Outside the Laboratory Assessment of Upper Limb Laterality in Patients With Stroke: A Cross-Sectional Study.

BACKGROUND: The rehabilitation of upper limb sensorimotor performance after stroke requires the assessment of daily use, the identification of key levels of impairment, and monitoring the course of recovery. It needs to be answered, how laboratory-based assessments and everyday behavior are connected, which dimension of metrics, that is, volume, intensity, or quality, is most sensitive to reduced function, and what sensor, that is, gyroscope or accelerometer, is best suited to gather such data.

METHODS: Performance in laboratory-based sensorimotor tests, as well as smartwatch-derived kinematic data of everyday life relative upper limb activity, during 1 day of inpatient neurorehabilitation (Germany, 2022) of 50 patients with stroke, was cross-sectionally assessed and resulting laterality indices (performance ratios) between the limbs were analyzed using ANCOVAs and principal component analysis.

RESULTS: Laboratory-based tests revealed the strongest laterality indices, followed by smartwatch-based (intensity>quality>volume) metrics. Angular velocity-based metrics revealed higher laterality indices than acceleration-based ones. Laterality indices were overall well associated; however, a principal component analysis suggested upper limb impairments to be unidimensional.

CONCLUSIONS: Our findings suggest that the use of sensors can deliver valid information of stroke-related laterality. It appeared that commonly used metrics that estimate the volume of use (ie, energy expenditure) are not the most sensitive. Especially reached intensities could be well used for monitoring, because they are more dependent on the performance of the sensorimotor system and less on confounders like age. The unidimensionality of the upper limb laterality suggests that an impaired limb with reduced movement quality and the inability to reach higher intensities will be used less in everyday life, especially when it is the nondominant side.