Modulation of Locomotor-Like EMG Activity in Subjects with Complete and Incomplete Spinal Cord Injury

Bruce Dobkin, MD, Susan Harkema, Ph.D., Phil Requejo, BA, Reggie Edgerton, Ph.D.

Summary by Patty Davis, MS, Mobility Research, 20 March 1997

The use of body weight supported treadmill training in the rehabilitation of spinal cord injured (SCI) patients is rapidly gaining credibility as an effective rehabilitation protocol. Early animal research found the development of rhythmic nerve potentials in isolated lumbar cords during forced stepping. These rhythmic patterns were attributed to central pattern generators for stepping, in response to the forced locomotion produced by a motorized treadmill. The promising results of body weight supported training in animals prompted the investigation of the effects and/or benefits of supported treadmill training in the human SCI patient populations.

Past research using body weight supported treadmill training (BWSTT) has described various aspects of EMG activity seen during the experimentation. These studies, however, have not compared the changes in EMG activity of complete and incomplete chronic spinal cord injured subjects during motorized treadmill training, nor have these studies discussed how EMG bursts are altered by specific sensory inputs associated with stepping. The purpose of this study was, therefore, to determine if automatic stepping patterns or step-like EMG bursts could be generated and consequently regulated by sensory feedback such as limb loading and proprioception, during body weight supported treadmill training.

Five mid-thoracic spinal cord injured subjects, manifesting complete and chronic sensory and motor loss, were uses as one test group for the study. All of the above mentioned subjects were classified Grade A on the American Spinal Injury Association Impairment (ASIAI) Scale. Also, four subjects having incomplete lesions were used for this study. One of the four subjects was classified Grade B and the other three subjects were classified Grade C on the American Spinal Injury Association Impairment Scale. (The ASIAI Scale grades five key muscle groups of each leg from L2 to S1 root.)

A suspension apparatus was used to support an average of 45-50% of the subject's body weight, thus allowing for maximal weight bearing by the fully extended knee in the stance phase, without causing a corresponding buckling of the knee joint. To mimic the normal gait cycle as closely as possible, trainers manually assisted the stepping pattern and modified joint angles when needed. Treadmill speeds ranged from 0.5 - 1.0 mph.

When looking at the data collected for both complete and incomplete SCI subjects, the timing of EMG bursts were consistent with the gait cycle period, and increased as the treadmill speed increased. Throughout the training sessions, the incomplete SCI subjects became more able to self-initiate the swing phase of stepping. The amplitude of the EMG bursts were found to be independent of the direction and magnitude of the joint motion and muscle length. The EMG activity seen, therefore, was thought to be independent of the stretch reflex elicited during BWSTT.

This study supports the results of previous studies which found that when provided with sensory feedback such as limb loading and proprioception, the human lumbosacral spinal cord can generate rhythmic locomotor EMG patterns, even in the absence of supraspinal influences. It is thought that bilateral load-bearing stepping provides segmental sensory inputs from the lower limbs. A series of responsive circuits then generates the rhythmic flexor and extensor EMG activity. Furthermore, this study provides evidence of the significance of sensory parameters such as treadmill speed, joint loading and hip position, which can alter the amplitude and duration of the EMG bursts. The results of this study support the idea that body weight supported treadmill training is a valuable, task-specific approach to the rehabilitation of complete and incomplete spinal cord injured patients.