Neural-enhanced motion-to-EMG: refining simulated muscle activity from musculoskeletal models using a Seq2Seq approach

Abstract

Electromyography (EMG) is essential for accurate assessment of motor function in rehabilitation, sports science, and robotics. However, its various time-consuming human operations (e.g., electromagnetic noise countermeasures) limit its widespread use. Meanwhile, motion capture technology has become more accessible, leading to increasing interest in musculoskeletal simulation models such as OpenSim. Although advances have been made in individualizing the model parameters, accurately estimating muscle activity remains a significant challenge. Previous efforts to optimize the parameters in musculoskeletal model simulators have yielded limited improvements in estimation accuracy. A key source of error that is identified in this study is the spatio-temporal distortion between the estimated and actual muscle activity, which has been inadequately addressed in previous research. To address this problem, this study proposes the Neural-Enhanced Motion-to-EMG (NEM2E) framework, which mitigates spatio-temporal distortions in simulated muscle activity using the Spatio-Temporal Distortion Refinement Network (STDR-Net). The STDR-Net is implemented via a Sequence-to-Sequence model with attention mechanisms to refine the estimates. Validation on two public datasets (walking and running motions) confirms significant accuracy improvements: enhanced estimations for all five muscles in the running dataset and for two of five muscles in the walking dataset. These findings demonstrate the potential of the NEM2E framework to refine OpenSim-generated muscle activity estimates and advance personalized applications in muscle activity analysis.