Multimodal 3D atlas of the macaque monkey motor and premotor cortex

Abstract

In the present study we reevaluated the parcellation scheme of the macaque frontal agranular cortex by implementing quantitative cytoarchitectonic and multireceptor analyses, with the purpose to integrate and reconcile the discrepancies between previously published maps of this region. We applied an observer-independent and statistically testable approach to determine the position of cytoarchitectonic borders. Analysis of the regional and laminar distribution patterns of 13 different transmitter receptors confirmed the position of cytoarchitectonically identified borders. Receptor densities were extracted from each area and visualized as its "receptor fingerprint". Hierarchical and principal components analyses were conducted to detect clusters of areas according to the degree of (dis)similarity of their fingerprints. Finally, functional connectivity pattern of each identified area was analyzed with areas of prefrontal, cingulate, somatosensory and lateral parietal cortex and the results were depicted as "connectivity fingerprints" and seed-to-vertex connectivity maps. We identified 16 cyto- and receptor architectonically distinct areas, including novel subdivisions of the primary motor area 4 (i.e. 4a, 4p, 4m) and of premotor areas F4 (i.e. F4s, F4d, F4v), F5 (i.e. F5s, F5d, F5v) and F7 (i.e. F7d, F7i, F7s). Multivariate analyses of receptor fingerprints revealed three clusters, which first segregated the subdivisions of area 4 with F4d and F4s from the remaining premotor areas, then separated ventrolateral from dorsolateral and medial premotor areas. The functional connectivity analysis revealed that medial and dorsolateral premotor and motor areas show stronger functional connectivity with areas involved in visual processing, whereas 4p and ventrolateral premotor areas presented a stronger functional connectivity with areas involved in somatomotor responses. For the first time, we provide a 3D atlas integrating cyto- and multi-receptor architectonic features of the macaque motor and premotor cortex. This atlas constitutes a valuable resource for the analysis of functional experiments carried out with non-human primates, for modeling approaches with realistic synaptic dynamics, as well as to provide insights into how brain functions have developed by changes in the underlying microstructure and encoding strategies during evolution.

Keywords: Agranular frontal cortex; Cytoarchitecture; Functional connectivity; Mapping; Receptor architecture.

Graphical abstract

Fig. 1

Schematic drawings of the lateral and medial views of a macaque monkey hemisphere depicting the parcellation schemes of the agranular frontal region proposed by (A) Brodmann, 1905; (B) Barbas and Pandya, 1987; (C) Preuss and Goldman-Rakic,1991; (D) Morecraft et al., 2012; (E) Matelli et al.,1985, ; and (F) Caminiti et al., 2017. Note, that in the map of Caminiti et al. (2017) cortical areas were defined on the basis of both architectonic and connectional criteria. Red arrow marks a small portion of area F5p on the surface, whereas black arrows indicate area F5s buried within the inferior arcuate sulcus.

Fig. 2

2D flat map depicting all identified areas on the medial and dorsolateral premotor surfaces (a total of 13 premotor and 3 motor areas). Areas are marked on the left hemisphere and microanatomical features on the right hemisphere. Black full lines mark the sulci and dimple borders on the surface, whereas dashed black lines represent fundus. The only dashed black line on the surface marks the midline, which segregates medial and dorsolateral cortical surface. Section number (every 40^th) indicated between the hemispheres. arcs – spur of the arcuate sulcus, asd – anterior supracentral dimple, cgs – cingulate sulcus, cs – central sulcus, ias – inferior arcuate branch, ips – inferior parietal sulcus, lf – lateral fissure, ps – principal sulcus, sas – superior arcuate branch, spcd – superior precentral dimple.

Fig. 3

(A) Cytoarchitecture of the medial (4m), anterior (4a) and posterior (4p) subdivisions of the primary motor cortex, area 4. Colored square over the scale bar indicates the color used in Fig. 2 to code the area in question. (B) Quantitative analysis of cytoarchitectonic borders. The position of each border verified by the statistical analysis of Mahalanobis distances is highlighted by a red line (and corresponding profile index) on the GLI-image, and the corresponding dot plot (depicted to the right of the GLI-image) reveals that the location of significant maxima in the distance function (x axis; indicated by each dot) does not depend on the block size (y axis), but remains constant over large block size intervals (highlighted by the red frame). Roman numerals indicate cortical layers. Scale bars 1 mm. cgs – cingulate sulcus, cs - central sulcus.

Fig. 4

(A) Cytoarchitecture of caudal medial premotor area F3, as well as of the subdivisions of caudal dorsolateral premotor area F2 (F2d and F2v). Colored square over the scale bar indicates the color used in Fig. 2 to code the area in question. (B) Quantitative analysis of cytoarchitectonic borders. For details see Fig. 3. Roman numerals indicate cortical layers. Scale bars 1 mm. cgs – cingulate sulcus, arcs – spur of the arcuate sulcus, lf – lateral fissure, spcd – superior precentral dimple.

Fig. 5

(A) Cytoarchitecture of rostral medial premotor area F6, as well as of the subdivisions of rostral dorsolateral premotor area F7 (F7d, F7i and F7s). Colored square over the scale bar indicates the color used in Fig. 2 to code the area in question. (B) Quantitative analysis of cytoarchitectonic borders. For details see Fig. 3. Roman numerals indicate cortical layers. Scale bars 1 mm. cgs – cingulate sulcus, ias – inferior arcuate branch, lf – lateral fissure, sas – superior arcuate branch.

Fig. 6

(A) Cytoarchitecture of the subdivisions of caudal ventrolateral premotor area F4 (F4s, F4d and F4v). Colored square over the scale bar indicates the color used in Fig. 2 to code the area in question. (B) Quantitative analysis of cytoarchitectonic borders. For details see Fig. 3. Roman numerals indicate cortical layers. Scale bars 1 mm. arcs – spur of the arcuate sulcus, cgs – cingulate sulcus, lf – lateral fissure, spcd – superior precentral dimple.

Fig. 7

(A) Cytoarchitecture of the subdivisions of rostral ventrolateral premotor area F5 (F5s, F5d and F5v). Colored square over the scale bar indicates the color used in Fig. 2 to code the area in question. (B) Quantitative analysis of cytoarchitectonic borders. For details see Fig. 3. Roman numerals indicate cortical layers. Scale bars 1 mm cgs – cingulate sulcus, ias – inferior arcuate branch, lf – lateral fissure, sas – superior arcuate branch.

Fig. 8

(A) Schematic drawing of a coronal section processed for receptor labeling at the level of the primary motor cortex showing the position of its medial (4m;), dorsolateral (4a;) and sulcal (4p;) subdivisions. Areal color coding as in Fig. 2. (B) Exemplary sections depicting the distribution of NMDA, GABA_B, M₁ and M₃ receptors. Lines represent borders between defined premotor areas. Scale bars code for receptor densities in fmol/mg protein. Distribution patterns of all 13 receptors are shown in Supplementary Fig. 1.

Fig. 9

(A) Schematic drawing of a coronal section processed for receptor labeling at the level of the posterior premotor region showing the position of medial (F3), dorsolateral (F2d and F2v) and ventrolateral (F4s, F4d and F4v) premotor areas. Areal color coding as in Fig. 2. (B) Exemplary sections depicting the distribution of kainate, GABA_B and α₁ receptors. Lines represent borders between defined premotor areas. Scale bars code for receptor densities in fmol/mg protein. Distribution patterns of all 13 receptors are shown in Supplementary Figs. 2 and 4.

Fig. 10

(A) Schematic drawing of a coronal section processed for receptor labelling at the level of the anterior premotor region showing the position of medial (F6), dorsolateral (F7d, F7i and F7s) and ventrolateral (F5s, F5d and F5v) premotor areas. Areal color coding as in Fig. 2. (B) Exemplary sections depicting the distribution of kainate, GABA_B and M₂ receptors. Lines represent borders between defined premotor areas. Scale bars code for receptor densities in fmol/mg protein. Distribution patterns of all 13 receptors are shown in Supplementary Figs. 3 and 5.

Fig. 11

Position and extent of the motor and premotor areas on lateral and medial views of the Yerkes19 surface (Donahue et al., 2016). The mean receptor densities of three exemplary receptor types (α₁, GABA_A and M₁) have been projected onto the corresponding area. Color bars code for receptor densities in fmol/mg protein. The projections of all receptor types onto the Yerkes 19 surface are shown in Supplementary Fig. 7. The files with the parcellation scheme and coding for the densities of all 13 receptors in each area are available via the EBRAINS platform of the Human Brain Project (https://kg.ebrains.eu/search/instances/Project/e39a0407-a98a-480e-9c63-4a2225ddfbe4) and the BALSA neuroimaging site (https://balsa.wustl.edu/study/g7qwN).

Fig. 12

Hierarchical cluster (A) and multidimensional scaling (MDS) (B) analyses of the receptor fingerprints of macaque primary motor and premotor areas. K-means clustering and elbow analysis showed three as the optimal number of clusters.

Fig. 13

Connectivity fingerprints of each examined area and showing their connectivity strength to areas of the resting state network. Green codes for prefrontal areas, blue for cingulate, yellow for somatosensory and red for parietal areas. Nomenclature of targeted areas is based on the Kennedy atlas (Markov et al., 2014), axis scaling is identical in all polar plots and indicates the z-score (following Fisher's r-to-z transformation) of the correlation coefficients.

Fig. 14

Hierarchical cluster (A) and multidimensional scaling (MDS) (B) analyses of the functional connectivity fingerprints of macaque primary motor and premotor areas. K-means clustering and elbow analysis showed four as the optimal number of clusters.