We describe a method to objectively determine if robot-assisted gait training optimally challenges a patient. This information is important because patients benefit most from rehabilitation training that positively challenges and motivates them while not being too stressful or boring. We used physiological recordings to automatically classify whether the patient is mentally engaged. While nondisabled subjects could report this information verbally, cerebral lesions caused by traumatic brain injury or stroke might limit other patients' communication or self-assessment capabilities. Our method is the first approach toward a gait-training environment that detects the mental state of a patient and automatically adapts the training to the patient's needs and demands.
Volume 48 Number 4, 2011
Pages 367 — 386
Last Reviewed or Updated Tuesday, May 10, 2011 1:39 PM