We investigated whether somatosensory feedback from contracting limb muscles exerts an inhibitory influence on the determination of central command during closed-loop cycling exercise in which the subject voluntarily determines his second-by-second central motor drive. Eight trained cyclists performed two 5 km time trials either without (5K_Ctrl) or with lumbar epidural anesthesia (5K_Epi; 24 ml of 0.5% lidocaine, vertebral interspace L3-L4). Percent voluntary quadriceps muscle activation was determined at rest using a superimposed twitch technique. Epidural lidocaine reduced pre-time trial maximal voluntary quadriceps strength (553 ± 45 N; MVC) by 22 ± 3%. Percent voluntary quadriceps activation was also reduced from 97 ± 1% to 81 ± 3% via epidural lidocaine and this was unchanged following the 5K_Epi indicating a sustained level of neural impairment throughout the trial. Power output was reduced by 9 ± 2% throughout the race (P < 0.05). We found three types of significant effects of epidural lidocaine which supported a substantial role for somatosensory feedback from the exercising limbs as a determinant of central command throughout high intensity closed-loop cycling exercise: a) relative integrated electromyogram of the vastus lateralis was significantly increased; b) pedal forces were similar despite reduced number of fast-twitch muscle fibers available for activation; c) ventilation was increased out of proportion to a reduced VCO₂ and heart rate and blood pressure were increased out of proportion to power output and VO₂. These findings demonstrate the inhibitory influence of somatosensory feedback from contracting locomotor muscles on the conscious and/or subconscious determination of the magnitude of central motor drive during high intensity closed-loop endurance exercise.