Clinicians Controlling Rehabilitation Technology

Neurological deficits due to stroke, cerebral palsy and spinal cord injury cause severe limitations in mobility. Clinical tests qualitatively assess impairments but suffer from low inter-rater agreement and lack specificity for prescribing clinical interventions. Also, the repertoire of interventions is limited. Emerging technologies create new possibilities to both objectively assess performance from ubiquitous sensors and to deliver new therapeutic treatments via (wearable) robotics. However, the required engineering skills restrict influence and control of clinicians during the design process. We aim to create a robust co-development platform where engineers and clinicians jointly develop performance metrics and novel interventions. The key is to offer accessible design interfaces and “translations” between engineering and clinical languages. Our proposal is motivated by two driving clinical problems: 1) gait impairments in adults with incomplete spinal cord injury, and 2) trunk and head stability of infants with cerebral palsy. We leverage biomechanical models to transform sensor signals into clinical metrics and to convert clinical actions into actuator forces through easy-to-control graphical interfaces. This empowers clinicians to explore and design interventions for individual patients, leading to improved outcomes and valuable labeled intervention and outcome data that can be used to recommend future treatments.