We investigated whether experimentally determined, suprathreshold pain sensitivity of pericranial musculature in patients with tension-type headache differs from that of migraine patients or from that of healthy subjects. Furthermore, we looked to see whether differences could be found in the effects of experimental pain induction on EMG activity of pericranial musculature and whether subgroups could be discovered with higher and lower pericranial pain sensitivity within the three diagnostic groups in terms of neurophysiological, psychological and clinical variables. In 20 patients with tension-type headache, 23 patients with migraine without aura, and 29 healthy individuals experimental pain was induced in the temporal muscle by mechanical pressure; pain sensitivity in the entire metrically subdivided suprathreshold pain sensitivity range was measured. Surface EMG activity of pericranial muscles was determined before, during and after experimental pain induction. In addition, headache characteristics as well as personality and mood states were determined and recorded in a standardized fashion. There were no significant differences in pain sensitivity of pericranial musculature between the three groups. Patients with tension-type headache showed significantly higher EMG scores during suprathreshold pain stimulation than did migraine patients. EMG scores of healthy subjects fell between these two groups. With respect to pericranial tenderness significant differences in clinical, neurophysiological and psychological variables were found only between subgroups within the group of patients with tension-type headache. The results indicate that significant differences in the examined groups are found not in pain perception but in the processing or reaction to experimental headache stimuli. In patients with tension-type headache subgroups evolve based on pericranial pain sensitivity with quantitatively and/or qualitatively impaired reactions; for this reason diagnostic grouping according to the IHS classification seems to be pathophysiologically relevant. The intraindividual phasic comparison of pain reactions appears to be more important than the absolute interindividual tonic comparison.