Deep-learning-based identification of individual motion characteristics from upper-limb trajectories towards disorder stage evaluation

TL;DR

The paper addresses identifying individuals from upper-limb movement trajectories to support rehabilitation tracking and potential disorder-stage assessment. It adopts a standardized center-out transport task and an adapted ResNet18-based deep time-series learner to classify participants from 3D trajectory data captured by optical motion capture. Key findings show high separability for a small subset (≈95% accuracy for 9 participants) and substantial but reduced accuracy for the full cohort (≈78% for 31 participants), with performance varying by target. The work suggests practical potential for portable, data-driven monitoring of rehabilitation and progression of movement disorders, with future extension to disorder stages and transfer to mobile sensing.

Abstract

The identification of individual movement characteristics sets the foundation for the assessment of personal rehabilitation progress and can provide diagnostic information on levels and stages of movement disorders. This work presents a preliminary study for differentiating individual motion patterns using a dataset of 3D upper-limb transport trajectories measured in task-space. Identifying individuals by deep time series learning can be a key step to abstracting individual motion properties. In this study, a classification accuracy of about 95% is reached for a subset of nine, and about 78% for the full set of 31 individuals. This provides insights into the separability of patient attributes by exerting a simple standardized task to be transferred to portable systems.

Figures (4)

• Figure 1: Experimental set-up
• Figure 2: Typical trial path
• Figure 3: Participant classification performance (10 folds) with respect to task, for scenarios of 9 and 31 participants.
• Figure 4: Rounded percentual confusion matrix for correct individual classifications (all 31 participants).