Postural control and gait measures derived from wearable inertial measurement unit devices in Huntington's disease: Recommendations for clinical outcomes

Background: Postural control impairments begin early in Huntington's disease yet measures most sensitive to progression have not been identified. The aims of this study were to: 1) evaluate postural control and gait in people with and without Huntington's disease using wearable sensors; and 2) identify measures related to diagnosis and clinical severity. Methods: 43 individuals with Huntington's disease and 15 age-matched peers performed standing with feet together and feet apart, sitting, and walking with wearable inertial sensors. One-way analysis of variance determined differences in measures of postural control and gait between early and mid-disease stage, and nonHuntington's disease peers. A random forest analysis identified feature importance for Huntington's disease diagnosis. Stepwise and ordinal regressions were used to determine predictors of clinical chorea and tandem walking scores respectively. Findings: There was a significant main effect for all postural control and gait measures comparing early stage, mid stage and non-Huntington's disease peers, except for gait cycle duration and step duration. Total sway, root mean square and mean velocity during sitting, as well as gait speed had the greatest importance in classifying disease status. Stepwise regression showed that root mean square during standing with feet apart significantly predicted clinical measure of chorea, and ordinal regression model showed that root mean square and total sway standing feet together significantly predicted clinical measure of tandem walking. Interpretations: Root mean square measures obtained in sitting and standing using wearable sensors have the potential to serve as biomarkers of postural control impairments in Huntington's disease.