Distribution and latency of muscle responses to transcranial magnetic stimulation of motor cortex after spinal cord injury in humans.

Noninvasive transcranial magnetic stimulation (TMS) of the motor cortex was used to evoke electromyographic (EMG) responses in persons with spinal cord injury (n = 97) and able-bodied subjects (n = 20, for comparative data). Our goal was to evaluate, for different levels and severity of spinal cord injury, potential differences in the distribution and latency of motor responses in a large sample of muscles affected by the injury. The spinal cord injury (SCI) population was divided into subgroups based upon injury location (cervical, thoracic, and thoracolumbar) and clinical status (motor-complete versus motor-incomplete). Cortical stimuli were delivered while subjects attempted to contract individual muscles, in order to both maximize the probability of a response to TMS and minimize the response latency. Subjects with motor-incomplete injuries to the cervical or thoracic spinal cord were more likely to demonstrate volitional and TMS-evoked contractions in muscles controlling their foot and ankle (i.e., distal lower limb muscles) compared to muscles of the thigh (i.e., proximal lower limb muscles). When TMS did evoke responses in muscles innervated at levels caudal to the spinal cord lesion, response latencies of muscles in the lower limbs were delayed equally for persons with injury to the cervical or thoracic spinal cord, suggesting normal central motor conduction velocity in motor axons caudal to the lesion. In fact, motor response distribution and latencies were essentially indistinguishable for injuries to the cervical or thoracic (at or rostral to T10) levels of the spine. In contrast, motor-incomplete SCI subjects with injuries at the thoracolumbar level showed a higher probability of preserved volitional movements and TMS-evoked contractions in proximal muscles of the lower limb, and absent responses in distal muscles. When responses to TMS were seen in this group, the latencies were not significantly longer than those of able-bodied (AB) subjects, strongly suggestive of "root sparing" as a basis for motor function in subjects with injury at or caudal to the T11 vertebral body. Both the distribution and latency of TMS-evoked responses are consistent with highly focal lesions to the spinal cord in the subjects examined. The pattern of preserved responsiveness predominating in the distal leg muscles is consistent with a greater role of corticospinal tract innervation of these muscles compared to more proximal muscles of the thigh and hip.