Going beyond PA: Assessing sensorimotor capacity with wearables in multiple sclerosis—a cross-sectional study

Abstract

Abstract Background Wearable technologies are currently clinically used to assess energy expenditure in a variety of populations, e.g., persons with multiple sclerosis or frail elderly. To date, going beyond physical activity, deriving sensorimotor capacity instead of energy expenditure, is still lacking proof of feasibility. Methods In this study, we read out sensors (accelerometer and gyroscope) of smartwatches in a sample of 90 persons with multiple sclerosis over the course of one day of everyday life in an inpatient setting. We derived a variety of different kinematic parameters, in addition to lab-based tests of sensorimotor performance, to examine their interrelation by principal component, cluster, and regression analyses. Results These analyses revealed three components of behavior and sensorimotor capacity, namely clinical characteristics with an emphasis on gait, gait-related physical activity, and upper-limb related physical activity. Further, we were able to derive four clusters with different behavioral/capacity patterns in these dimensions. In a last step, regression analyses revealed that three selected smartwatch derived kinematic parameters were able to partially predict sensorimotor capacity, e.g., grip strength and upper-limb tapping. Conclusions Our analyses revealed that physical activity can significantly differ between persons with comparable clinical characteristics and that assessments of physical activity solely relying on gait can be misleading. Further, we were able to extract parameters that partially go beyond physical activity, with the potential to be used to monitor the course of disease progression and rehabilitation, or to early identify persons at risk or a sub-clinical threshold of disease severity.