The ALE procedure follows three steps, which we will only briefly describe here (see refs. [37, 38, 43] for detailed information). First, GingerALE tests for the spatial uncertainty of the reported coordinates by modeling them with a Gaussian function, thereby accounting for the sample size of each study. This is followed by the construction of a whole-brain map for each study, whereby each voxel gets a value assigned that is equal to the probability of WM volume or integrity alterations within it. Subsequently, these maps are merged across all studies resulting in an ALE image with ALE values representing the likelihood of these alterations and tested for statistical significance with correction for multiple comparisons.
