Comparison of Regression Models for Estimation of Isometric Wrist Joint Torques Using Surface Electromyography

Background: Several regression models have been proposed for estimation of isometric joint torque using surfaceelectromyography (SEMG) signals. Common issues related to torque estimation models are degradation of modelaccuracy with passage of time, electrode displacement, and alteration of limb posture. This work compares theperformance of the most commonly used regression models under these circumstances, in order to assistresearchers with identifying the most appropriate model for a specific biomedical application.Methods: Eleven healthy volunteers participated in this study. A custom-built rig, equipped with a torque sensor,was used to measure isometric torque as each volunteer flexed and extended his wrist. SEMG signals from eightforearm muscles, in addition to wrist joint torque data were gathered during the experiment. Additional data weregathered one hour and twenty-four hours following the completion of the first data gathering session, for thepurpose of evaluating the effects of passage of time and electrode displacement on accuracy of models. AcquiredSEMG signals were filtered, rectified, normalized and then fed to models for training.Results: It was shown that mean adjusted coefficient of determination (R2a) values decrease between 20%-35% fordifferent models after one hour while altering arm posture decreased mean R2avalues between 64% to 74% fordifferent models.Conclusions: Model estimation accuracy drops significantly with passage of time, electrode displacement, andalteration of limb posture. Therefore model retraining is crucial for preserving estimation accuracy. Data resamplingcan significantly reduce model training time without losing estimation accuracy. Among the models compared,ordinary least squares linear regression model (OLS) was shown to have high isometric torque estimation accuracycombined with very short training times.