Open-loop control of the freely-swinging paralyzed leg.

An experimental model has been used to study issues that are relevant to the use of electrical stimulation to help paralyzed individuals walk. Modulated stimulation sequences for the quadriceps muscles were manually selected using an iterative trial-and-error procedure to cause the knee angle to follow a specific movement pattern (desired trajectory). Four paraplegic subjects were tested before and after an eight-week program in which the quadriceps were exercised daily with electrical stimulation. It was found that 12.6 +/- 2.9 iterations were required to approximate the desired trajectory. The average error of the final match between the actual and desired trajectories was 2.1 degrees +/- 0.7. Repeated responses were extremely consistent; the average difference between successive trials was less than 1 degree in 81 percent of the trials. When the stimulation sequence was repeated every 3 s for 50 cycles, however, there was a progressive degradation in the response, even in exercised legs, that demonstrated the limitations of open-loop control. Stimulus modulation envelopes for all four subjects were similar in shape (although varied in amplitude) indicating that the iterative process can be shortened by starting with an "average" modulation envelope. Stimulation sequences achieved accurate matches of the desired trajectory on subsequent days when adjusted by a simple gain factor. The relevance of these results to multichannel control of walking is discussed.