Passive Shoulder Exoskeleton Use Over Days During a Variable Visuomotor Wiring Task

Abstract

This study investigated the use of shoulder exoskeletons in a visuomotor upper extremity wiring task over 3 days. Twenty-four participants were randomly assigned to an exoskeleton or control group. Their task performance, perceived exertion and workload, erceived acceptance, and neural activation patterns over 3 days were obtained. Despite similar exertion ratings between groups, the exoskeleton group reported lower perceived workload and higher task accuracy, suggesting that users benefited from the exoskeleton’s use, allowing for greater resources to attend to task-related cognitive processes. Neuroergonomics analysis revealed differences in neural signatures between the two groups indicating that exoskeleton group exhibited increased activation over days in motor function-related brain regions, while the control group demonstrated greater connectivity within the executive control network. These findings suggest that shoulder exoskeletons offer physical and cognitive benefits in tasks that require complex cognitive processing through specific neuromotor adaptations over days, which has training implications for exoskeleton use.

Extended Abstract

Shoulder work-related musculoskeletal disorders (WRMSDs) account for one-third of all WRMSDs in the manufacturing industries (Punnett & Wegman, 2004). Repetitive overhead motions, overexertion, and prolonged awkward postures associated with these tasks increase the risk of musculoskeletal injuries (Maurice et al., 2019) . Passive shoulder exoskeletons have been introduced as an ergonomic intervention to mitigate the risk by supporting the weight of the upper arm and redistributing muscular load from shoulder to back, thereby reducing strain on the muscles (De Looze et al., 2016). Exoskeleton use has been primarily evaluated for physical benefits through muscle activity, task performance, and user experience. However, it is also important to consider the cognitive effort in addition to physical demands required for industrial tasks and how exoskeleton use impacts the human-exoskeleton interaction (Bequette et al., 2020; Zhu et al., 2021). An increase in cognitive effort with the use of exoskeletons has previously been assessed by evaluating performance, perceived workload, and user satisfaction (Stirling et al., 2018). Zhu et al. (2021) used functional near-infrared spectroscopy (fNIRS) to identify workload-related changes in brain activity, especially between the functionally independent frontal and motor regions of the brain with lower-back exoskeleton use. However, the specific effects of shoulder exoskeletons on cognitive demands and adaptation processes over time remain understudied. This study aims to quantify the neural and perceived impacts of passive shoulder exoskeleton use during a complex wiring task over 3 days, with a hypothesis that the exoskeleton would affect task performance, increase front-motor connectivity, and demonstrate adaptation through improved performance and reduced demands. In this study, 24 participants were recruited and assigned to either a control or exoskeleton group. Participants were asked to perform a Trail-Making Test B on a whiteboard simulating a wiring task on 3 days. The task required cognitive processing, visual search, and arm movements. The exoskeleton group performed the tasks with the assistance of the EksoVest exoskeleton. Measures such as ratings of perceived exertion (RPE) and subjective workload (NASA-TLX) were collected. Technology Acceptance Model (TAM) questionnaire was used to assess participants’ perceptions of the exoskeleton. Performance was evaluated based on the average target acquisition time, the average standard deviation for acquiring each target from the average time, and the number of correct responses. fNIRS was used to collect brain activity in the frontal and motor regions of the brain to assess peak activation in functionally independent regions of the brain and functional connectivity between them. Results indicate RPE decreased over days (p = .001) for both groups, indicating an adaptation to task over time. Participants in the exoskeleton group reported lower mental demand (p = .05), effort (p < .001), and frustration (p < .001), indicating a decreased mental workload associated with the exoskeleton intervention. However, comparable ratings of RPE indicate no influence of exoskeleton use on the perception of physical effort despite lowered perceived mental workload. Both groups demonstrated similar motor skill acquisition, but on day 3, the exoskeleton group showed significantly improved accuracy (p = .005) compared to the control group. The comparative analysis of the neural activation pattern and functional connectivity reveals enhanced activation among the functionally distinct regions of the brain in the exoskeleton group, while the greater functional connection between front-motor in the control group over days suggests an effective translation of cognitive processing into motor commands, therefore improving motor skills. However, as the strategies were developed to improve performance available cognitive resources declined therefore comparable cognitive performance over days. However, the exoskeleton group developed different neural strategies resulting in improved task accuracy and lower perceived mental demands. These findings highlight the potential benefits of exoskeleton use for complex tasks requiring dual motor-cognitive processing and inform training strategies and user acceptance of such technologies.