Seventy-eight patients with confirmed primary cervical dystonia CD were studied. All patients underwent CT of the soft tissues of the neck using slices at cervical vertebrae and MRI images of the cervical spine and of soft tissues. MRI images of 50 patients who did not have CD were used for comparison. This was followed by measuring the largest diameter along with the description of the shape of all observable muscles including the small muscles of the occipital area. In lateral flexion and rotation, 19% of patients showed disorders of muscles acting on head joints (laterocaput/torticaput). Muscles that act on the cervical spine were affected (laterocollis/torticollis) in 20% of patients. Both types of the disorder, but with various degrees of the caput- and collis- involvement, were presented in 61% of patients. Consequently, the ratio for these forms was approximately estimated as 1:1:3. The following conclusions have been made: In lateral flexion, clinical differentiation between laterocollis and laterocaput is possible. Lateral shift is always a result of laterocollis on one side and laterocaput on the opposite side. In rotation, clinical differentiation between torticollis and torticaput is not always possible. CT slices at levels C1 and C2 are advisable in these cases. Comparing the positions of vertebrae on both levels will provide a safe differentiation between torticollis and torticaput. Analysis of forward flexion (differentiation between antecollis and antecaput) can be accomplished by lateral observation of the angles between the cervical spine and the thoracic spine, respectively, and between the cervical spine and the base of skull. The same applies to the analysis of backward flexion (differentiation between retrocollis and retrocaput). Sagittal shift forwards usually does not require further diagnosis: it is almost always caused by bilateral dystonic activities of the Mm. sternocleidomastoidei.