Develop methods to determine the patient cognition during equine-assisted therapy (EAT). EAT is a form of a therapy that involves interactions between patients and horses during therapeutic riding. EAT has been known to be effective in enhancing gait and balance of neurologically-impaired patients (e.g., cerebral palsy, autism) and in alleviating post-traumatic stress disorder (PTSD) symptoms in military veterans. The effectiveness of EAT significantly depends on the quality of cognitive states of all participating bodies including patients, horses and the physical therapists. For example, when patients are distracted and cannot focus on the activities during riding, the expected outcome of the therapy worsens. The same concept applies for both the horses and the therapists. In this proposal, we confine our scope only to the patients since patients are the main target for the therapy and their active involvement in the therapy session is known to be one of the most important factors. Therefore, it is important to determine the patient cognition during the EAT session. Currently, there are no algorithms that can quantify the cognition online except the qualitative estimation from the experience of the therapists during the session.