Epidural motor cortex stimulation (MCS) could achieve good pain control in patients with drug-resistant chronic neurogenic pain. In the search for parameters associated with the favourable outcome of this surgical procedure, quantitative sensory testing was performed in a series of 31 patients treated by MCS for chronic pain. Non-nociceptive and nociceptive sensory thresholds were measured in the painful area and its contralateral homologous zone with the stimulator in 'off' and in 'on' position. All 13 patients who exhibited normal or quite normal non-nociceptive thermal thresholds within the painful area benefited from MCS. Of the remaining 18 patients with altered thermal sensory thresholds, eight patients nevertheless experienced good pain control by MCS. In these eight 'good responders', sensory thresholds were improved by switching 'on' MCS. In contrast, the last 10 patients showed abnormal thermal thresholds that were not modified by switching 'on' MCS, and did not respond clinically to MCS. Therefore, 'good responders' to MCS could be identified by the absence of alteration of non-nociceptive sensory modalities within the painful area, or by abnormal sensory thresholds that could be improved by MCS. These results additionally suggest that MCS acts on neural pathways involved in sensory discrimination that, in turn, are able to modulate the transmission of pain signals.