BACKGROUND In clinical practice, analgesic drug doses applied during general anaesthesia are considered sufficient when clinical responses (e.g. movement, blood pressure and heart rate elevations) are suppressed during noxious stimulation. We investigated whether absent clinical responses are indicative of suppressed spinal and brain responsiveness to noxious stimulation in anaesthetised subjects. METHODS Ten healthy volunteers were investigated during deep propofol anaesthesia supplemented with increasing doses of remifentanil in a stepwise manner. Noxious electrical stimuli at an intensity comparable with surgical stimulation were repeatedly administered at each targeted remifentanil concentration. During stimulation, we monitored both clinical responses (blood pressure, heart rate, and movement) and neuronal responses. Neuronal responses were assessed using functional magnetic resonance imaging, spinal reflex responses, and somatosensory evoked potentials. RESULTS This monitoring combination was able to faithfully detect brain and spinal neuronal responses to the noxious stimulation. Although clinical responses were no longer detected at analgesic dosages similar to those used for general anaesthesia in clinical practice, spinal and brain neuronal responses were consistently observed. Opioid doses that are significantly larger than is usually used in clinical practice only reduced neuronal responses to 41% of their maximal response. CONCLUSIONS Nociceptive activation persists during deep general anaesthesia despite abolished clinical responses. Absent clinical responses are therefore not indicative of absent nociception-specific activation. Thus, commonly accepted clinical responses might be inadequate surrogate markers to assess anti-nociception during general anaesthesia. Further research is required to investigate whether persistent nociception causes adverse effects on patient outcome.