There has thought to of been a central pattern generator in humans for some time.
Wise Young has written about this in the past on his blog, the extract below is dated 2009:
Many scientists speculated that humans have a central pattern generator (CPG) for locomotion in the human spinal cord and proposed the use of spinal cord stimulation to activate locomotion. Hadi, et al. showed that destruction of interneurons neurons in the L2 spinal cord produced lasting loss of walking ability in rats, suggesting that this is where the mammalian CPG may be located. Ribotta, et al. showed that transplantation of loecus coreulus cells to provide serotonergic innervation of L1-L2 spinal cord segment restored well-defined locomotor patterns in rats. The serotonin agonist 8-OH-DPAT induces locomotor movements in low thoracic transected mice. Yakovenko, et al showed that intraspinal stimulation of spinal cords of transected rats significantly increased locomotor scores.
Dimitrijevic, et al. reported evidence of a spinal central pattern generator and showed that stimulation of this center improved gait in people after spinal cord injury. The center can be readily stimulated with epidural electrodes placed over the L2 spinal cord. The stimulation initiates extension of lower limb and can produce stepping-like movements in subjects with complete spinal cord injury. In both complete and incomplete spinal cord injury, reorganization of the spinal cord circuitry provides alternative mechanisms of locomotor and motor control. In 2007, Mianssian, et al. found that human lumbar cord circuitry can be activated by extrinsic tonic input to generate locomotor-like activity. Early application both FES and CPG stimulation may be helpful in preventing disuse atrophy of both muscle and neurons.
CPG stimulation helps locomotor training. Barriere, et al. pointed out the important role of the central pattern generator in the recovery of voluntary locomotion. Herman, et al. has shown that epidural stimulation of the T10/T12 central pattern generator in humans facilitates walking recovery in people with chronic spinal cord injury, improving walking speed and endurance while reducing the sense of effort by the subject. In particular, Carhart, et al. described a patient with chronic incomplete tetraplegia in whom partial weight-bearing therapy alone was insufficient to achieve functional ambulation. However, application of the T10/T12 stimulation not only produced further improvements in treadmill walking but facilitated transfer of these gains to overground walking. The participant initially reported a reduction in the sense of effort for overground walking from 8/10 to 3/10 (Borg scale) and was able to double his walking speed to 0.35 m/sec and distance to 325 m.
Spinal cord stimulation facilitates walking not by directly activating neurons that control walking but by lowering the threshold to activate walking and to recruit additional muscles that may not be under voluntary control. The stimulation not only increases the efficiency and speed of walking but reduces both perceived and actual effort required for walking, including metabolic energy utilization during walking. Recent studies suggest that the electrodes don’t even have to be placed epidurally but can activate the spinal cord when placed on skin overlying the lumbar cord. Note that rhythmic auditory stimulation likewise has strong entrainment effects on gait cadence, velocity, and stride length in patients with incomplete spinal cord injury.