Positron emission tomography (PET) and accumulation of H(2)(15)O as a marker of neuronal activity were used to create maps of cerebral blood-flow changes evoked by painful heat stimulation in 10 subjects. Two levels of painful tonic and phasic heat stimuli were applied with use of a newly developed contact heat thermode on the volar surface of the dominant (right) arm. The subjects participated in two separate PET sessions. Maps reflecting low and high levels of painful tonic heat were obtained in the first session, and low and high levels of painful phasic heat in the second session. The subjects scored their peak pain intensity and unpleasantness on 10-cm visual analogue scales. For each subject, PET images were aligned to nuclear magnetic resonance (NMR) images and remapped into the standardized co-ordinate system of Talairach. After normalization of the PET volumes, subtraction images were formed voxel-by-voxel and converted to a t-statistic volume. The perceived pain intensity and unpleasantness were identical with painful tonic and phasic heat stimulation. Directed searches revealed significant blood-flow increases in the contralateral primary sensorimotor cortex (MI/SI), SII, insular cortex and cingulate cortex when the low tonic heat map was subtracted from the high. A similar, but not identical, pain-processing network was observed for the maps representing the subtraction of low and high phasic heat. In this subtraction, the blood-flow increases in MSI/SI did not reach statistical significance, and significant blood flow decreases were found in the contralateral middle temporal gyrus. Finally, the location of the activation site in the cingulate cortex was different from that observed during tonic heat pain. This study has provided more evidence for the existence of a common pain-processing network engaged during the perception of different levels of toxic and phasic heat pain. Copyright 1998 European Federation of Chapters of the International Association for the Study of Pain.