Abstract Text: Mild traumatic brain injury (mTBI) can be difficult to identify due to lack of radiographic findings and non-specific symptoms. Missing early diagnosis of mTBI can lead to severe sequelae with a risk of insufficient medical intervention and potential exposure to repetitive injury for the patient, resulting in increased risk of progressive neurodegeneration. Therefore, there is a need for early, objective, and sensitive mTBI diagnostics and measurements in identification of mTBI. Using the Vestibular/Ocular-Motor Screening (VOMS) concussion evaluation test as a guide, we devised virtual reality (VR) tests that objectively measure movement/coordination, balance, auditory perception, memory, and oculometrics. In addition to specific task-related metrics, data is collected from 19-contact EEG (standard 10-20 system), electrocardiogram, insole pressure sensors, eye-tracking, and three inertial measurement units. We aim to collect data from 100 subjects with mTBI (within 6 months of injury) and 100 age/gender-matched healthy controls in collaboration with Dr. Richard Zorowitz and Kathaleen Brady at Medstar Health. Applying ML/AI techniques to this dataset, we aim to identify characteristic signal features that are sensitive to mTBI and ultimately tune model hyperparameters for an on-site/deployable mTBI diagnostic tool. Providing rapid and sensitive biomarkers may improve clinical diagnoses and faster treatment of mTBI patients. Additionally, regulatory reviewers and device manufacturers will have access to validated biomarkers as benchmarks for developing future devices with indications for diagnosis of TBI. These biomarkers can supplement other outcome measures (e.g., CT scan and subjective neurological examinations) within the medical brain computing interface and VR spaces.