Somatosensory evoked potentials (SEPs) of the median nerve were recorded in 40 patients with the tonic and tremulous form of torticollis and in 40 healthy volunteers. Polymyographic recordings of the activity of cervical muscles were performed in all patients with cervical dystonia to determine the dystonic and antagonistic muscles. Patient SEPs were recorded during abnormal head movement. SEPs in 20 healthy volunteers were recorded with the head in the middle position. SEPs in another 20 healthy volunteers were recorded with the head rotated 60 degrees to the right. The mean peak-to-peak amplitude values of the precentral P22/N30 and the postcentral N20/P25 complexes and their mean side-to-side ratios were calculated in the F3 (F4), C3' (C4'), and C3+ (C4+) electrode positions in all four groups. In patients with the tonic form of torticollis (group I), an apparent mean P22/N30 amplitude increase was found above the hemisphere contralateral to the direction of head deviation in both precentral electrode positions, F3(4) and C3(4)'. A statistically significant difference was observed between group I and other patient and control groups. In patients with the tremulous form of torticollis (group II), an increase in the mean P22/N30 amplitude was found above both hemispheres in both precentral electrode positions F3(4) and C3(4)'; a significant difference was found between group II and both control groups. Lateralization of the P22/N30 component was found only in patients with the tonic form of torticollis. The mean side-to-side ratio of the precentral P22/N30 component amplitude was significantly different when group I was compared with either group II or control groups. No significant difference between group II and either control group was found. No significant abnormalities in the postcentral N20/P25 component were found in either the dystonic patients or in healthy control subjects. These results might indicate a different pattern of cortex excitability in patients with tonic versus tremulous forms of torticollis and therefore may implicate different underlying pathophysiological mechanisms in these two forms of disorder.