Figure 1.

Overview of the different methods used in our analysis. In addition to the gray matter (GM) and white matter (WM) volume maps provided by the PAC competition, we also pre-processed the data in order to obtain the regional volume, thickness and mean curvature information of the brain using Freesurfer. We then used different strategies that involved creating a gram matrix, dimensionality reduction algorithms (e.g., PCA) and TPOT (an automated machine learning framework) to train different models. In addition, to using different pre-processing, we also trained different models for the different sites where the data was recorded. All models that had a mean absolute error (MAE) lower than 7 years were used to build a weighted ensemble.