This study investigates whether use of muscle excitation controls constructed from subjectspecific muscle synergy information can improve optimization prediction of knee contact forces and muscle excitations during walking. Muscle synergies quantify how a large number of experimental muscle electromyographic (EMG) signals can be reconstructed by linearly mixing a much smaller number of neural commands generated by the nervous system. Our hypothesis was that controlling all muscle excitations with a small set of experimentally calculated neural commands would improve prediction of knee contact forces and leg muscle excitations compared to using independently controlled muscle excitations.