A reduced sensor configuration for upper extremity monitoring after stroke

Abstract Background Home rehabilitation for stroke survivors is crucial for promoting upper extremity (UE) movements, improving functional ability, and enhancing independence and quality of life. Wearable sensors enable monitoring of movement trends and progress during home rehabilitation. The feasibility of using sensors in the homes of stroke survivors depends both on user acceptability and the accuracy of the sensors in capturing the complexity of movement. The aim of this study was to identify a reduced and more practical sensor configuration that preserves the ability to capture movement complexity for monitoring UE movements in stroke survivors during daily activities at home. This was informed by aligning user acceptability data with movement complexity measures derived from a nine-sensor reference configuration in a home-like environment. Methods Eleven chronic stroke survivors were observed in a natural or simulated home environment while attempting to attach and detach sensors and wearing them while performing self-chosen activities of daily living. Nine inertial measurement unit sensors were placed on the participants’ UEs and sternum. Acceptability was assessed with a custom-made questionnaire consisting of 14 items scored on a 1–5 Likert scale, and five open-ended questions. Furthermore, information entropy was calculated to determine the minimal number of sensors needed to capture movement complexity. Results Total acceptability of wearing the 9 sensors was high, with a median score of 57 out of 70 (81%; IQR = 11), whereas usability was moderate, with a median score of 11 out of 20 (55%; IQR = 5.5). The usability challenges were related mostly to attaching and detaching sensors on the hands and non-affected UE. The minimal sensor configuration to capture behavior complexity consisted of three-to-four sensors, placed on the sternum, non-affected and affected forearms, with or without affected upper arm. Conclusions The 9-sensor configuration had high acceptability, but lower usability. A reduced configuration with three-to-four sensors was sufficient to maintain the accuracy of the sensors, while potentially increasing stroke survivors’ usability. In turn, this may increase the feasibility of wearing sensors at home for stroke survivors.