Edited by: Kathleen S. Rockland, Boston University School of Medicine, United States
Reviewed by: Anita Disney, Vanderbilt University, United States; Floris G. Wouterlood, VU University Amsterdam, Netherlands
*Correspondence: Karl Zilles
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
We measured the densities (fmol/mg protein) of 15 different receptors of various transmitter systems in the supragranular, granular and infragranular strata of 44 areas of visual, somatosensory, auditory and multimodal association systems of the human cerebral cortex. Receptor densities were obtained after labeling of the receptors using quantitative
Densities of transmitter receptors vary between areas of human cerebral cortex. Multi-receptor fingerprints segregate cortical layers. The densities of all examined receptor types together reach highest values in the supragranular stratum of all areas. The lowest values are found in the infragranular stratum. Multi-receptor fingerprints of entire areas and their layers segregate functional systems Cortical types (primary sensory, motor, multimodal association) differ in their receptor fingerprints.
Cortical layers—as defined in classical architectonic studies (Brodmann,
Cortical layers also differ by their input and output, as well as by the preferred direction of connections with other cortical areas (Rockland and Pandya,
Also important for the present analysis of receptor fingerprints is the location of the terminal fields of connections which build most of the synapses, and thus must contain most of the receptors required for signal processing. Cortico-cortical neurons of the supragranular layers extend their apical dendrites up to layer I, where they form tufts, whereas not all of the infragranular neurons reach layer I with their apical dendrites, but have most of their dendritic arborizations in supragranular layers (Lund et al.,
Brains of three donors without any record of neurological or psychiatric diseases or of long-term drug treatment (age range: 72–77 years; 2 males, 1 female) were removed at autopsy. Subjects had given written consent before death and/or had been included in the body donor program of the Department of Anatomy, University of Düsseldorf, Germany, which also requires written consent by the donor. All procedures complied with the requirements defined by the local ethical committee. Post mortem delay before deep freezing was between 8 h and 18 h. Causes of death were cardiac arrest, lung edema, and myocardial infarction. After having separated each hemisphere into approximately 3 cm thick slabs, the slabs were shock frozen in isopentane at −40°C and stored at −80°C in airtight plastic bags until further processing. Thus, brain tissue was not treated with any chemical fixation substances. The total post mortem delay, including the deep freezing step, varied between 8 h and 18 h.
The unfixed frozen slabs were serially sectioned in the coronal plane (section thickness 20 μm) with a large scale cryostat microtome. Alternating sections were processed for quantitative
The labeled sections were exposed against tritium-sensitive films (Hyperfilm, Amersham, Braunschweig, Germany) together with plastic (Microscales®, Amersham) or brain tissue scales (with a known protein density) containing step-wise increasing radioactivity concentrations. Protein content in the homogenate used to create the brain tissue scales had previously been determined by means of the Lowry method (Lowry et al.,
where
Equidistant receptor profiles oriented vertically to the cortical surface were extracted by means of a minimum length algorithm from the linearized autoradiographs (Schleicher et al.,
We examined laminar distributions of 15 different receptors in 44 iso- and periallocortical areas (Figure Isocortical prefrontal areas 11, 8, 9, 10L, 10M, 46 and 47 (Brodmann, 4 and 6 (primary motor and premotor cortices; Brodmann, 3b, 1, 2, and 3a (primary somatosensory cortex; Brodmann, V1 (primary visual cortex, cytoarchitectonical area 17; Brodmann, Dorsal (V2d) and ventral (V2v) parts of the secondary visual cortex (cytoarchitectonical area 18; Brodmann, V3d, V3A, V3v, V4v, FG1 and FG2 (higher visual areas, cyto- and receptorarchitectonically defined areas hOc3d, hOc4d, hOc3v, hOc4v, FG1 and FG2, respectively; Rottschy et al., 44 and 45 (receptorarchitectonically defined ventral and anterior portions of Brodmann’s areas 44 (44v) and 45 (45a) in Broca’s region; Amunts et al., superior parietal areas 5L and 5M (cyto- and receptorarchitectonically identified areas of the higher unimodal somatosensory cortex; Scheperjans et al., inferior parietal areas PFt, PFm, PGa, and PGp (cyto- and receptorarchitectonically defined; Caspers et al., primary and higher unimodal auditory areas 41, 42, and 22 (cytoarchitectonically defined Te1, Te2 and 22; Morosan et al., lateral, medial and basal portions of the parieto-temporo-occipital regtion (37L, 37M and 37B, respective parts of area 37; Brodmann, multimodal temporal areas 20, 21, 36 and 38 (Brodmann, periarchicortical cingulate area 24 and isocortical cingulate areas 23, 31 and 32 (Brodmann,
Location of the analyzed cortical regions in the single subject template brain of the Montreal Neurological Institute. Location of areas: 10M, 10L from Bludau et al. (
Analysis of the region-specific balance between the densities of multiple receptors in a single cortical area prompted the introduction of the term receptor fingerprint (Zilles et al.,
To determine whether the density (over all layers) of each receptor type separately was homogeneously or not homogeneously distributed over the 44 areas, ANOVA tests were carried out and
For the question whether the 44 areas significantly differed in their receptor fingerprints, a discriminant analysis was carried with “area” as a grouping factor. This enables the determination of homogeneity or inhomogeneity of the fingerprints between areas. Since the discriminant analysis over all areas indicated a highly significant inhomogeneity of the fingerprints between the areas (
The color coded images of receptor densities give a first impression of their heterogeneous regional and laminar receptor distribution (Figure
Neighboring coronal sections through the human occipital lobe. Left: distribution of the cholinergic muscarinic M2 receptors. Color bar codes for receptor densities in fmol/mg protein. Right: Cell body stained section. Red contoured inset from primary visual cortex (V1) and blue contoured inset from V2v. The high magnifications in cell body stained sections were used to define the borders of the supragranular (sg), granular (g), and infragranular (ig) strata, which were then manually traced in the neighboring receptor autoradiograph. calc, calcarine sulcus; d, dorsal direction; ips, intraparietal sulcus; l, lateral direction; m, medial direction; v, ventral direction.
The regional densities (fmol/mg protein) of 15 different transmitter receptors for glutamate, GABA, acetylcholine, dopamine, noradrenaline and serotonin were measured in the three strata (supragranular, granular and infragranular strata) of 44 cytoarchitectonically defined iso- and periarchicortical areas of the human brain. Additionally, the mean density of each receptor over all cortical layers (mean areal density) was calculated. All original data are provided in Supplementary Table S2.
Figure
Absolute densities (fmol/mg protein) of each of the studied receptor types in 44 cortical areas. Dotted line: mean areal receptor density; blue line: receptor density in the supragranular stratum; red line: receptor density in the granular stratum; green line: receptor density in the infragranular stratum; straight black line indicates the mean areal density of each receptor averaged over all 44 areas. Areas with maximal or minimal mean areal receptor densities are indicated by their names in the respective graphs. In the case of the nicotinic α4β2 receptor, the absolute density of this receptor in the granular stratum of the three primary sensory areas (V1, 3b and 41) is highlighted. Asterisks indicate maxima and/or minima in the densities of supragranular (in blue), granular (in red), infragranular (in green) or all layers (in black) which differ significantly from the mean density of the given receptor averaged over all examined areas.
Receptor | All layers | Supragranular | Granular | Infragranular |
---|---|---|---|---|
AMPA | 7.190 | 7.468 | 1.2936 | 0.981 |
NMDA | ||||
Kainate | 11.386 | 13.877 | 13.651 | 12.354 |
GABAA | 0.787 | |||
GABAA/BZ | 3.917 | 6.892 | 6.093 | 7.391 |
GABAB | 0.176 | 0.104 | ||
M1 | 0.477 | 0.092 | 0.171 | |
M2 | 0.106 | |||
M3 | 5.617 | 0.472 | 0.274 | 0.088 |
α4β2 | 0.246 | 0.258 | 5.837 | |
α1 | 0.687 | |||
α2 | 0.532 | 1.273 | 8.385 | |
5-HT1A | 9.792 | |||
5-HT2 | 12.009 | 4.544 | 8.523 | 12.998 |
D1 | 1.231 | 2.040 | 0.188 | 5.412 |
If we focus on the
Large differences between 5-HT1A receptor densities of the different strata are visible throughout all regions studied (Figure
In most areas a canonical sequence of receptor densities from highest values in the supragranular stratum, intermediate values in the granular layer IV, and lowest values in the infragranular stratum is found. Exceptions from this rule are seen for the kainate receptor, which shows highest densities in the infragranular stratum, the M2 receptor, which reaches highest densities in the granular stratum of most areas, and the α1 and 5-HT1A receptors, which in some areas present higher densities in the infragranular than in the granular stratum.
Each cortical area expressed all receptor types, but at different mean areal and laminar densities (Supplementary Table S2). The regional-specific expression of all receptors studied constitutes the receptor fingerprint of an area or a stratum. We generated four receptor fingerprints per cortical area, visualizing the mean areal density, as well as the density in its supragranular, granular and infragranular strata. This is done for both the absolute (for selected areas see Figure
Absolute multi-receptor fingerprints of 15 different receptor types in each of the 44 cortical areas. From these areas, V1, V2v and V3v (visual system), 3b, 1 and 2 (somatosensory system), 41, 42 and 22 (auditory system), 4 and 6 (motor cortex), PFm (inferior parietal cortex), 10L (lateral part of the frontopolar cortex) and 44 (part of Broca’s region) were chosen as typical fingerprints representing different functional systems. The fingerprints of all other areas are found in Supplementary Figure S1. Scaling of the absolute fingerprints in fmol/mg protein is the same in all areas.
Normalized multi-receptor fingerprints of 15 different receptor types in each of the 44 cortical areas. For further information see Figure
Discriminant analyses of the fingerprints over all layers, and separately for the three strata shows that the fingerprints of all 44 areas are heterogeneous (
Since a pure visual comparison between the different fingerprints depends on the interpretation by the observer, we quantified the size of the mean areal and strata-specific fingerprints by computing the sum of the densities of all receptors over all layers or in each of the three strata.
The ranges of the sizes overlap slightly between the strata, if the standard deviations are taken as measure (Figure
Sum of the absolute densities of all receptors examined in each area and stratum as well as over all layers of each brain region. Mean values over all areas of the three strata and the total cortical depth and their standard deviations are indicated by straight lines and dotted lines, respectively.
Although the absolute densities of the different receptors vary between the examined brains (as revealed by the SD values and the variation coefficients specified in Supplementary Table S2), the proportional changes in densities between cortical areas remain constant when comparing different brains; e.g., in all examined brains, V1 contained higher overall NMDA, GABAA, or α2, but lower α1, or 5-HT1A receptor densities than did V2. Likewise, the relationship between receptor densities in the examined strata was also constant in the different brains examined; e.g., in area V1 of all brains highest 5-HT1A receptor densities were always found in the supragranular stratum, and lowest ones in the infragranular stratum.
In conclusion, the sizes of absolute fingerprints, and thus the density of all receptors together in each area or stratum, are regional-specific and show a canonical sequence (supragranular to granular to infragranular) of the strata from large to small fingerprints.
Using the normalized fingerprints, differences in the
The shapes of the normalized fingerprints (Figure
The fingerprints of areas of the primary auditory (area 41) and the secondary and multimodal auditory/temporal areas (areas 42, 20–22, 36) systematically differ. Particularly, the normalized density of the α1 receptor is higher in the association areas 20–22 and 36 compared to the unimodal auditory areas 41 and 42. The fingerprint of the temporo-polar area 38 differs in shape from all other temporal areas studied here. Likewise, the fingerprints of the temporo-occipital transition region (areas 37B, 37L and 37M) differ from those of the areas of the temporal and occipital lobes (Figure
The normalized fingerprints of motor areas 4 and 6 completely contrast with those of all sensory areas. Additionally, the high density of α1 receptors in the premotor cortex (area 6) contributes to the segregation of this area from the primary motor cortex (area 4).
Multimodal association regions are located in the prefrontal, temporal and parietal lobes including the precuneus region. Areas of the inferior parietal lobule are clearly segregated by different shapes of receptor fingerprints into two different groups: the supramarginal group with areas PFm and PFt, and the angular group with areas PGa and PGp. Both groups of fingerprints are separated by the high to very high density of the muscarinic M3 receptors in the angular group, and a lower density of this receptor in the supramarginal group which resembles only that of the average over all 44 areas. Notably, the fingerprints of PGa and the postcentral areas 5L and 5M are very similar, although the latter areas are not located in the inferior parietal lobule. Additional to their high M3 receptor density, the supragranular and granular strata of 5L, 5M, PGa and PGp show a GABAA/BZ density clearly above average. Thus, both receptors largely contribute to the similarity of the fingerprints between these four parietal areas and segregate them from PFm and PFt.
In a next step the area- and stratum-specific fingerprints were tested to answer two questions:
Do the fingerprints of the three strata build separate, strata-specific clusters if all cortical areas are compared? Do the fingerprints over all layers and/or the stratum-specific fingerprints systematically differ by their shapes and sizes between cortical areas? Do these variations indicate principal aspects of functional and topographical segregation, as well as hierarchical organization?
A multidimensional scaling analysis of the stratum-specific fingerprints shows three clusters (Figure
Multidimensional scaling analysis to visualize the different clusters of receptor fingerprints extracted from the supragranular, granular and infragranular strata of all examined areas. Note the exceptional positions of both motor areas 4 and 6, and of the Broca areas 44 and 45. These are the only areas whose supragranular and granular fingerprints take positions outside the cluster of all other areas.
The hierarchical cluster analysis of fingerprints of mean areal receptor densities separates all early visual areas from the rest of the cortex already after the first branching in the dendrogram (Figure
Hierarchical clustering of all regions except the agranular cortices of areas 4, 6 and 24 based on the receptor fingerprints extracted from their supragranular
Location of the clusters with areas of similar fingerprints in the single subject template brain of the Montreal Neurological Institute. Clusters were identified by hierarchical and k-means cluster analyses (see Figure
The fingerprints of the supragranular stratum (Figures
The fingerprints of the granular stratum (Figures
The fingerprints of the infragranular stratum (Figures
The cluster analyses of stratum-specific fingerprints (Figures
In summary, the fingerprints of the different strata differed greatly and enabled a separation of these strata-specific fingerprints into three clusters (Figure
The new aspect of this study is the large scale cross area comparison of stratum-specific receptor fingerprints. It has been demonstrated that the densities of various transmitter receptors vary considerably between different cytoarchitectonically defined areas in the human cerebral cortex (Cortés et al.,
As previously stated (Mash et al.,
One main focus of the present study was on the layer specificity of single receptor densities. We found a canonical sequence from high to low receptor densities when moving from the supragranular, to the granular and finally the infragranular stratum in nearly all regions. This general rule is only violated by the kainate, muscarinic M2, noradrenergic α1 and serotonin 5-HT1A receptors. Kainate receptors reach their highest densities in the infragranular stratum. However, also in this case a regional segregation into visual, multimodal temporal, parietal and prefrontal association areas can be confirmed by the variable kainate receptor densities in the different cortical areas. The M2 receptor shows the highest and the α1 receptor the lowest densities in the granular stratum of numerous areas. The most extreme situation was found for the 5-HT1A receptor, which reached by far highest densities in the supragranular stratum and lowest densities in the granular stratum. Very high absolute densities of this receptor were found in the supragranular stratum of temporal association areas 36 and 38 as well as of unimodal somatosensory area 2, auditory area 42 and anterior cingulate area 24.
In accordance with previous reports (e.g., Rakic et al.,
Laminar distribution of the NMDA receptor in human primary (V1) and secondary (V2) visual cortex as well as in areas 44 and 45 of the Broca region.
The present findings on the laminar distribution of GABAA receptors labeled with [3H] muscimol in human area V1 displayed a high density of this receptor in layer IVC that stood out due to the considerably lower densities in layers IVB and V. Notable is the patchy appearance of layer IVC with distributed GABAA receptor maxima along this layer. This is visible both in human and macaque V1 (Figure
Laminar distribution of the GABAA/BZ binding sites and the GABAA receptor in human and macaque primary visual cortex V1. Color bars code for receptor densities in fmol/mg protein. Scale bars 1 mm. Asterisks highlight the portion of layer IV with higher GABAA receptor densities in macaque than in human V1.
We found a bilaminar distribution pattern of the GABAA/BZ binding sites in human area V1 with the highest density in layer IVC, somewhat lower but still high densities in layers III–IVA, and the lowest densities in layers IVB and V–VI (Figure
The calcium-binding proteins parvalbumin, calbindin and calretinin label almost exclusively inhibitory GABAergic interneurons in the macaque visual cortex (Van Brederode et al.,
The inhibitory GABAA receptors are activated by the agonist muscimol. [3H]muscimol binds at the orthosteric α1/β2 subunit interface and interacts with the receptor via a high-affinity binding site. The α1 subunit has a strong influence on all binding properties, including desensitization, while the β2 subunit has minor impact (Baur and Sigel,
The α1 receptor reaches highest densities in layers I–II of human V1, moderate values in upper layer III. Lowest receptor densities are found in layers IVA–IVC and VI. Layer V shows a density intermediate between layers III and VI. The macaque V1 shows a very similar laminar distribution pattern (Rakic et al.,
Whereas most receptors show a parallel course of the mean areal densities and the layer-specific densities, the nicotinic α4β2 receptor is a notable exception. In all three primary sensory areas, this receptor displayed distinctly higher maxima in the granular stratum exceeding those of other strata and of mean areal densities (Figure
Since all cortical layers expressed all receptor types studied here, we analyzed the aspect of multi-receptor expression by calculating receptor fingerprints, a major tool for characterizing the balance between receptor expressions in the different strata. Here, the densities of 15 different receptor types in each of the 44 areas with a brain-wide distribution were visualized separately for each stratum. The discriminant analyses demonstrated the regional inhomogeneity of the fingerprints averaged over all layers and separately for each of the three strata. The results of the subsequent pairwise comparisons between areas, which resulted in significant differences only for a subsample of pairs, is not surprising because only three brains could be included in the present study of 15 different receptors in 44 cortical areas and their three strata. An enlargement of the sample of brains is currently not possible due to practical limitations (shortage of adequate human brain tissue with short post mortem delay, technical difficulties associated with the serial sectioning of entire deep frozen and unfixed human hemispheres and the financial requirements for autoradiographical processing of thousands of sections). The size of the fingerprints, scaled by absolute mean areal or stratum-specific densities, reflects the sum of the densities of all receptors, and shows considerable regional variations. Since the absolute densities of AMPA, NMDA, GABAA, GABAA/BZ and GABAB receptors are much higher than those of all other receptor types, we additionally calculated normalized fingerprints. These normalized fingerprints also show a regional heterogeneity of multi-receptor expression both for the mean areal densities and the stratum-specific densities.
The size of the absolute fingerprint of a cortical area is defined by the sum of the densities of all receptors measured in that area. Since we measured all receptor types in all areas, and the absolute fingerprints are identically scaled, the sizes of fingerprints enable a comparison of the densities (fmol mg/protein) between all 44 areas, and reveal area- as well as strata-specific differences, thus demonstrating the regional heterogeneity of receptor expression (Figure
The smallest fingerprints are found in the primary motor cortex area 4 (as previously described in Zilles et al.,
Interestingly, the motor and Broca areas are both highly myelinated in the adult brain (Hopf,
In general, unimodal sensory areas show a higher inter-laminar divergence of total receptor densities than multimodal association areas. This suggests a greater difference of receptor-mediated processing mechanisms between the cortical layers in the unimodal sensory areas, particularly in the primary visual cortex, compared to multimodal association areas.
The principal differences in the total densities of all receptors in a cortical area may be associated with the different prevailing connections of cortical layers. Whereas the supragranular strata are preferred by cortical-cortical connections, the granular stratum is that of thalamo-cortical and lateral connections, and the infragranular stratum gives raise mainly to subcortical projections (Rockland and Pandya,
In contrast to the absolute size of fingerprints, which is caused by the total density of all receptors in a given area, the shape of a fingerprint reflects the balance between the different receptors in an area. Since the absolute fingerprints are based on the same scaling of the densities of all receptors, the peculiarities of modulatory receptors are difficult to recognize by visual inspection, because their absolute densities are lower than those of the GABAergic receptors, sometimes by two orders of magnitude. Therefore, we additionally calculated normalized fingerprints.
E.g., the high M2 receptor density compared to those of all other receptors in the visual cortex, particularly in V1, is reflected by the shapes of their normalized fingerprints. i.e., V1 contains the highest M2 receptor density of all areas examined here. Thus, the impact of the M2 receptor on the specific shape of a fingerprint, and possibly on the balance between all receptors, is relatively the highest in V1. Also the contribution of α2 receptors on the shape of the fingerprint clearly stands out in all primary sensory areas when compared to the other receptors in these areas. Additionally, the above mentioned high absolute density of the nicotinic α4β2 receptor in the granular stratum of the primary somatosensory and auditory areas is well recognizable in the normalized fingerprint. Another example is also the high relative density of α1 receptors in all three strata of the premotor cortex. This suggests that the influence of this receptor is the highest in this area compared to the other areas examined here. A further exceptional position is observed for the M3 receptor in the inferior parietal area PGp (Supplementary Figure S2). In conclusion, the normalized fingerprints clearly demonstrate locally specific roles of certain receptor types which lead to different balances between the receptors in the here studied areas and possibly to such different balances in all other regions of the cerebral cortex. That is a hint to a regional specificity of the receptor-mediated mechanisms of information processing.
Since both the different sizes and shapes of fingerprints reflect the regional and stratum-specific balances between receptors, a multidimensional scaling analysis of the fingerprints was performed. It shows that each stratum and each area have a characteristic fingerprint and thus, specific levels of total receptor densities and balances between the receptors (Figure
Only exceptions from this general finding are the positions of the fingerprints of the granular stratum in motor areas 4 and 6, which are shifted into the range of the infragranular cluster, and the concomitant shift of the supragranular fingerprints into the granular cluster. This shift of stratum-specific fingerprints of areas 4 and 6 to the “wrong” clusters may be caused by the method with which we defined the position of their layer IV (see “Materials and Methods” Section). In the vast majority of the literature, areas 4 and 6 are described as being agranular, i.e., lacking a layer IV (Brodmann,
The separation of the layer-specific fingerprints further supports the existence of the above discussed divergence of receptor supported processing mechanisms between the three strata. The principal separation of the layers by their receptor fingerprints does not exclude a vertically organized interaction between the layers according to the concept of functional columns. Rather, it emphasizes that the stratum-specific receptor balances occurring at the different levels of a column are embedded in a vertically organized interaction which enables the area-specific functions.
The degree of similarity between the fingerprints of the three strata and of all layers together was analyzed by hierarchical cluster analyses which revealed a considerable regional heterogeneity, but also some notable general rules. Early visual areas were clearly separated from the rest of the cortex if we focus on the supragranular stratum and on all layers together, thus emphasizing similar receptor balances in this stratum over all these areas. The analysis of the fingerprints of the granular stratum emphasizes the exceptional organization of layer IV in V1 and the similaritiy of these fingerprints in the ventral visual stream in contrast to the dorsal visual stream, since areas V3d and V3A are clearly separated from the cluster of the ventral stream areas. The infragranular stratum does not separate dorsal and ventral stream areas from each other. Therefore, we can conclude, that organizational principles of the visual cortex (primary vs. higher unimodal sensory areas; ventral vs. dorsal stream areas) are recognizable by the fingerprints, but to different degrees in the different strata. A further general finding was the close similarity of the receptor fingeprints of the primary and secondary somatosensory and auditory areas. This is most clearly seen in the supra- and infragranular strata, as well as in the fingerprints of all layers together. This leads again to the conclusion that primary and secondary sensory areas are clearly different from higher unimodal or multimodal cortices by their receptor balances. As discussed above, specific receptor types (M2, nicotinic α4β2, noradrenergic α2, and serotonergic 5-HT2) play an important role here for the special position of particularly primary secondary areas. Although the function of the single receptor types has been intensely studied, the here important aspect of the role of the different receptor types within the cortical microcircuitry is, however, presently largely unknown. Our results on strata- and area-specific fingerprints and their relationships to general classification schemes of the cortex may be seen as a stimulus to study the specific functional role of these recepotr types in a systemic environment. Beside the distinct position of primary and early sensory unimodal areas, it must be emphasized that also the separation of the cluster of all prefrontal areas (analyses of all strata and all layers together) from other multimodal association areas with separate clusters for the infraparietal and temporal regions is a strong hint to the analytical potential of the hierarchical receptor fingerprint analysis as a tool to understand principal rules of cortical segregation also in higher functional and multimodal systems.
It is remarkable that the receptor fingerprints of the primary somatosensory area 3b and the primary auditory cortex (area 41) are very similar and cluster together, while both fingerprints largely differ from that of the primary visual cortex V1. Therefore, fingerprints seem to reflect differences in modalities of sensory systems. The similarity of the fingerprints of areas 41 and 3b may be explained by their similar functional properties, i.e., mechanoreception, whereas the input in V1 is clearly different. Thus, V1 has to comply with different functional requirements compared to the auditory and somatosensory systems. Comparable differences can also be found for the unimodal sensory areas and their fingerprints. The fingerprint of area 2 (somatosensory cortex) is very similar to that of area 42 (secondary auditory cortex), and both clearly differ from the fingerprint of the secondary visual cortex V2. Finally, the fingerprints of areas 44 and 45, which are subdivisions of Broca’s language region, consistently cluster together in all strata and in the entire cortical width. All these findings suggest a modality-specific component of the shape of receptor fingerprints.
The different levels of branching are indicated in the hierarchical cluster analyses. The highest possible number of clusters after k-means analysis was determined. All areas belonging to the same cluster in the supragranular, granular or infragranular strata, as well as to a cluster defined at the level of the mean over all layers are then labeled with the identical color. The resulting map indicates a subdivision of the cortex based on the similarity or dissimilarity of the regional-specific fingerprints in the different strata or entire areas. These maps are remarkable similar, since they assign in most cases the same areas to one cluster irrespective of the stratum in which the fingerprints have been yielded. Notably, the clusters comprise neighboring areas in many cases, but it must be emphasized, that the criterion of topographical neighborhood expresses at the same time a grouping of the areas according to different modalities or principal classifications into primary sensory, motor or higher multimodal association areas. Interestingly, area 38 does not cluster with the other multimodal temporal areas, but with areas on the supramarginal gyrus (PFm, PFt) and with the temporo-occipital transition zone (areas 37B, 37L, 37M). The latter areas have been attributed to the language network of Wernicke’s region (Mesulam et al.,
In conclusion, the present results provide evidence for the fact that the regional and laminar heterogeneity of multi-receptor expression patterns in the cerebral cortex is not random. Rather, transmitter receptor densities vary sytematically between cortical areas depending on the functional networks, or subdivisions thereof, to which they can be assigned. Furthermore, a general canonical sequence of densities from highest values in the supragranular stratum, intermediate values in the granular layer IV, and lowest values in the infragranular stratum is found in most areas, and for most receptor types. The stratum-specific differences in the patterns of multi-receptor balances, point at divergent receptor supported processing mechanisms between the three strata. Finally, area- and stratum-specific multi-receptor expression patterns (i.e., receptor fingerprints) reflect the segregation of the cerebral cortex into functionally and topographically definable groups of cortical areas (visual, auditory, somatosensory, limbic, motor), and reveal their hierarchical position (primary and unimodal (early) sensory to higher sensory and finally to multimodal association areas) within sensory systems.
This study was carried out in accordance with the recommendations of “Experimentelle wissenschaftliche Studien an Gewebeproben von Gehirnen und Organen von Körperspendern, Ethics Committee of the Medical Faculty of the Heinrich-Heine University Düsseldorf” with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the “Ethics Committee of the Medical Faculty of the Heinrich-Heine University Düsseldorf.”
KZ and NP-G conceived of and designed the study, analyzed data, drafted the manuscript and figures. NP-G acquired data. KZ obtained funding.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This project has received funding from the European Union’s Horizon 2020 Framework Programme for Research and Innovation under Grant Agreement No. 720270 (Human Brain Project SGA1).
The Supplementary Material for this article can be found online at: