Effect of neural mode on model estimates of muscle contributions to knee loading in individuals with anterior cruciate ligament reconstruction

This study examined the effect of neural mode (i.e., static optimisation (SO) or Calibrated Electromyography-Informed Neuromusculoskeletal Modelling (CEINMS)) on model estimated muscle contributions to knee loading in individuals with anterior cruciate ligament reconstruction (ACLR) during side-step cutting. Eleven individuals with ACLR completed pre-planned sidestep cutting in a gait laboratory while their whole-body motions, body-ground reaction forces, and electromyography (EMG) from major knee-spanning muscles were acquired. Neuromusculoskeletal modelling simulations were performed using both SO and CEINMS methods to estimate muscle contributions to anteroposterior force, varus/valgus moments, and internal/external rotation moments at the knee. Both SO and CEINMS neural modes accurately predicted body-ground reaction forces. Moderate-to-strong coefficients of determination were found between CEINMS and SO estimates of muscle contributions to knee loads. Irrespective of neural mode, quadriceps were the dominant cause of anterior force, valgus moment, and external rotation moment at the knee, while hamstrings contributed most to the knee's posterior force and internal rotation moment. Non-knee-spanning muscles also contributed to knee loads, with small between-mode differences in these contributions. In conclusion, both neural modes well predicted whole-body mechanics (i.e., body-ground reaction forces) and differences between neural modes became more pronounced as the focus of analysis moved from general (e.g., net body-ground loads) to knee-specific mechanics.