Assessment of a two-mass ground reaction force model applied to indoor overground running in adult recreational runners

Abstract

Outdoor running kinetic measurements like vertical ground reaction force (vGRF) need simple and accurate models. A previous study assessed a two mass model (2MM) on an athletic adult population during treadmill running, but not recreational adults during overground running. The objectives were to compare accuracy of the overground 2MM and an optimized version to the reference study and force platform (FP) measurements. Overground vGRF, ankle position, and running speed were collected on 20 healthy subjects in a laboratory. The subjects ran at three self-selected speeds and with an opposite foot strike strategy. Reconstructed 2MM vGRF curves were calculated with the original parameter values (Model1), with parameters optimized each strike (ModelOpt), and with group-based optimal parameters (Model2). Root mean square error (RMSE), optimized parameters, and ankle kinematics were compared to the reference study; peak force and loading rate were compared to FP measurements. The original 2MM showed decreased accuracy with overground running. ModelOpt overall RMSE was lower than Model1 (p > 0.001, d = 3.4). ModelOpt overall peak force was different but most like FP signals (p < 0.01, d = 0.7) and Model1 was most different (p < 0.001, d = 1.3). ModelOpt overall loading rate was similar to FP signals and Model1 was different (p < 0.001, d = 2.1). Optimized parameters were different (p < 0.001) from the reference study. 2MM accuracy was largely attributable to curve parameter choice. These may be dependent on extrinsic factors like running surface and protocol and intrinsic factors like age and athletic caliber. Rigorous validation is needed if the 2MM is to be used in the field.