Temporal and intensity coding of pain in human cortex

We used a high-resolution functional magnetic resonance imaging (fMRI) technique in healthy right-handed volunteers to demonstrate cortical areas displaying changes of activity significantly related to the time profile of the perceived intensity of experimental somatic pain over the course of several minutes. Twenty-four subjects (ascorbic acid group) received a subcutaneous injection of a dilute ascorbic acid solution into the dorsum of one foot, inducing prolonged burning pain (peak pain intensity on a 0-100 scale: 48 +/- 3, mean +/- SE; duration: 11.9 +/- 0.8 min). fMRI data sets were continuously acquired for similar to 20 min, beginning 5 min before and lasting 15 min after the onset of stimulation, from two sagittal planes on the medial hemispheric wall contralateral to the stimulated site, including the cingulate cortex and the putative foot representation area of the primary somatosensory cortex (SI). Neural clusters whose fMRI signal time courses were positively or negatively correlated (P < 0.0005) with the individual pain intensity curve were identified by cross-correlation statistics in all 24 volunteers. The spatial extent of the identified clusters was linearly related (P < 0.0001) to peak pain intensity. Regional analyses showed that positively correlated clusters were present in the majority of subjects in ST, cingulate, motor, and premotor cortex. Negative correlations were found predominantly in medial parietal, perigenual cingulate, and medial prefrontal regions. To test whether these neural changes were due to aspecific arousal or emotional reactions, related either to anticipation or presence of pain, fMRI experiments were performed with the same protocol in two additional groups of volunteers, subjected either to subcutaneous saline injection (saline: n = 16), inducing mild short-lasting pain (peak pain intensity 23 +/- 4; duration 2.8 +/- 0.6 min) or to nonnoxious mechanical stimulation of the skin (controls: n = 16) at the same body site. Subjects did not know in advance which stimulus would occur. The spatial extent of neural clusters whose signal time courses were positively or negatively correlated with the mean pain intensity curve of subjects injected with ascorbic acid was significantly larger (P < 0.001) in the ascorbic acid group than both saline and controls, suggesting that the observed responses were specifically related to pain intensity and duration. These findings reveal distributed cortical systems, including parietal areas as well as cingulate and frontal regions, involved in dynamic encoding of pain intensity over time, a process of great biological and clinical relevance.