Ground Reaction Forces and Moments in Stroke Survivors: Experimental versus AnyBody Model Predictions

Multibody dynamics models and simulations offer an efficient alternative to traditional methods, such as force plates, pressure sensing mats, and instrumented treadmills, for computing ground reaction forces (GRF) and moments (GRM), valuable in the quantification of the gait of neurological patients. Accurate determination of GRF and GRM magnitudes, with a specific focus on the disruptive shear components affecting gait, is essential for post-stroke gait assessment and rehabilitation. This study explored the predictive capability of musculoskeletal models equipped with foot contact elements, by comparing them with experimental data from a published dataset. The results yielded peak normalized Root Mean Square Errors (n-RMSE) of 0.51±0.31% and 0.46±0.28% for mediolateral shear components, 0.4±0.13% and 0.35±0.16% for anteroposterior shear components, and 0.34±0.16% and 0.32±0.12% for compressive components of GRF for the right and left foot respectively. For GRM, nRMSE peaks in the sagittal plane were 4.72±3.55%, followed by 2.51±2% in the frontal plane, and 2±1.44% in the transverse plane for the right foot. On the left side, nRMSE peaks were 3.73±3.12% in sagittal plane, 2.75±2.7% in frontal plane, and 2.65±2% in the transverse plane. This study underscores the potential of musculoskeletal modeling and simulation software, such as AnyBody, as a time and cost-effective alternative for evaluating the biomechanics of stroke survivors.