GROUP-WISE CONSISTENT CORTICAL PARCELLATION BASED ON DTI-DERIVED CONNECTIONAL PROFILES

Seeking common and corresponding anatomical/functional regions across individuals via cortical parcellation has been studied for decades. Essentially, many cortical parcellation approaches are hampered by unclear cortical boundaries, individual variability, and difficulties in establishing correspondences during parcellation of multiple brains. In this paper, based on DICCCOLs, which is a map of discrete cortical landmarks demonstrated to possess group-wise consistent white matter fiber connection patterns across individuals, we presented a novel approach for group-wise consistent parcellation of the cortices under a hierarchical scheme. Experimental results show that this approach can achieve consistent fine-granularity cortical parcellation with intrinsically-established correspondences across individuals. Functional validation based on both resting state fMRI and task-based fMRI data are promising.