We analyzed the twitch and summated torque (tetanus) during repetitive activation and recovery of the human soleus muscle in individuals with spinal cord injury. Thirteen individuals with complete paralysis (9 chronic, 4 acute) had the tibial nerve activated every 1,500 ms with a 20-Hz train (7 stimuli) for 300 ms and a single pulse at 1,100 ms. The stimulation protocol lasted 3 min and included 120 twitches and 120 tetani. Minimal changes were found for the acute group. The chronic group showed a significant reduction in the torque and a significant slowing of the contractile speeds of both the twitch and tetanus. The decrease in the peak twitch torque was significantly greater than the decrease in the peak tetanus torque early during the fatigue protocol for the chronic group. The twitch time to peak and half relaxation time were prolonged during fatigue, which was associated with improved fusion of the tetanus torque. At the end of the fatigue protocol, the decrease in the peak twitch torque was not significantly different from the decrease in the peak tetanus torque. After 5 min of rest, the contractile speeds recovered causing the tetanus to become unfused, but the tetanus torque became less depressed than the twitch torque. The differential responses for the twitch and the tetanus suggest an interplay between optimal fusion created from contractile speed slowing and excitation contraction coupling compromise. These issues make the optimal design of functional electrical stimulation systems a formidable task.