The timed up & go test sit-to-stand transition: Which signals measured by inertial sensors are a viable route for continuous analysis?

Abstract

Background: The Timed Up and Go test is a well-known clinical test for assessing of mobility and fall risk. It has been shown that the IMU which use an accelerometer and gyroscope are capable of analysing the quantitative parameters of the sit-to-stand transition.

Research question: Which signals obtained by the inertial sensors are suitable for continuous Timed Up & Go test sit-to-stand transition analysis?

Methods: In the study we included 29 older adult volunteers and 31 de-novo Parkinson disease (PD) patients. All subjects performed an instrumented extended TUG wearing a gyro-accelerometer. The sit-to-stand transition was detected from an angular velocity signal. The sit-to-stand signal pattern within the subject group was analyzed via an intra-class correlation between curves. Inter-subjects' variability was visualized using prediction bands.

Results: The angular velocity about the pitch axis exhibited the best signal match across subjects in both groups (0.50 < ICC < 0.75). When analysing acceleration, the acceleration along the antero-posterior axis showed moderate inter-subjects signal pattern match (0.50 < ICC < 0.75) in the reference group. The analysis of other signals revealed a poor signal pattern in both subject groups.

Significance: For optimal interpretation of the analysis of continuous curves, the signal pattern must be considered. Also, the inter-subject variability along this pattern can be informative and useful.

Keywords: Acceleration; Angular velocity; Prediction bands; Repeatability.