Feature Prediction

To evaluate whether our 14-dimensional Gaussianized local image filter feature set encompasses information captured using more typically utilized local features, we trained quadratic regression models to predict several typical features (Figure 2). Predictions were accurate with large effect sizes for FreeSurfer measures of curvature and sulcal depth (r= 0.80 and r= 0.73 respectively), gray-white contrast (r2 = 0.61), and cortical thickness (r2 = 0.46), as well as myelin contrast (r2 = 0.39). Spatial distributions appear similar between measured and predicted features.