The unlabeled black arrows on the right of Fig. 4 indicate curvature-like features in the parcellation. These lines follow the fundus and crown of the cingulate sulcus and gyrus respectively. The emergence of these macroscopic landmarks could be associated with sampling errors at areas of high curvature, where partial voluming is more prevalent. Alternatively, it may reflect a relationship between gyrification and diffusion anisotropy, as suggested by several groups—e.g., . Deeper cortical layers appear to thin in sulci, which has been suggested to be a way of maintaining equal local volume by folding-induced tangential stretching [31, 32]. By contrast, upper cortical layers appear to puff out and become more myelinated on gyri. These systematic folding-correlated effects may give rise to detectable differences in grey matter diffusion patterns. The initial detection of a correlation between gyrification and T1 was similarly initially dismissed as a depth-sampling artifact, but then subsequently suggested to be partly due to real myelination differences between sulcal and gyral cortex. An additional complication is that partial volume errors may be detecting systematic differences in fiber direction near the grey/white matter border; for example, the dominant diffusion direction is expected to be highly radial in areas of high curvature, such as gyral crowns, and more tangential in along the banks of gyri due to the angle at which u-fibres project into the cortex [20, 21].