Fig. 1
From: Advanced structural brain aging in preclinical autosomal dominant Alzheimer disease
Overview of study design. (A) Whole-brain T1 structural MR images from DIAN were processed with brain extraction and linear registration to an atlas template, and were then used as input data for brain age analyses. (B) DeepBrainNet (DBN) is a publicly available convolutional neural network that has previously been trained to predict age in 11,729 MRI scans from independent control samples [20]. (C) Model prediction accuracy of age was evaluated by calculating Pearson’s correlation coefficient (r), coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) in non-carrier participants. (D) Brain Age Gap (BAG) for each DIAN participant was calculated as the difference between DBN-predicted and true chronological age. Figure created in part with BioRender.com and modified from [33]