Dystonia is a neurological movement disorder characterized by repetitive, unintentional movements and disabling postures that result from sustained or intermittent muscle contractions. The basal ganglia and cerebellum have received substantial focus in studying DYT1 dystonia. It remains unclear how cell-specific ∆GAG mutation of torsinA within specific cells of the basal ganglia or cerebellum affects motor performance, somatosensory network connectivity, and microstructure. In order to achieve this goal, we generated two genetically modified mouse models: in model 1 we performed Dyt1 ∆GAG conditional knock-in (KI) in neurons that express dopamine-2 receptors (D2-KI), and in model 2 we performed Dyt1 ∆GAG conditional KI in Purkinje cells of the cerebellum (Pcp2-KI). In both of these models, we used functional magnetic resonance imaging (fMRI) to assess sensory-evoked brain activation and resting-state functional connectivity, and diffusion MRI to assess brain microstructure. We found that D2-KI mutant mice had motor deficits, abnormal sensory-evoked brain activation in the somatosensory cortex, as well as increased functional connectivity of the anterior medulla with cortex. In contrast, we found that Pcp2-KI mice had improved motor performance, reduced sensory-evoked brain activation in the striatum and midbrain, as well as reduced functional connectivity of the striatum with the anterior medulla. These findings suggest that (1) D2 cell-specific Dyt1 ∆GAG mediated torsinA dysfunction in the basal ganglia results in detrimental effects on the sensorimotor network and motor output, and (2) Purkinje cell-specific Dyt1 ∆GAG mediated torsinA dysfunction in the cerebellum results in compensatory changes in the sensorimotor network that protect against dystonia-like motor deficits.