We tested the hypothesis that family risk for psychosis (FR) and clinical risk for psychosis (CR) are associated with structural brain abnormalities, with increased deficits in those at both family risk and clinical risk for psychosis (FRCR). The study setting was the Oulu Brain and Mind Study, with subjects drawn from the Northern Finland 1986 Birth Cohort (n = 9479) using register and questionnaire based screening, and interviews using the Structured Interview for Prodromal Symptoms. After this procedure, 172 subjects were included in the study, classified as controls (n = 73) and three risk groups: FR excluding CR (FR, n = 60), CR without FR (CR, n = 26), and individuals at both FR and CR (FRCR, n = 13). T1-weighted brain scans were acquired and processed in a voxel-based analysis using permutation-based statistics. In the comparison between FRCR versus controls, we found lower greymatter volume (GMV) in a cluster (1689 voxels at−4.00,−72.00,−18.00 mm) covering both cerebellar hemispheres and the vermis. This cluster was subsequently used as a mask to extract mean GMV in all four groups: FR had a volume intermediate between controls and FRCR. Within FRCR there was an association between cerebellar cluster brain volume and motor function. These findings are consistent with an evolving pattern of cerebellar deficits in psychosis risk with the most pronounced deficits in those at highest risk of psychosis.