Like many neurodegenerative diseases, the clinical symptoms of Parkinsons disease (PD) do not manifest until significant progression of the disease has already taken place, motivating the need for sensitive biomarkers of the disease. While structural imaging is a potentially attractive method due to its widespread availability and non-invasive nature, global morphometric measures (e.g., volume) have proven insensitive to subtle disease change. Here we use individual surface displacements from deformations of an average surface model to capture disease related changes in shape of the subcortical structures in PD. Data were obtained from both the University of British Columbia (UBC) [n = 54 healthy controls (HC) and n = 55 Parkinsons disease (PD) patients] and the publicly available Parkinsons Progression Markers Initiative (PPMI) [n = 137 (HC) and n = 189 (PD)] database. A high dimensional non-rigid registration algorithm was used to register target segmentation labels (caudate, putamen, pallidum, and thalamus) to a set of segmentation labels defined on the average-template. The vertex-wise surface displacements were significantly different between PD and HC in thalamic and caudate structures. However, overall displacements did not correlate with disease severity, as assessed by the Unified Parkinson's Disease Rating Scale (UPDRS). The results from this study suggest disease-relevant shape abnormalities can be robustly detected in subcortical structures in PD. Future studies will be required to determine if shape changes in subcortical structures are seen in the prodromal phases of the disease.