Indirect evidence suggests that the thalamus contributes to abnormal movements occurring in patients with dystonia (dystonia patients). The present study tested the hypothesis that thalamic activity contributes to the dystonic movements that occur in such patients. During these movements, spectral analysis of electromyographic (EMG) signals in flexor and extensor muscles of the wrist and elbow exhibited peak EMG power in the lowest frequency band [0-0.78 Hz (mean: 0.39 Hz) dystonia frequency] for 60-85% of epochs studied during a pointing task. Normal controls showed low-frequency peaks for <16% of epochs during pointing. Among dystonia patients, simultaneous contraction of antagonistic muscles (cocontraction) at dystonia frequency during pointing was observed for muscles acting about the wrist (63% of epochs) and elbow (39%), but cocontraction was not observed among normal controls during pointing. Thalamic neuronal signals were recorded during thalamotomy for treatment of dystonia and were compared with those of control patients without motor abnormality who were undergoing thalamic procedures for treatment of chronic pain. Presumed nuclear boundaries of a human thalamic cerebellar relay nucleus (ventral intermediate, Vim) and a pallidal relay nucleus (ventral oral posterior, Vop) were estimated by aligning the anterior border of the principal sensory nucleus (ventral caudal, Vc) with the region where the majority of cells have cutaneous receptive fields (RFs). The ratio of power at dystonia frequency to average spectral power was >2 (P < 0.001) for cells in presumed Vop often for dystonia patients (81%) but never for control patients. The percentage of such cells in presumed Vim of dystonia patients (32%) was not significantly different from that of controls (31%). Many cells in presumed Vop exhibited dystonia frequency activity that was correlated with and phase-advanced on EMG activity during dystonia, suggesting that this activity was related to dystonia. Thalamic somatic sensory activity also differed between dystonia patients and controls. The percentage of cells responding to passive joint movement or to manipulation of subcutaneous structures (deep sensory cells) in presumed Vim was significantly greater in patients with dystonia than in control patients undergoing surgery for treatment of pain or tremor. Dystonia patients had a significantly higher proportion of deep sensory cells responding to movement of more than one joint (26%, 13/52) than did 'control' patients (8%, 4/49). Deep sensory cells in patients with dystonia were located in thalamic maps that demonstrated increased representations of parts of the body affected by dystonia. Thus dystonia patients showed increased receptive fields and an increased thalamic representation of dystonic body parts. The motor activity of an individual sensory cell was related to the sensory activity of that cell by identification of the muscle apparently involved in the cell's receptive field. Specifically, we defined the effector muscle as the muscle that, by contraction, produced the joint movement associated with a thalamic neuronal sensory discharge, when the examiner passively moved the joint. Spike X EMG correlation functions during dystonia indicated that thalamic cellular activity less often was related to EMG in effector muscles (52%) than in other muscles (86%). Thus there is a mismatch between the effector muscle for a thalamic cell and the muscles with EMG correlated with activity of that cell during dystonia. This mismatch may result from the reorganization of sensory maps and may contribute to the simultaneous activation of multiple muscles observed in dystonia. Microstimulation in presumed Vim in dystonia patients produced simultaneous contraction of multiple forearm muscles, similar to the simultaneous muscle contractions observed in dystonia. These observations are consistent with a model in which sensory input to Vim in dystonia is transmitted through altered sensory maps to activate multiple muscles in the periphery, producing the overflow of muscle activation that is characteristic of dystonia.