The aim of this study was to investigate the contribution of proprioceptive feedback to the activation of the soleus muscle during the stance phase of walking in spastic patients. Twelve hemiparetic spastic patients (cerebral stroke) and 12 age-matched able-bodied controls participated in the study. Antispastic medication was interrupted 24 hr before the test and the spasticity level was evaluated with the Ashworth score. All subjects walked on a treadmill at a comfortable speed (~ 1 km/h) with the left leg attached to a robotic actuator that can dorsiflex and plantarflex the ankle. Fast plantar flexion perturbations (6 deg, 500 deg/s) were applied in the mid-stance phase to investigate the contribution of feedback from ankle plantar flexors to the soleus EMG. To investigate the afferent-mediated modulation of the soleus EMG, the ankle dorsiflexion during the stance phase was slightly enhanced or reduced mimicking variations in the ankle kinematics during normal walking. Four levels of ankle dorsiflexion enhancements and reductions (± 2 deg/± 6 deg/s; ± 4 deg/± 10 deg/s) were applied 250 ms after heel contact (300 ms duration). A control step was recorded before each perturbed step (25 each). MANOVA tests were used to compare responses between the groups and regression analysis was used to test the relationship between the velocity of the ankle dorsiflexion and the soleus EMG. The amplitude of the short latency stretch reflex response was significantly increased in the patients compared with the controls (p=0.02). The fast plantar flexion perturbations produced a drop in the soleus EMG that was significantly larger in the healthy volunteers than the patients (p=0.03). The slow-velocity dorsiflexion enhancements and reductions generated in the control subjects’ increments and decrements, respectively, in the soleus EMG. However, in the spastic patients these responses were decreased. Moreover, there was an inverse relationship between the Ashworth scale and the amplitude of the responses to the slow perturbations (slope of the regression lines). There was no relationship between walking speed and the slope of the regression lines. These results confirm that afferent feedback contributes to the soleus activity during the stance phase of human walking, and that afferent feedback enhances and modulates this activity according to the ankle kinematics. While the short latency soleus stretch reflex was hyper-excited in the patients, the afferent feedback contribution to the soleus activity during spastic walking was depressed. We speculate that part of the walking impairment in spastic patients is caused by either reduced afferent feedback or impaired integration of the afferent feedback within CNS.