies how leveraging engineering, biomechanics, and machine learning can lead to significant advancements in healthcare diagnostics.

source: https://medicalxpress.com/news/2025-09-guesswork-concussion.html