Fine-Grained Topography and Modularity of the Macaque Frontal Pole Cortex Revealed by Anatomical Connectivity Profiles

The frontal pole cortex (FPC) plays key roles in various higher-order functions and is highly developed in non-human primates. An essential missing piece of information is the detailed anatomical connections for finer parcellation of the macaque FPC than provided by the previous tracer results. This is important for understanding the functional architecture of the cerebral cortex. Here, combining cross-validation and principal component analysis, we formed a tractography-based parcellation scheme that applied a machine learning algorithm to divide the macaque FPC (2 males and 6 females) into eight subareas using high-resolution diffusion magnetic resonance imaging with the 9.4T Bruker system, and then revealed their subregional connections. Furthermore, we applied improved hierarchical clustering to the obtained parcels to probe the modular structure of the subregions, and found that the dorsolateral FPC, which contains an extension to the medial FPC, was mainly connected to regions of the default-mode network. The ventral FPC was mainly involved in the social-interaction network and the dorsal FPC in the metacognitive network. These results enhance our understanding of the anatomy and circuitry of the macaque brain, and contribute to FPC-related clinical research. Electronic supplementary material The online version of this article (10.1007/s12264-020-00589-1) contains supplementary material, which is available to authorized users.


Fig. S2 Anatomical connectivity patterns between each subarea and cortical structures (right hemisphere).
The connectivity of each cluster yielded by tractography-based parcellation shown on the F99 surface using Caret helps to qualitatively identify differential connections. Anatomical connectivity fingerprints quantitatively identify the differences of the connectivity patterns between each subarea and cortical structures. For the fingerprints, we classified the connected regions on the periphery of the ellipse based on the different structure to which they belong, and displayed them using different color fonts (starting from area AI, and anticlockwise, the regions with different color fonts belong to the insular cortex, cingulate cortex, occipital cortex, temporal cortex, frontal cortex, and orbitofrontal cortex). Each subarea is named C1, C2, ..., C8.

Fig.S3
Anatomical connectivity patterns between each subarea and subcortical structures (right hemisphere). Population maps of the whole brain anatomical connectivity patterns shown in CIVM space using MRIcron help to qualitatively identify differential connections, and the connection pattern of each area is colored differently. Anatomical connectivity fingerprints quantitatively identify differences in the connectivity patterns between each subarea and subcortical structures. For the fingerprints, we classified the connected regions on the periphery of the ellipse based on the different structures to which they belong, and display them using different color fonts (starting from area LV, and anticlockwise, the brain regions with different color fonts belong to the lateral ventricles, midbrain, hypothalamus, central subpallium, pallium, paraseptal subpallium, striatum, subpallial amygdala, subpallial septum, lateral pallium, ventral pallium, and medial pallium). Each subarea is named C1, C2, ..., C8.
4 Similarity analysis and repeatability of connected brain regions, and FPC modularity structure

Sensitivity analysis of the parcellation results to the number of samples
To explore the sensitivity of the parcellation results to the number of samples, we parcellated the macaque FPC with three different values of streamlines/samples (15000, 13000, and 12000), and then calculated the overlap of the parcellation results. Here, we named the three parcellation results PR15, PR13, and PR12. In particular, PR15 was the current result and was regarded as the object of comparison. The other two parcellation results, PR13 and PR12, were each compared with PR15.
We calculated the number of all non-zero voxels in each post-processed maximum probability map (ppMPM), the number of overlapping voxels for the entire ppMPM, and the number of overlapping voxels of each subregion. Then we calculated the degrees of overlap were 95.12% (15000 vs 13000) and 91.64% (15000 vs 12000). Qualitatively, there was good consistency between PR15 and 6 PR13 (Fig. S5) and between PR15 and PR12 (Fig. S7). In addition, we calculated the degree of overlap for each subregion, for which we calculated the proportion of overlapping voxels of each subregion to itself. In particular, we calculated the number of voxels each subregion and the number of overlapping voxels of each subregion. Then we obtained eight results that represented the degree of overlap of each subregion (Figs S6 and S8, Tables S1 and S2). Besides, the distribution pattern of parcellation results was approximately consistent. All these results showed that a reasonable value of samples appears to have no clear effect on the accuracy of the parcellation results.   We also noted that the degree of overlap of cluster 5 between PR15 and PR12 was 56.62% (PR12 vs PR15) and recognized that there was a difference in the position of the dorsal FPC. The reason is that this subregion was a small subdivision, and the total numbers of voxels in cluster 5 were 675 (PR15) and 1185 (PR12). This number (671) represents the intersection of cluster5 between PR15 and PR12, which means that the maximum degree of overlap was 56.96% (675/1185) and the low value of 56.62% was acceptable. Also, the distribution patterns of parcellation results were approximately consistent and the total overlap was 91.64%, which suggested good consistency.