Figure 1.

Comparison of different analysis steps necessary for the measurement of the cortex with and without lesion filling. The first row (a) illustrates the analysis strategy without lesion filling, while the second (b) illustrates the approach with lesion filling. In both rows, a T1w MRI (a1 and b1) with segmented lesions, the corresponding tissue classification derived from CIVET (a2 and b2), the WM surface transformed back to volume space (a3 and b3), the representation of the cortex (a4 and b4), and a magnified view of the cortical thickness assessment (a5 and b5) are shown. The figure shows that the misclassification of WM lesions, which occurs using the approach without lesion filling (a2) produces an inaccurate WM surface (a3) and, consequently, an incorrect estimation of cortical thickness (a4) especially in the proximity of juxtacortical lesions (a5). Using the approach without lesion filling, the estimated cortex in fact includes also lesional voxels.