Reformulation in the Phase Plane Enhances Smoothness Rater Accuracy in Stroke

ABSTRACT

To improve the characterization of motor impairment, we compared the sensitivities of a phase plane metric with temporal domain measures derived from integrated squared jerk (ISJ). Five subjects with stroke and a cohort of 21 neurologically intact volunteers performed self-paced, isolated elbow flexions. Analysis of angular trajectories from the stroke group revealed that temporal domain metrics failed to detect a performance deficit at the p < .05 level, while the phase plane metric did resolve a deficit (p < .01). When applied to a subset of movements with arrest periods, the phase measure also uniquely identified impairment (Wilcoxon rank-sum test, p < .001). Finally, when tested on a data-driven model, the phase measure, but not temporal metrics, increased monotonically with the severity of trajectory distortions. We conclude that motion smoothness can be accurately measured in the phase plane.

Keywords:

• jerk
• movement
• phase plane
• single joint
• stroke

ACKNOWLEDGMENTS

The authors gratefully acknowledge Steven Escaldi, DO, and Eileen Crandall, PT, for patient administration, David I. Shreiber, PhD; Troy Shinbrot, PhD; and Don Yungher, PhD, for helpful suggestions, and an anonymous reviewer for guidance in the completion and writing of this work. This work was funded in part by NSF-IGERT fellowship to Michael Wininger (DGE 033196, Prabhas Moghe, PI), and a NIDRR RERC grant to William Craelius.