Epidural motor cortex stimulation (MCS) has been proposed as a treatment for chronic, drug-resistant neuropathic pain of various origins. Regarding pain syndromes due to peripheral nerve lesion, only case series have previously been reported. We present the results of the first randomized controlled trial using chronic MCS in this indication. Sixteen patients were included with pain origin as follows: trigeminal neuralgia (n = 4), brachial plexus lesion (n = 4), neurofibromatosis type-1 (n = 3), upper limb amputation (n = 2), herpes zoster ophthalmicus (n = 1), atypical orofacial pain secondary to dental extraction (n = 1) and traumatic nerve trunk transection in a lower limb (n = 1). A quadripolar lead was implanted, under radiological and electrophysiological guidance, for epidural cortical stimulation. A randomized crossover trial was performed between 1 and 3 months postoperative, during which the stimulator was alternatively switched 'on' and 'off' for 1 month, followed by an open phase during which the stimulator was switched 'on' in all patients. Clinical assessment was performed up to 1 year after implantation and was based on the following evaluations: visual analogue scale (VAS), brief pain inventory, McGill Pain questionnaire, sickness impact profile and medication quantification scale. The crossover trial included 13 patients and showed a reduction of the McGill Pain questionnaire-pain rating index (P = 0.0166, Wilcoxon test) and McGill Pain questionnaire sensory subscore (P = 0.01) when the stimulator was switched 'on' compared to the 'off-stimulation' condition. However, these differences did not persist after adjustment for multiple comparisons. In the 12 patients who completed the open study, the VAS and sickness impact profile scores varied significantly in the follow-up and were reduced at 9-12 months postoperative, compared to the preoperative baseline. At final examination, the mean rate of pain relief on VAS scores was 48% (individual results ranging from 0% to 95%) and MCS efficacy was considered as good or satisfactory in 60% of the patients. Pain relief after 1 year tended to correlate with pain scores at 1 month postoperative, but not with age, pain duration or location, preoperative pain scores or sensory-motor status. Although the results of the crossover trial were slightly negative, which may have been due to carry-over effects from the operative and immediate postoperative phases, observations made during the open trial were in favour of a real efficacy of MCS in peripheral neuropathic pain. Analgesic effects were obtained on the sensory-discriminative rather than on the affective aspect of pain. These results suggest that the indication of MCS might be extended to various types of refractory, chronic peripheral pain beyond trigeminal neuropathic pain.