Studies suggest that the human lumbosacral spinal cord can generate st
eplike oscillating electromyographic (EMG) patterns, but it remains un
clear to what degree these efferent patterns depend on the phasic peri
pheral sensory information associated with bilateral limb movements an
d loading. We examined the role of sensory information related to lowe
r-extremity weight bearing in modulating the efferent motor patterns o
f spinal-cord-injured (SCI) subjects during manually assisted stepping
on a treadmill. Four nonambulatory subjects, each with a chronic thor
acic spinal cord injury, and two nondisabled subjects were studied. Th
e level of loading, EMG patterns, and kinematics of the lower limbs we
re studied during manually assisted or unassisted stepping on a treadm
ill with body weight support. The relationships among lumbosacral moto
r pool activity [soleus (SOL), medial gastrocnemius (MG), and tibialis
anterior (TA)], limb load, muscle-tendon length, and velocity of musc
le-tendon length change were examined. The EMG mean amplitude of the S
OL, MG, and TA was directly related to the peak load per step on the l
ower Limb during locomotion. The effects on the EMG amplitude were qua
litatively similar in subjects with normal, partial, or no detectable
supraspinal input. Responses were most consistent in the SOL and MG at
load levels of <50% of a subject's body weight. The modulation of the
EMG amplitude from the SOL and MG, both across steps and within a ste
p, was more closely associated with limb peak load than muscle-tendon
stretch, or the velocity of muscle-tendon stretch. Thus stretch reflex
es were not the sole source of the phasic EMG activity in flexors and
extensors during manually assisted stepping in SCI subjects. The EMG a
mplitude within a step was highly dependent on the phase of the step c
ycle regardless of level of load. These data suggest that level of loa
ding on the lower limbs provides cues that enable the human lumbosacra
l spinal cord to modulate efferent output in a manner that may facilit
ate the generation of stepping. These data provide a rationale for gai
t rehabilitation strategies that utilize the level of load-bearing ste
pping to enhance the locomotor capability of SCI subjects.