Trends and properties of human cerebral cortex: correlations with cortical myelin content

Abstract

"In vivo Brodmann mapping" or non-invasive cortical parcellation using MRI, especially by measuring cortical myelination, has recently become a popular research topic, though myeloarchitectonic cortical parcellation in humans previously languished in favor of cytoarchitecture. We review recent in vivo myelin mapping studies and discuss some of the different methods for estimating myelin content. We discuss some ways in which myelin maps may improve surface registration and be useful for cross-modal and cross-species comparisons, including some preliminary cross-species results. Next, we consider neurobiological aspects of why some parts of cortex are more myelinated than others. Myelin content is inversely correlated with intracortical circuit complexity - in general, more myelin content means simpler and perhaps less dynamic intracortical circuits. Using existing PET data and functional network parcellations, we examine metabolic differences in the differently myelinated cortical functional networks. Lightly myelinated cognitive association networks tend to have higher aerobic glycolysis than heavily myelinated early sensory-motor ones, perhaps reflecting greater ongoing dynamic anabolic cortical processes. This finding is consistent with the hypothesis that intracortical myelination may stabilize intracortical circuits and inhibit synaptic plasticity. Finally, we discuss the future of the in vivo myeloarchitectural field and cortical parcellation--"in vivo Brodmann mapping"--in general.

Keywords: Aerobic glycolysis; Cerebral cortex; Cortical area; Cortical parcellation; Myelin map; PET.

Figure 1

Single subject unsmoothed myelin map from the 0.7mm isotropic publically released HCP data. Medial wall is masked in black. Color scale ranges from 4% (black) to 96% (red).

Figure 2

Pre-scan normalized (i.e. Siemens scanner corrected for B1− receive field) T2w (A) and T1w (C) images. sqrt(T1w X T2w) bias field corrected T2w (B) and T1w (D) images. The corrected images are more homogeneous because the B1+ transmit field is correlated between the T1w and T2w images.

Figure 3

Group average myelin maps from 69 humans, 29 chimpanzees, and 19 macaques scaled according to brain size. Many areal homologies are suggested by similarities in the maps across species, but differences are also apparent. More of the expansion in brain size has come from lightly myelinated (blue) areas vs from heavily myelinated (red/orange/yellow) or very lightly myelinated areas (black/purple). Adapted from Glasser, M.F., Preuss, T.M., Nair, G., Rilling, J.K., Zhang, X., Li, L., Van Essen, D.C., 2012b.

Figure 4

Histograms of myelin map values across humans, chimpanzees, and macaques. The peak of the distribution is in the lightly myelinated areas in humans but in the heavily myelinated areas in macaques, with chimpanzees intermediate but closer to macaques than to humans. Adapted from Glasser, M.F., Preuss, T.M., Nair, G., Rilling, J.K., Zhang, X., Li, L., Van Essen, D.C., 2012b.

Figure 5

Myelin maps from Glasser and Van Essen (2011) (A–D) 69-subject MRI data. Partial volume corrected and surface regressed maps of aerobic glycolysis (GI) (E–H) from Vaisnavi et al (2010) 33-subject PET data.

Figure 6

Mean myelin map (A–D) and aerobic glycolysis (GI) values (E–H) across the Yeo et al 2011 functional network parcellation.