The organization of dorsal frontal cortex in humans and macaques

Abstract

The human dorsal frontal cortex has been associated with the most sophisticated aspects of cognition, including those that are thought to be especially refined in humans. Here we used diffusion-weighted magnetic resonance imaging (DW-MRI) and functional MRI (fMRI) in humans and macaques to infer and compare the organization of dorsal frontal cortex in the two species. Using DW-MRI tractography-based parcellation, we identified 10 dorsal frontal regions lying between the human inferior frontal sulcus and cingulate cortex. Patterns of functional coupling between each area and the rest of the brain were then estimated with fMRI and compared with functional coupling patterns in macaques. Areas in human medial frontal cortex, including areas associated with high-level social cognitive processes such as theory of mind, showed a surprising degree of similarity in their functional coupling patterns with the frontal pole, medial prefrontal, and dorsal prefrontal convexity in the macaque. We failed to find evidence for "new" regions in human medial frontal cortex. On the lateral surface, comparison of functional coupling patterns suggested correspondences in anatomical organization distinct from those that are widely assumed. A human region sometimes referred to as lateral frontal pole more closely resembled area 46, rather than the frontal pole, of the macaque. Overall the pattern of results suggest important similarities in frontal cortex organization in humans and other primates, even in the case of regions thought to carry out uniquely human functions. The patterns of interspecies correspondences are not, however, always those that are widely assumed.

Figures

Figure 1.

a, The investigation consisted of three parts. First, we used DW-MRI tractography to parcellate human DFC. This approach has previously been shown to identify many of the key cytoarchitectonic regions when applied to other brain areas. Second, we examined resting-state BOLD coupling patterns between the dorsal frontal clusters we had found in the first step and regions in other parts of the human brain. Finally, we examined resting-state BOLD coupling patterns between dorsal frontal regions in the macaque. We examined BOLD coupling patterns because, in comparison with DW-MRI tractography, they are less affected by the distance between brain regions. They can therefore be used to illustrate in which cortical networks the areas participate in both humans and macaques. The networks associated with different frontal areas can then be compared and the resting-state functional coupling patterns in the macaque can be compared with the known anatomical connections in the same species. b, The dorsal frontal region investigated extended from the inferior frontal sulcus on the lateral surface to the cingulate sulcus on the medial surface. Rostrally, it included the frontal pole, and caudally it extended to the superior precentral sulcus or to a similar position on the medial surface. The region investigated therefore included all the tissue commonly called dorsolateral PFC, dorsomedial PFC, pre-SMA, SMA, FEF, and dorsal parts of the frontal pole and parts of PMd. The degree of overlap in the area investigated in nine subjects after registration to standard MNI space is indicated by the color (scale bar shown at center). c, Summary of the 10 right DFC regions revealed by the tractography-based parcellation in the nine human subjects studied. The results are presented in more detail in Figures 5, 6, 7, 8, 10, 11, 12, 14, and 15. We refer to cluster 1 as SMA, cluster 2 as preSMA, cluster 3 as area 9, cluster 4 as area 10, cluster 5 as area 9/46d, cluster 6 as area 9/46v, cluster 7 as area 46, cluster 8 as area 8A, cluster 9 as the rostral PMd, and cluster 10 as area 8B.

Figure 2.

a–f, ROIs in humans (a–c) and macaques (d–f). The resting-state functional coupling between each DFC regions (blue dots) and each of the target regions (purple dots) was estimated in both humans and macaques to determine how closely DFC networks corresponded in the two species. Note that the size of the dots does not refer to the actual size of the ROIs.

Figure 3.

Tractography-based parcellation revealed four clusters in human dorsomedial frontal cortex. a–d, In each case the clusters' BOLD coupling patterns with other brain regions suggested similarities with particular areas of macaque DFC: cluster 1 resembled SMA (a), cluster 2 resembled pre-SMA (b), cluster 3 resembled area 9 (c), and cluster 4 resembled area 10 (d). Colors indicate the degree of overlap in the cluster placement across subjects (scale bar shown at bottom).

Figure 4.

The initial first-pass parcellation of the three more posterior dorsomedial frontal areas, clusters 1, 2, and 3, could be further parcellated each into two areas. In each case, the parcellation resulted in a more dorsal area and a more ventral area. The ventral areas all corresponded to regions of the cingulate cortex that have been previously described (Beckmann et al., 2009) and were not investigated further in the current study. The dorsomedial clusters 1, 2, and 3 shown in Figures 5, 6, and 7, together with their associated coupling patterns, correspond to the dorsal tier of areas shown here.

Figure 5.

Cluster 1/SMA. a, b, The BOLD coupling patterns of cluster 1 in human subjects (a) and SMA in macaques (b) illustrated as z-statistic maps overlain on cortex. c, BOLD coupling patterns between human cluster 1 and macaque SMA and a set of regions in other parts of the brains in each species provided fingerprints of “functional connectivity.” d, The summed absolute differences between the functional coupling scores of human cluster 1 and 10 areas in macaque DFC suggested that the BOLD coupling pattern of human cluster 1 was least different to (in other words, it was most similar to) the coupling pattern associated with macaque SMA (lowest bar). Ars, Arcuate sulcus; CS, central sulcus; CingS, cingulate sulcus; IFS, inferior frontal sulcus; IPS, intraparietal sulcus; LF, lateral fissure; LuS, lunate sulcus; PCS, precentral sulcus; PS, principal sulcus; SFS, superior frontal sulcus.

Figure 6.

Cluster 2/pre-SMA. a–c, The BOLD coupling patterns of cluster 2 in human subjects (a) and pre-SMA in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 2 resembled macaque pre-SMA (lowest bar). All other conventions are as in Figure 5.

Figure 7.

Cluster 3/area 9. a–c, The BOLD coupling patterns of cluster 3 in human subjects (a) and area 9 in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 3 resembled macaque area 9 (lowest bar). All other conventions are as in Figure 5.

Figure 8.

Cluster 4/area 10. a–c, The BOLD coupling patterns of cluster 4 in human subjects (a) and area 10 in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 4 resembled macaque area 10 (lowest bar). All other conventions are as in Figure 5.

Figure 9.

Tractography-based parcellation revealed six clusters in human dorsolateral frontal cortex. a–f, In each case, the clusters' topological position and BOLD coupling patterns with other brain regions suggested similarities with particular areas of macaque DFC: cluster 5 resembled area 9/46d (a), cluster 6 resembled area 9/46v (b), cluster 7 resembled area 46 (c), cluster 8 resembled 8A (d), cluster 9 resembled rostral PMd (e), and cluster 10 resembled area 8B (f). All other conventions are as in Figure 5.

Figure 10.

Cluster 5/area 9/46d. a–c, The BOLD coupling patterns of cluster 5 in human subjects (a) and area 9/46d in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 5 resembled macaque area 9/46d (lowest bar). All other conventions are as in Figure 5.

Figure 11.

Cluster 6/area 9/46v. a–c, The BOLD coupling patterns of cluster 6 in human subjects (a) and area 9/46v in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 6 resembled macaque area 9/46v (lowest bar). All other conventions are as in Figure 5.

Figure 12.

Cluster 7/area 46. a–c, The BOLD coupling patterns of cluster 7 in human subjects (a) and area 46 in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 7 resembled macaque area 46 (lowest bar). All other conventions are as in Figure 5.

Figure 13.

Cluster 8/8A. a–c, The BOLD coupling patterns of cluster 8 in human subjects (a) and 8A in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 8 resembled macaque 8A (lowest bar). All other conventions are as in Figure 5.

Figure 14.

Cluster 9/rostral PMd. a–c, The BOLD coupling patterns of cluster 9 in human subjects (a) and rostral PMd in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 9 resembled macaque rostral PMd (lowest bar). All other conventions are as in Figure 5.

Figure 15.

Cluster 10/area 8B. a–c, The BOLD coupling patterns of cluster 10 in human subjects (a) and area 8B in macaques (b) and their associated functional connectivity fingerprints in humans and macaques (c). d, The summed absolute differences between the functional coupling scores suggested that human cluster 10 resembled macaque area 8B (lowest bar). All other conventions are as in Figure 5.

Figure 16.

a, Activations (light blue dots) associated with theory of mind and the making of inferences about others' beliefs are found in area 10/cluster 4 and area 9/cluster 3 (Amodio and Frith, 2006; Behrens et al., 2008; Van Overwalle and Baetens, 2009). b, Activations associated with the monitoring of information in memory (dark blue dots) are found in anterior lateral prefrontal cortex (Sakai et al., 2002; Owen et al., 2005; Amiez and Petrides, 2007; Champod and Petrides, 2007, 2010). Activity associated with cognitive branching and the representation and monitoring of counterfactual actions (red dots) is situated nearby but a little more ventrally and anteriorly (Daw et al., 2006; Gilbert et al., 2006; Boorman et al., 2009; Amiez et al., 2012). Activity associated with cognitive control demands is associated with more posterior DFC (green dots) (Kerns et al., 2004; Dosenbach et al., 2007; Duncan, 2010).

Figure 17.

Summary of the full matrix of similarity values between the functional networks associated with human DFC regions (horizontal axis) and macaque DFC regions (vertical axis). Blue colors indicate that there is little difference between the functional networks associated with a macaque and a human DFC area, while red colors indicate greater disparity between areas. Similarity/disparity estimates were based on the summed absolute differences between the functional coupling scores (also shown in panel c of Figs. 5, 6, 7, 8, 10, 11, 12, 13, 14, 15). The dashed line indicates the comparison for which the distance value is minimum.

Figure 18.

Summary of the DFC parcellation, presented in more detail in Figures 3 and 9. We refer to cluster 1 as SMA, cluster 2 as pre-SMA, cluster 3 as area 9, cluster 4 as area 10, cluster 5 as area 9/46d, cluster 6 as area 9/46v, cluster 7 as area 46, cluster 8 as area 8A, cluster 9 as the rostral PMd, and cluster 10 as area 8B. The probabilistic tractography was run from vertices at the gray matter/white matter boundary surface. Therefore, for the purpose of this figure, the clusters were dilated (performed using the FSL tool fslmaths) before being transformed (performed using the FSL tool FLIRT) to the MNI space.

Figure 19.

Replication of DFC parcellation in an additional group of 25 subjects in right and left hemisphere. We refer to cluster 1 as SMA, cluster 2 as pre-SMA, cluster 3 as area 9, cluster 4 as area 10, cluster 5 as area 9/46d, cluster 6 as area 9/46v, cluster 7 as area 46, cluster 8 as area 8A, cluster 9 as the rostral PMd, and cluster 10 as area 8B.