The temporal and spatial processing of pain perception in human was tr
aced by magnetoencephalography (MEG). We applied a painful CO2 laser b
eam to the forearm of 11 normal subjects, and estimated the activated
areas using a single equivalent current dipole (ECD) at each time poin
t, and a brain electric source analysis (BESA) as a spatio-temporal mu
ltiple source analysis method. The four-source model was found to be t
he most appropriate; sources 1 and 2 at the secondary sensory cortex (
SII) contralateral and ipsilateral to the stimulation, and sources 3 a
nd 4 at the anterior medial temporal area (probably the amygdalar nucl
ei or hippocampal formation) contralateral and ipsilateral to the stim
ulation, respectively. Activities in all 4 areas were temporally overl
apped. Activity in the primary sensory cortex (SI) contralateral to th
e stimulated site was not identified. Activity in the cingulate cortex
was also not clearly identified. These results are probably due to on
e or more of the following factors; (1) the cingulate cortex is too de
ep, (2) the ECDs generated in the cingulate cortex are mainly oriented
radially, and (3) the ECDs generated in bilateral hemispheres interfe
re with each other. No significant or consistent magnetic fields were
recorded after 500 msec following the stimulation, probably due to the
complicated spatial and temporal overlapping of activities in multipl
e areas.