Workflow of the iterative normalization process and the whole brain-based spatial statistics.

<p>(<b>A</b>) (i) recorded data were transformed into a 50 µm isogrid (nearest neighbor interpolation). (ii) after linear transformation according to manually set landmarks identified using a stereotaxic mouse atlas (iii) scanner- and sequence specific b0- and FA-templates were created. (iv) data were non-linearily normalized and templates were created (v) in an iterative process. (<b>B</b>) whole brain-based spatial statistics (WBSS) was performed after a quality check (vi) to eliminate motion corrupted volumes. After eddy-current correction, DTI-metrics were calculated (vii). Smoothing of the DTI-metrics maps (viii), statistical voxelwise comparison (ix), correction for multiple comparisons (x), and clustering (xi) lead to significant group differences.