Heat and electrical detection thresholds were assessed in 72 patients suffering from painful temporomandibular disorder. Employing widely accepted criteria, 44 patients were classified as suffering from temporomandibular joint (TMJ) arthralgia (i.e. pain originating from the TMJ) and 28 from myalgia (i.e. pain originating from the muscles of mastication). Electrical stimulation was employed to assess thresholds in large myelinated nerve fibers (Abeta) and heat application to assess thresholds in unmyelinated nerve fibers (C). The sensory tests were performed bilaterally in three trigeminal nerve sites: the auriculotemporal nerve territory (AUT), buccal nerve territory (BUC) and the mental nerve territory (MNT). In addition, 22 healthy asymptomatic controls were examined. A subset of ten arthralgia patients underwent arthrocentesis and electrical detection thresholds were additionally assessed following the procedure. Electrical detection threshold ratios were calculated by dividing the affected side by the control side, thus reduced ratios indicate hypersensitivity of the affected side. In control patients, ratios obtained at all sites did not vary significantly from the expected value of 'one' (mean with 95% confidence intervals; AUT, 1:0.95-1.06; BUC, 1.01:0.93-1.11; MNT, 0.97:0.88-1.05, all areas one sample analysis P>0.05). In arthralgia patients mean ratios (+/-SEM) obtained for the AUT territory (0.63+/-0.03) were significantly lower compared to ratios for the MNT (1.02+/-0.03) and BUC (0.96+/-0.04) territories (repeated measures analysis of variance (RANOVA), P<0.0001) and compared to the AUT ratios in myalgia (1.27+/-0.09) and control subjects (1+/-0.06, ANOVA, P<0.0001). In the myalgia group the electrical detection threshold ratios in the AUT territory were significantly elevated compared to the AUT ratios in control subjects (Dunnett test, P<0.05), but only approached statistical significance compared to the MNT (1.07+/-0.04) and BUC (1.11+/-0.06) territories (RANOVA, F(2,27)=3.12, P=0.052). There were no significant differences between and within the groups for electrical detection threshold ratios in the BUC and MNT nerve territories, and for the heat detection thresholds in all tested sites. Following arthrocentesis, mean electrical detection threshold ratios in the AUT territory were significantly elevated from 0.64+/-0.06 to 0.99+/-0.04 indicating resolution of the hypersensitivity (paired t-test, P=0.001). In conclusion, large myelinated fiber hypersensitivity is found in the skin overlying TMJs with clinical pain and pathology but is not found in controls. In patients with muscle-related facial pain there was significant elevation of the electrical detection threshold in the AUT region.