MRI-based surface-assisted parcellation of human cerebellar cortex: an anatomically specified method with estimate of reliability

Makris N, Schlerf JE, Hodge SM, Haselgrove C, Albaugh MD, Seidman LJ, Rauch SL, Harris G, Biederman J, Caviness VS, Kennedy DN, Schmahmann JD

Neuroimage 2005 May;25(4):1146-60

PMID: 15850732


We revisit here a surface assisted parcellation (SAP) system of the human cerebellar cortex originally described in Makris, N., Hodge, S.M., Haselgrove, C., Kennedy, D.N., Dale, A., Fischl, B., Rosen, B.R., Harris, G., Caviness, V.S., Jr., Schmahmann, J.D., 2003. Human cerebellum: surface-assisted cortical parcellation and volumetry with magnetic resonance imaging. J Cogn Neurosci 15, 584-599. This system preserves the topographic and morphologic uniqueness of the individual cerebellum and allows for volumetric analysis and representation of multimodal structural and functional data on the cerebellar cortex. This methodology integrates features of automated routines of the program FreeSurfer as well as semi-automated and manual procedures of the program Cardviews to create 64 cerebellar parcellation units based on fissure information and anatomical landmarks of the cerebellar surface. Using this technique, we undertook the parcellation of ten cerebella by two independent raters. The reliability of the resulting parcellation units (64 total) was high, with an average Intraclass Correlation Coefficient (ICC) of 0.724 in the vermis and 0.853 in the hemispheres. Clusters of parcellation units were then created, based on lobar and connectivity data and functional hypotheses. These 36 clusters, when treated as anatomical units, had an average ICC of 0.933. Whereas the individual units provide a high level of detail and anatomical specificity, the clusters add flexibility to the analysis by providing higher reliability.