6.3 In search of a detailed cartography of the cortex

The search for a kind of more detailed cartography, a modern ‘parcellation’, of the cortex as a way to segregate areas of the cortex as a mosaic of discrete areas or networks, according to their distinguishing features (eg, microstructure, connectivity, or functional properties) is a current vigorous area of neuroscience research¹.

Starting in the early 20th century, the human cerebrum was mostly divided by histological differences between regions. One of the first systematic studies based on structure was performed by Brodmann. Brodmann’s map ²includes 43-52 regions according to regional cytoarchitectural differences in cells and their laminar organisation, including cell size, spacing or packing density, and arrangement of myelinated nerve fibres (myeloarchitecture). The Vogts (husband and wife team, Oskar and Cécile) increased the number of areas to 182³. The similarities of the architecture in layers across neighbouring sections of the cortex supported the idea that it is organised in arrays of modules with neighbours of similar structure and functionality. However, the primary somatosensory cortices differ from the high-order association cortices in terms of their laminar organization and afferent and efferent connections⁴.

A recent line of research has taken a simplified view of cortical microstructure by reducing it to a single quantity: the number of neurons within a unit of surface area⁵. These studies have demonstrated a rostro-caudal gradient in neuron number in the cortices of a broad range of mammalian species, including several rodents, marsupials, and non-human primates. Neuron numbers are generally high in caudal portions of the cortex, such as the occipital lobe, and gradually decrease toward more rostral regions⁶.

In parallel, a myeloarchitectonic approach has emerged with high definition MRI with excellent white/grey matter contrast in the brain due to the presence of myelin, and thus uniquely suited for in vivo studies of cortical myeloarchitecture⁷.

Initial identification of cortical columns 

Following the initial proposal that the visual cortex is organised anatomically in vertical columns across its thickness by Rafael Lorente de Nó⁸, Vernon Mountcastle revealed a vertical organisation of the cortical layers in columns⁹. Each module is composed of minicolumns, about 30 µm in diameter and containing about 100 cells. Larger columns (1 mm diameter) may be composed of hundreds of minicolumns¹⁰.

The Nobelists David Hubel and Torsten Wiesel proposed that the units of organisation of the visual cortex are cortical columns, as identified by electrophysiological recordings and axonal tracing individual neurons in the primary visual cortex in cats. They showed that minimal visual stimuli such as short lines selectively activated neurons within small columns of cortex. In the primary visual cortex, there are at least ~5300 columns with 400,000 neurons/mm². Each column comprises a few hundred neurons comprising different classes with very specific synaptic connections¹¹. Pyramidal cells have dense recurrent synaptic connections, each cell receiving from 6000 to 13,000 synaptic inputs. Approximately 80% of these connections come from within the same cortical area, 95% of which arise from neurons within 2 mm. Horizontal fibres in the most superficial (external) ‘supra-granular’ layers span only a few millimetres across. Most of the thousands of synapses formed on the dendrites of cortical neurons come from their neighbouring excitatory neurons located in the same cortical area. Thus these axons connect neurons spanning multiple adjacent functional columns within the cortex. 

Cortical columns are connected with neighbouring ones via short intracortical pathways involving as many as 10¹¹ axon collaterals of pyramidal cortical cells which run tangentially within the grey matter for a few mm. Connections a few cm long run between folds within the same lobe.

Using retrograde and anterograde tracing comparing their distribution pattern plus serial two-photon tomography imaging in fixed brains, Watanabe et al have revealed interconnections between layers and different areas of the prefrontal cortex in macaques¹². The retrograde signals exhibited striking colocalization with the anterograde signals and confirmed that reciprocal connectivity applies to both highly selective and diffuse projections. Although the existence of short internal loops within the cortex is undeniable, the general architecture of the cortex appears a continuous sheet of interconnected columns, making it difficult to identify sharp boundaries between levels.

Several cooperating projects over the past few years have addressed the issues of a suitable cartography of the cortex. In 2017, The National Institutes of Health (NIH) announced US$500 million in funding for the Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative – Cell Census Network (BICCN), a funding mechanism with the aim to develop a comprehensive description of brain cell types and to build a comprehensive atlas of the human brain compared with other species. 

These studies not surprisingly revealed variations in the relative thickness of the different layers and the cellular composition of the cortex in different areas. Applying single-cell RNA sequencing to samples from nearly 100 anatomical locations in the adult human brain demonstrated that, although some brain regions have distinct cell types, many regions differ primarily in the relative proportions of a shared group of cell types¹³. 

Overall, the data support an evolutionary strategy for the diversification of regional function that largely does not depend on the generation of new cell types but rather uses small variation within cell types and changes in their relative distribution to create distinct circuitry. “Not unexpectedly, besides the variation among brain regions, there is also variation between the brains of different individuals. There is no single prototypical human”¹⁴.

Each area of the cortex is also interconnected with more distant areas via intercortical connections running in the tracts of white matter. In addition, every area of the cortex in one hemisphere is interconnected to the corresponding area of the other hemisphere via the corpus callosum which in humans contain 100-200 million fibres, a small proportion of all connections.

The Brainnetome Project was launched in 2014 in China to investigate the hierarchy of the human brain from genetics to neuronal circuits to behaviours, conceptualising two components (nodes and connections) as the basic research units. The NIH Human Connectome Project, launched in 2009, used brain scans to map out brain circuitry and the large fibre tracts connecting different brain regions.

There are currently at least 15 groups building Human Brain Atlases based on a variety of classification principles which include levels of cortical organisation, cellular composition, pattern of connectivity, myelin content, enzyme activity, magnitude of gene expression, pharmacological receptor architecture, neurochemical markers or functional properties.¹⁵ 

Despite this collective effort, what emerges from these modern ‘parcellation’ studies is that identifying the number of loops within the cortex appears beyond reach at present. 

Hierarchical organisation of the neural circuits involved in language.

Another example of hierarchical organisation of the cortex is the neural architecture underlying the complex function of language. While all sensory inputs related to the ability to read and listen to language including visual, auditory, and even tactile inputs arrive to the corresponding primary visual, auditory and somatosensory cortices, these primary inputs project to higher cortical levels where they are then integrated. Most important of these is Wernicke’s area near the angular gyrus, which is where the primary comprehension of language occurs. This association area is clearly several levels above the primary sensory cortices processing incoming information. From Wernicke’s area, intercortical connections project to Broca’s area in the frontal lobe, which is a premotor area for speech but which is also important for language comprehension. These intra-hemispheric connections, forming the arcuate fasciculus, involve no more than 25,000 neurons. This clearly indicates that higher cognitive functions linked to language require higher levels in the hierarchy of the cortical ladder.

Hierarchy in the motor-frontal lobes

A clear indication for a hierarchical organisation of internal loops in the frontal cortex is shown by the different levels connected with the primary motor cortex: at its base, it is progressively connected to levels anatomically more anterior and functionally more removed from the external world. Thus the pathways go from premotor areas for execution of motor actions and planning of movements, to prefrontal areas responsible for cognitive and executive functions, goal-directed behaviours, and motivational and emotional processing¹⁶. 

In an extensive historical review on the cortex, Petrides (2005)¹⁷ concludes “The lateral frontal cortex appears to be functionally organized along both a rostral-caudal axis and a dorsal-ventral axis. The most caudal frontal region, the motor region on the precentral gyrus, is involved in fine motor control and direct sensorimotor mappings, whereas the caudal lateral prefrontal region is involved in higher order control processes that regulate the selection among multiple competing responses and stimuli based on conditional operations. Further rostrally, the mid-lateral prefrontal region plays an even more abstract role in cognitive control.”

The association cortex comprises the majority of the human cerebral cortex and is made up of multiple, interdigitated association networks. The properties of association networks are quite different from that of sensory and motor cortices. Sensory and motor areas are embedded within cerebral networks that are organised in a topographic fashion forming preferentially local networks, meaning that adjacent areas tend to show strong functional coupling with one another. By contrast, multiple association networks involve areas distributed throughout the cortex, always including discrete regions within prefrontal, parietal, temporal, and midline cortices. These distributed association networks are interdigitated in a manner that yields complex functional zones, particularly in parietal and prefrontal association cortices¹⁸. 

Despite convincing evidence for a hierarchical organisation of the cerebral cortex, the boundaries of each level are harder to identify, not least because the number of synaptic distances representing the external world in the cortex may be more flexible (ie plastic) rather than completely hard wired.

In summary, all internal neural loops above the lower spinal sensory-motor neuromechanical loop are nested in a hierarchical manner, forming a complex system of superimposed parallel interconnected neural loops with well recognised ascending/feedforward and descending/feedback neural pathways.

Grillner, one of the foremost neuroscientists of modern era, supports this perspective when he states that that to understand the brain “a multi-level approach is required in which the different levels link into each other”¹⁹.

The multiple superimposed loops of the brain usually operate in parallel. At every level, competition and integration between multiple ascending and descending inputs within each internal loop shapes the final behaviour. Consequentially, inner experiences are most likely to involve multiple loop levels. It follows that any action, any behaviour, cannot be easily attributed to a single level of command or location. 

---

---

1. MM Shira et al (2023) HumanBrainAtlas: an in vivo MRI dataset for detailed segmentations. Brain Structure and Function 228, 1849-1863
   SM Nelson et al (2010) A parcellation scheme for human left lateral parietal cortex. Neuron 67, 156-170. ↩︎
2. K Brodmann (1908) Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. ↩︎
3. See: R Nieuwenhuys & CAJ Broere (2023) A new 3D myeloarchitectonic map of the human neocortex based on data from the Vogt–Vogt school. Brain Structure and Function 228, 1549-1559;
   K Amunts & K Zilles (2015) Architectonic mapping of the human brain beyond Brodmann. Neuron 6, 1086-1107. ↩︎
4. N Kanwisher (2010) Functional specificity in the human brain: a window into the functional architecture of the mind. Proceedings of the National Academy of Science USA 107,11163–11170;
   DN Pandya & B Seltzer (1982) Association areas of the cerebral cortex. Trends in Neuroscience 5, 386–390. ↩︎
5. CE Collins et al (2010) Neuron densities vary across and within cortical areas in primates. Proceedings of the National Academy of Science USA 107, 15927–15932;
   CJ Charvet et al (2015) Systematic, cross-cortex variation in neuron numbers in rodents and primates. Cerebral Cortex 25, 147–160. ↩︎
6. JM Huntenburg, P-L Bazin, DS Margulies (2018) Large-Scale Gradients in Human Cortical Organization. Trends in Cognitive Sciences 22, P21-31. ↩︎
7. EL Barbier et al (2002) Imaging cortical anatomy by high-resolution MR at 3.0T: Detection of the stripe of Gennari in visual area 17. Magnetic Resonance in Medicine 48, 735–738. ↩︎
8. R Lorente de Nó (1938) Analysis of the activity of the chains of internuncial neurons. Journal of Neurophysiology 1, 207-244. ↩︎
9. VB Mountcastle (1957) The columnar organization of the brain. Brain 120, 701-722. ↩︎
10. AM Bastos et al (2012) Canonical microcircuits for predictive coding. Neuron 76: 695-711;
    K Amunts & K Zilles (2015) Architectonic mapping of the human brain beyond Brodmann. Neuron 6, 1086-1107. ↩︎
11. DH Hubel & TN Wiesel (1962) Receptive fields, binocular interaction, and functional architecture in the cat’s visual cortex. Journal of Physiology 160, 106–54;
    DH Hubel & TN Wiesel (1972). Laminar and columnar distribution of geniculo-cortical fibers in the macaque monkey. Journal of Comparative Neurology 146, 421–450;
    CD Gilbert & TN Wiesel (1983) Functional organization of the visual cortex. Progress in Brain Research 58, 209–218;
    RJ Douglas & KA Martin (2004) Neuronal circuits of the neocortex. Annual Review of Neuroscience 27, 419-451. ↩︎
12. A Watanabe et al (2023) Local and long-distance organization of prefrontal cortex circuits in the marmoset brain. Neuron 111, 2258-2273. ↩︎
13. NL Jorstad et al (2023) Transcriptomic cytoarchitecture reveals principles of human neocortex organization. Science 382, adf6812;
    K Siletti et al (2023) Transcriptomic diversity of cell types across the adult human brain. Science 382, add7046. ↩︎
14. A Weninger & P Arlotta (2023) A family portrait of human brain cells. Science 382, 168-169. ↩︎
15. For example: The Allen Brain Map – https://portal.brain-map.org
    The Human Brain Project – https://www.humanbrainproject.eu/en/science-development/focus-areas/brain-atlases/
    The Human Protein Atlas – https://www.proteinatlas.org/humanproteome/brain
    BrainSpan Atlas of the Developing Human Brain – https://www.brainspan.org
    ↩︎
16. D Badre & M D’Esposito (2009) Is the rostro-caudal axis of the frontal lobe hierarchical? Nature Reviews Neuroscience 10, 659–669. ↩︎
17. M Petrides (2005). Lateral prefrontal cortex: architectonic and functional organization. Philosophical Transactions of the Royal Society of London, Series B Biological Sciences 360, 781–795. ↩︎
18. BTT Yeo et al (2011) The organization of the human cerebral cortex estimated by intrinsic functional connectivity. Journal of Neurophysiology 106, 1125–1165. ↩︎
19. S Grillner, A Kozlov, JH Kotaleski (2005) Integrative neuroscience: linking levels of analyses. Current Opinion in Neurobiology 15, 1–8. ↩︎