Hip proprioceptors preferentially modulate reflexes of the leg in human spinal cord injury

Abstract

Stretch-sensitive afferent feedback from hip muscles has been shown to trigger long-lasting, multijoint reflex responses in people with chronic spinal cord injury (SCI). These reflexes could have important implications for control of leg movements during functional activities, such as walking. Because the control of leg movement relies on reflex regulation at all joints of the limb, we sought to determine whether stretch of hip muscles modulates reflex activity at the knee and ankle and, conversely, whether knee and ankle stretch afferents affect hip-triggered reflexes. A custom-built servomotor apparatus was used to stretch the hip muscles in nine chronic SCI subjects by oscillating the legs about the hip joint bilaterally from 10° of extension to 40° flexion. To test whether stretch-related feedback from the knee or ankle would be affected by hip movement, patellar tendon percussions and Achilles tendon vibration were delivered when the hip was either extending or flexing. Surface electromyograms (EMGs) and joint torques were recorded from both legs. Patellar tendon percussions and Achilles tendon vibration both elicited reflex responses local to the knee or ankle, respectively, and did not influence reflex responses observed at the hip. Rather, the movement direction of the hip modulated the reflex responses local to the joint. The patellar tendon reflex amplitude was larger when the perturbation was delivered during hip extension compared with hip flexion. The response to Achilles vibration was modulated by hip movement, with an increased tonic component during hip flexion compared with extension. These results demonstrate that hip-mediated sensory signals modulate activity in distal muscles of the leg and appear to play a unique role in modulation of spastic muscle activity throughout the leg in SCI.

Keywords: reflexes; spasticity; spinal cord injury.

Fig. 1.

A: experimental apparatus used to impose bilateral hip oscillations and record sagittal plane torques during hip movements. A linear motor was aligned with the patellar tendon, and a small motor was strapped around the subject's ankle to vibrate the Achilles tendon. The knee was braced with padding to minimize leg movement relative to the linear motor. B: patellar tendon tapping (top) and Achilles tendon vibration (bottom) perturbations were applied every half-cycle of hip motion [i.e., either when the hip was flexing (left) or extending (right)]. Only the first 3 cycles are depicted.

Fig. 2.

Average normalized peak-to-peak (p-p) rectus femoris (RF) and vastus medialis (VM) reflex response during tendon tap percussion during hip extension (tap-ext) and hip flexion (tap-flx). *Statistical significance between comparisons (P < 0.05). Error bars indicate SE.

Fig. 3.

Peak knee extension (top) and flexion (bottom) torque during active and passive hip oscillations for tendon tapping conditions. Tendon tapping during hip extension produced larger knee extension responses in spinal cord injury (SCI) subjects than noninjured (NI) subjects. *Statistical significance (P < 0.05). Error bars indicate SE.

Fig. 4.

A: example ankle EMG data from subject S3 during passive hip oscillations with Achilles tendon vibration [extension (vibe-ext) or flexion (vibe-flx)]. Subject S3 is representative of 4 of the 9 SCI subjects for tibialis anterior (TA) EMG responses. MG, medial gastrocnemius. B: average MG EMG area of SCI subjects comparing control, vibe-ext, and vibe-flx between active (assisted) and passive tests. *Significant differences between active and passive and between test conditions (P < 0.05). Error bars indicate SE.

Fig. 5.

A: average hip torque data from 1 SCI subject (S6) and 1 NI subject representing torque responses across all test conditions during the assisted hip movements. The hip torque traces depicted were averaged per cycle across the 3 trials for each test condition. Shaded region represents flexion torque, and nonshaded region represents extension torque. B: interaction between effort and subject. SCI subjects showed higher peak torque amplitude during passive hip oscillations but lower amplitudes when assisting the hip movements (effort × subject interaction: hip extension and flexion, P < 0.001). Error bars indicate SE.

Fig. 6.

Normalized medial hamstrings (MH) EMG for SCI subjects during assisted and passive hip movements. There were no significant differences between the conditions; however, the MH EMG during the assisted conditions was greater than during the passive conditions.