ASNR meeting Abstracts

Background: Wearable sensors (e.g. accelerometers) for tracking human physical activity have allowed for measurement of objective activity performance of the upper limb in daily life. Data extracted from accelerometers can be used to quantify multiple variables measuring different aspects of upper limb performance in one or both limbs. Work to date has focused on single variables, but upper limb performance is likely multidimensional. Here, we propose multivariate categories of upper limb performance, derived from wearable sensor data, as a potential solution for improving stroke rehabilitation care. Methods: This study analyzed data extracted from bimanual, wrist-worn triaxial accelerometers in adults from three previous cohorts (N=211), two samples of persons with stroke and one sample from neurologically intact adult controls. Data used were upper limb performance variables calculated from accelerometer data, associated clinical measures, and participant demographics. A total of 12 cluster solutions (3-, 4-, or 5-clusters based with 12, 9, 7, or 5 input variables) were calculated to systematically evaluate the most parsimonious solution. Quality metrics and principal component analysis of each solution were calculated to arrive at a locally-optimal solution with respect to number of input variables and number of clusters. Data from earlier time points will be evaluated for their potential to predict eventual cluster membership. Results/Anticipated Results: Across different numbers of input variables, two principal components consistently explained the most variance. Across the models with differing numbers of upper limb input performance variables, a 5-cluster solution explained the most overall total variance (79%) and had the best model-fit (AIC improvement of 184, compared to the next best model). The clusters are named by the amount of overall upper limb activity and integration of the upper limbs into daily activity. The category names in order of increasing upper limb performance are: Minimal Activity/Rare Integration, Minimal Activity/Limited Integration, Moderate Activity/Moderate Integration, Moderate Activity/Full Integration, and High Activity/ Full Integration. People from the stroke cohorts ended up in all 5 categories while the adult controls ended up in the moderate to high categories. Variables that may predict eventual cluster membership of those in stroke cohort will be provided. Discussion/Significance: We identified 5 categories of upper limb performance formed from 5 upper limb performance variables in cohorts with and without neurological upper limb deficits. Following validation on a larger, heterogeneous sample, these categories may be used as outcomes in upper limb stroke research and implemented into clinical rehabilitation practice.