Validity Assessment of Sensors for Movement Tracking With Children/Youth With Diverse Neuromotor Abilities

Problems in prenatal brain development or brain injury in infancy can cause cerebral palsy (CP), a non-progressive disorder associated with impaired movement, posture, and balance. CP is the most common motor disability in children affecting two to three children for every 1000 births.

Although CP is a permanent disability, occupational and physical therapy can be implemented to help improve motor function or manage the symptoms over time. However, access to these therapies can be limited by financial, time, and geographical constraints. This motivates interest in home-based rehabilitation programs. Several studies have evaluated the possibility of active video games (AVGs) to complement conventional in-person rehabilitation interventions and boost motivation in children. The user interacts with the environment within the AVG via different motion-tracking interfaces ranging in complexity from inertial measurement units (i.e. Wii control stick) to 3D depth sensors (i.e. Microsoft Kinect). From a meta-analysis conducted by Ren et al., AVGs have been found to enhance gross motor function for children with CP.

Given the individual rehabilitation goals and abilities of each child, AVGs are most effective when they can be calibrated to target specific body segments for motor skill development. Ideally, AVGs would adapt to the child's physical performance in the same way that a therapist observes a patient's progress and adjusts the intensity and frequency of the exercise. Data collected from technological systems, such as an optoelectronic motion tracking system, has the potential to follow changes in movement abilities to support progress tracking. Gold standard optoelectronic motion capture (e.g., Vicon, Motion Analysis) are too expensive for the home, but there are low-cost, commercial products that are practical for home use (e.g., Microsoft Kinect, Orbbec Persee). While the movement-tracking capabilities of these devices are sufficient to support game play, it is not known how accurately they track movements (i) during game play and (ii) for children with diverse motor skills and patterns. This information is pertinent to validate the use of low-cost optoelectronic, motion-tracking to (i) provide high quality feedback and support practice of therapy exercises (e.g. to what extent can these systems recognize and provide feedback on exercises in the same way that a therapist would?) and (2) track changes in motor performance (i.e. to what extent can these systems be used to accurately quantify and track changes in motor skills?)

To increase accessibility for a larger population of children with CP, a series of interactive video games have been created at the Possibility Engineering and Research Lab (PEARL), which tracks rehabilitation exercises and integrates these into gameplay. The video games, Bootle Blast (BB) and Bootle Boot Camp (BC), utilize a 3D depth sensor, the Orbbec Persee+. Bootle Blast has been successfully piloted by 4 children with CP over 12 weeks in their homes and used in clinics at Holland Bloorview Kids Rehabilitation Hospital since 2017. To date, body tracking has been used primarily to support children's interactions in the games. Its accuracy for tracking therapy exercises, potential to deliver high quality feedback, and validity for quantifying qualities of movement (e.g., reach envelope, smoothness of movement, movement trajectory) has yet to be established.

The overall goal of this study is to establish the clinical utility and accuracy of markerless motion captures systems (e.g. the Orbbec Persee+/Astra 2, Apple LiDAR, 2D cameras + body tracking software) for tracking therapy exercises and movement during game play.