Surface morphometry substantially lower sample sizes to assess reduction in observed atrophy compared to volumes.
Sample sizes required to detect significant longitudinal changes in both surface deformation and volumes over 24 months. The Thalamus StatROI corresponds to clusters associated with negative sign changes in deformations. In contrast, the vertexwise analysis of the Midbrain produced clusters and corresponding StatROIs of both signs relative to the direction of the deformation along the surface normal. In both cases, the corresponding volumetric measures require a much greater sample sizes than deformations for detecting significant longitudinal changes.
This table shows the sample sizes required per arm to detect changes ranging from 20% to 80% of the observed changes from M24 to BL in PD for the StatROI surface deformation.
For comparison purposes, we also included the sample sizes required to detect the same changes in matching ROI volumes in parentheses. Notice that the Thalamus volumes in both the left and right hemispheres require a very large sample size to detect significant changes over 24 months from baseline. In contrast, the negative sign deformations require a relatively small sample size per arm to detect a significant change of about 60% of the observed longitudinal changes. For the Midbrain, both positive and negative sign deformations require even a lower sample size to detect a significant reduction of the observed changes over 24 months.