The pain matrix and neuropathic pain

Two articles published in this and a recent edition of Brain (Iadarola et al., 1998; Peyron et al., 1998) raise some issues about what functional imaging studies can teach us about nociceptive and non-nociceptive processing in the brain. Both papers address whether brain structures activated during neuropathic and acute experimental nociceptive pain are different. In Peyron et al.'s paper, the neuropathic pain is clinical mechanical allodynia resulting from a brainstem lesion, and in Iadorola et al.'s paper it is capsaicin-induced allodynia in normal volunteers. Both studies suggest similar brain structures are involved in processing the two types of pain. However, Iadorola et al.'s study suggests that inferior prefrontal cortex may be 'uniquely' involved in processing allodynic pain. Until relatively recently, our understanding of the functional anatomy of human cerebral nociceptive processing was based on post-mortem anatomical studies (Bowsher et al., 1957), effects of cortical and subcortical stimulation during crani-otomy procedures, and lesion data Albe-Fessard et al. (1985) postulated a lateral pain system based on the cortical projections of the lateral thalamic nuclei to the primary somatosensory cortex and a medial pain system based on the medial thalamic group of nuclei with unspecified cortical projections. He proposed a division of function in which the lateral, fast and somatotopic system processes acute nociceptive inputs and the medial system processes chronic nociceptive inputs. However, failure to elicit anything resembling pain during stimulation of the somatosensory cortex, Penfield and Boldrey (1937) raised the possibility that pain was not processed cortically at all. In 1992, Sikes and Vogt identified a nociceptive region of Brodman area 24 of the anterior cin-gulate cortex, and suggested that this was one of the main cortical projections of the medial pain system. Since then the key role of the cingulate cortex in the integration of sensory, affective and motivational processes has been recognized (Devinsky et al., 1995). Human PET studies have confirmed Vogt and colleagues' original finding and suggest that the cingulate cortex is involved in nociceptive processing related to both acute phasic and tonic experimental and chronic pain (Jones et al. The other brain structures involved are the periaqueductal grey, thalamus, lentiform nucleus, insula, anterior cingulate (areas 24, 25, 32 and 24Ј, 32Ј) and prefrontal cortex (areas 9, 10, 44), inferior parietal cortex (area 39/40) and, somewhat variably, primary and secondary somatosensory cortices. How do these areas contribute to the experience and behavioural consequences of pain? At first glance …