brainPAD

Predict brain age and calculate brain-predicted age difference scores from T1 MRIs

If you use this code, please cite: https://doi.org/10.1007/s11682-020-00260-3.

Relevant parameters (e.g. beta weights, training set intercept, training set slope, and training set voxelwise data - for z-scoring) are available for request from Zenodo through the following link: https://doi.org/10.5281/zenodo.2819646

Steps:

1) Run auto_reorient.m

2) Visually inspect orientation and check for artefacts

    a) Manually reorient poorly oriented images and discard images with artefacts
    OR
    b) Discard images with poor orientation and artefacts

3) Run complete_SPM_preprocessing_job.m

4) Visually QC/inspect preprocessed grey matter images for segmentation

5) Load all smwc*.nii files into a new single folder

6) Download model parameters from Zenodo https://doi.org/10.5281/zenodo.2819646 and load into MATLAB workspace

7) Run getBrainPADs.m
   
  [brainPAD, subid, prediction] = getBrainPADs(niftiFolder, saveto,...
  betas, test_outcome, test_subid, training_data, int, slope)
     
     niftiFolder = path to folder containing nifti images
     saveto = output folder where data should be saved
     betas = averaged_betas.mat
     test_outcome = chronological ages for each scan (1 * m array)
     test_subid = subids for each scan (1 * m cell)
     training_data = voxelwise_data.mat
     int = betaInt.mat
     slope = training_slope.mat

The following must be added to your MATLAB path for this code to work:

1) SPM
2) all code downloaded from this repository