Cerebral generators of long latency brain potentials in response to painful heat stimuli were identified from potential distributions in 31 EEG leads, using the brain electrical source analysis (BESA) programme in the multiple spatio-temporal dipole mode. Data were taken from a study with 10 young healthy male subjects who participated in 3 identical sessions, 1 week apart, with 4 blocks of 40 stimuli (randomized intensities above mean pain threshold). Brief infrared laser heat pulses were applied to the right temple; laser evoked brain potentials (LEPs) were averaged over 40 stimuli per block. BESA was applied to the grand mean maps averaged over the 10 subjects, 3 sessions and 4 stimulus blocks per session, as well as to the individual maps. In all cases 4 generators could consistently be identified by BESA, which were able to explain up to 98.8% of the total variance in scalp distributions at certain time intervals: dipole I with a maximum activity at 106.3 msec in the contralateral somatosensory trigeminal cortex, 19.0 mm beneath the surface; dipole II with a maximum activity at 112.1 msec at the corresponding ipsilateral area at a depth of 13.6 mm; dipole III with a maximum activity at 130.4 msec in the frontal cortex; dipole IV with 2 relative maximum activities at 150.6 and 220.5 msec, localized centrally under the vertex at a depth of 33.1 mm, which described both the late vertex negativity and the consecutive positivity. BESA applied to the individual LEP maps of each individual and session yielded again 4 major generators with sites, strengths and orientations comparable to those of the grand mean evaluations. The standard deviation (S.D.) of site coordinates within subjects was less than 3 mm for dipoles I, II and IV (5 mm for dipole III). The between-subject standard deviation was considerably larger (15 mm), which was attributed to individual differences in head geometry, size and anatomy. Dipoles I and II are assumed to be generators in secondary somatosensory areas of the trigeminal nerve system with bilateral representation, though significantly stronger in the contralateral site. Dipole III in the frontal cortex may be related to attention and arousal processes, as well as to motor cortical initiation for eye movements and muscle effects. The central dipole IV describing all late activity between 150 and 220 msec is probably a representative of perceptual activation and cognitive information processing; it was located in deep midline brain structure, e.g., the cingular gyrus.