How Can Robotic Devices Help Clinicians Determine the Treatment Dose for Post-Stroke Arm Paresis?

Abstract

Upper limb training dose after stroke is usually quantified by time and repetitions. This study analyzed upper limb motor training dose in stroke participants (N = 36) using a more comprehensive approach. Participants, classified by initial motor severity (severe/moderate/mild) and recovery trajectory (good/poor), received daily robotic and occupational therapy. Treatment dose was reported using a multidimensional framework. Fugl-Meyer Assessment (FMA) score and robot-derived kinematic parameters (reach distance (cm), velocity (cm/s), accuracy (cm) and smoothness (number of velocity peaks)) were analyzed pre- and post-intervention. FMA scores (mean (SD)) improved significantly post-intervention in severe (+11 (12) pts; p < 0.001) and moderate (+13 (6) pts; p ≤ 0.01) impairment groups. In the severe group, good recoverers showed greater improvement (+18 (12) pts) than poor recoverers (+4 (4) pts). Despite similar robotic therapy duration (34 min/session) and number of movements (600-900/session) between good and poor recoverers, both groups experienced very different therapeutic plans in the use of physical modalities: good recoverers gradually moved from assisted to the unassisted then resisted modality. Kinematic analysis showed distinct patterns of motor improvement across severity levels, ranging from quantitative (reach distance/velocity) to qualitative (accuracy/smoothness) changes. This approach provides a more accurate description of the therapeutic dose by characterizing the movements actually performed and can help personalize rehabilitation strategies.