UNLABELLED: This study was designed to assess the value of repetitive transcranial magnetic stimulation (rTMS) to predict the efficacy of epidural motor cortex stimulation (EMCS) to treat neuropathic pain. We have included 59 patients treated by EMCS for more than 1 year and in whom active and sham 10Hz-rTMS sessions were performed as preoperative tests, targeted over the cortical representation of the painful area. Analgesic effects were rated on a visual analogue scale. The real rTMS efficacy was determined by subtracting the effect of the sham stimulation on pain scores from that of the active stimulation (active-sham calculation). Pain scores were significantly reduced by active rTMS and EMCS, but not by sham rTMS. Twenty-six of the 33 patients (79%) who responded to active rTMS and all the 21 patients (100%) who responded for active-sham calculation also responded to EMCS. The response observed in active-sham calculation had a positive predictive value of 1.0, but a negative predictive value of .6 regarding EMCS outcome. The analgesic effect of rTMS or EMCS was not influenced by the side, origin, or duration of pain or by the presence of motor or sensory deficit in the painful area. Poorer results were observed in case of lower limb pain for rTMS and in older patients for EMCS. This study confirms that neuropathic pain can be significantly relieved by motor cortex rTMS or EMCS. A positive outcome of EMCS can be predicted by a real response to rTMS, but not on clinical grounds. PERSPECTIVE: Single sessions of sham-controlled preoperative rTMS tests can be used to confirm the indication of EMCS therapy but have no value to exclude patients from this therapy. New rTMS protocols remain to be assessed to improve the usefulness of preoperative rTMS in EMCS practice.