To reveal areas in the central nervous system of importance for electrodermal control, regional cerebral blood flow (rCBF) was correlated to nonspecific skin conductance fluctuations (NSF) during aversive and nonaversive conditions. Participants viewed a TV screen displaying white noise or snake videotapes presented both with and without electric shocks given to the right hand. H215O positron emission tomography was used to measure rCBF, and the constant voltage technique was used to measure NSF from the left hand. Electrodermal activity was positively related to rCBF in the left primary motor cortex (MI, Brodmann's Area 4) and bilaterally in the anterior (Areas 24 and 32) and posterior cingulate cortex (Area 23). Negative relations were observed bilaterally in the secondary visual cortex (Areas 18 and 19) and the right inferior parietal cortex (Area 39), with a tendency also for the right insular cortex (Areas 13, 15, and 16). Because results from lesion and stimulation studies in humans converge with the present imaging results, we conclude that the cingulum and the motor cortex, in addition to the parietal and possibly the insular cortex, form part of one or several distributed neural network(s) involved in electrodermal control. Because these areas also support anticipation, affect, and locomotion, electrodermal responses seem to reflect cognitively or emotionally mediated motor preparation.