Project Topics
Get Complete Project Material File(s) Now! »
Evolution and specificities of the primate cerebral cortex
Evolution has produced a wide array of brain sizes and shapes (Figure 1.4). Brain mass varies by a factor of 100 000 in the mammalian order (Count, 1947) ranging from 0.060g for the insectivorous white-toothed pygmy shrew to 9.2kg for the sperm whale.
Do species with a larger brain have higher cognitive abilities? Many studies have tackled this issue using parameters relying on brain and body sizes but without establish-ing a clear relationship between their parameter and the actual intelligence degree of the di↵erent species. A recent study investigated the direct correlation between anatomical parameters and cognitive abilities in a meta-analysis of 24 primates (Deaner et al., 2007). They found that the best anatomical indicator of cognitive abilities was the absolute brain size and thus the number of neurons, independent of the body size. The authors claim that this rule applies to the primate order but elephants and cetaceans have much larger brains than humans but do not appear to have as high cognitive functions. There could be several non mutually exclusive explanations. Brain size might not be directly corre-lated to neuron numbers in a similar fashion in di↵erent mammalian clades. That is, even though cetaceans brains can weigh 7 times more than humans they might not contain 7 times more neurons. Indeed some recent studies showed that there might be clade specific scaling rules in mammalian brains (Herculano-Houzel, 2011; Charvet et al., 2013). It is also possible that having a bigger brain and more neurons is a necessary but not suffi-cient attribute for a more intelligent brain. Computational machine efficiency depends on the number of its processors (neurons) and on the speed of the information flux between these processors, thus on the properties of the connections and on the arrangement of the whole structure that will determine the distances to travel. In the next paragraphs we will review the di↵erent pieces of evidence we have for how primate brains get bigger and the qualitative di↵erences of the primate brain organization.
Are all brains made the same: a common scaling rule for cell number in all mammalian brains?
As seen in Figure 1.4 brain sizes are very heterogeneous in mammals. Do all brains result from a scaled up or a scaled down version of a same basic organization plan? In other words do brains of the same size contain the same number of neurons? Previous comparative neuroanatomical studies based on cellular densities and volumetric analyses argued for a common scaling rule for mammals (Haug, 1987; Stevens, 2001). In the brain, the cerebral cortex takes up a higher relative proportion and bigger brains are mainly constituted of cortex. In these studies, the calculation of cell densities has been carried out without taking into account the di↵erent clades and biased towards the cerebral cortex. The authors derived a general rule for a large group of mammals according to which bigger brains have more neurons, smaller neuronal densities and a higher glia to neuron ratio. They claimed that glial cells are the most numerous cell type of the brain outnumbering the neurons by up to 50 times (Kandel et al., 2000). However, these studies have been carried out in a restricted number of species and cell density measurements are subject to high discrepancies among studies, for example neuronal densities of the cerebral cortex varies up to 6-folds in di↵erent studies. Moreover, as brain tissue is heterogeneous in cell density, modern stereological techniques would necessitate segmenting the brain into hundreds of pieces of homogeneous cell density which would be a tedious and error-prone task.
A new technique has been recently been developed which overcomes the caveats of the above mentioned protocols and enables the determination of absolute numbers of neuronal and non-neuronal populations in di↵erent brain regions (Herculano-Houzel and Lent, 2005). The isotropic fractionator is a nonstereological approach which relies on transforming the brain tissue into a homogeneous isotropic cell suspension. Such suspen-sions can then easily be stained with a neuronal marker and neuronal and non-neuronal cells can be automatically counted. Contrarily to stereological neuron counting on brain sections, the isotropic fractionator relies on dissecting unstained brain tissue where mor-phological landmarks are spare. This technique can thus only be applied to large pieces of tissue, and is not adapted to resolve inter cortical/ inter layer di↵erences in neuron numbers, which can be a non negligible issue knowing the gradient of number of neurons in a cortical column along the rostro-caudal axis (Charvet et al., 2013). Moreover, several steps of the process need to be checked to fully validate the technique. The isotropic fractionator relies on generating a suspension of nuclei, but no definite proof of the re-tention of all nuclei has been given. In addition, the neuronal marker used NeuN is not an all-or-none staining, meaning that the discrimination between neuronal and non neu-ronal cells based solely on this marker might be error prone. More precise comparisons with stereological analyses need to be carried out in order to confirm the validity of this technique (Carlo and Stevens, 2013; Charvet et al., 2013). The following results need to be taken with care before they had been confirmed using other techniques taking into account possible local variation of neuron numbers.
The isotropic fractionator technique allowed the determination of brain cell numbers in various mammalian species and might lead to the revision of old dogmas. Over several years a dataset of 28 mammalian species belonging to 3 di↵erent clades has been gathered (Herculano-Houzel et al., 2006, 2007; Azevedo et al., 2009; Gabi et al., 2010; Sarko, 2009). This dataset enabled them for the first time to directly assess the cellular scaling rules at the gross structure level (cerebral cortex, cerebellum, rest of the brain) of di↵erent orders of mammals in a meta-analysis (Herculano-Houzel, 2011). The authors found di↵erent scaling rules for primates and rodents. For an equivalent body mass primates have a larger brain than rodents. The number of neurons scales also more rapidly with body mass for primates compared to rodents. Brain mass of rodents increases with the number of neurons raised to the power 1.5 whereas primate brain mass increases almost linearly. This clearly demonstrates that neuron numbers do not scale universally with brain mass (Figure 1.5), this is also true for the cerebral cortex. The larger the brain, the bigger the di↵erence of neurons between rodent and primate brains. This data invalidates brain size as a good indicator of neuron numbers. A rodent brain of 100g will contain 470 millions neurons, a primate brain of the same mass will contain 7.3 billions so more than 15 times more. Primate brains seem to be able to add neurons in a more economic way, less volume consuming than rodents.
Figure 1.5. Number of neurons are not well predicted by brain size among mammalian orders. The Owl monkey whose brain mass is slightly smaller than the Agouti contains almost twice as many neurons.Along the same line, the Capucin monkey with a 52g brain possesses more than twice as many neurons as the Capybara whose brain weight 76g. Figure from (Herculano-Houzel, 2009).
The distribution of mass in the brain does not reflect the number of neurons of the structures. A relatively larger cerebral cortex does not contain a higher proportion of neurons in bigger brains. The cerebral cortex, although accounting for up to 80% of brain mass in humans, usually contains about 20% of the brain neurons and the cerebellum 70-80% (Azevedo et al., 2009). Interestingly the cerebellum and the cerebral cortex have also clade specific scaling rules with neuronal number in which the structure mass increases faster with neuron numbers in rodents (Herculano-Houzel, 2011). The two structures seem to gain neurons in a coordinated fashion at an average rate of 4.2 neurons in the cerebellum for one in the cerebral cortex. This result supports the hypothesis of an integrated function of both structures and suggests that a common selective pressure might be at the origin of a concerted evolution (Balsters et al., 2010).
The neuronal density seems to be very heterogeneous among species and structures (Herculano-Houzel, 2011). The authors found the neuronal density to decrease as ro-dent brain size increases whereas contrarily to previous results primate neuronal density decreases only slightly when the brain size increases. Unlike neuron numbers, glial cell numbers have been reported to share common scaling rules in the di↵erent mammalian clades and even in the di↵erent structures. Glial cell number increases linearly with brain mass and their density is similar in the di↵erent brain structures. Brain size is actually a good indicator of glial cell number rather than neuron number. As glial cell density can be considered constant, the lower rodent neuronal density suggests a bigger neuronal size (soma and neuronal fibers) in rodents compared to primate as brain size increases.
Although the ratio glia/neuron had been shown to increase in bigger brains (Sherwood et al., 2006), using the isotropic fractionator the authors did not report this trend. They found that the ratio glial/neuron is heterogeneous among the di↵erent brain structures (in human, 0.99 on average but 3.76 in the cortex, 0.35 in the cerebellum and 11.35 in the rest of the brain; (Azevedo et al., 2009)). The ratio is found to increase only in the rodent brain as it increases in size. In primates the glia/neuron ratios remain constant in the di↵erent brain structures as brains scale up in size (Herculano-Houzel, 2011). Moreover, the glia/neuron ratio seems to be correlated to the average neuronal size. From this results Herculano-Houzel et al. proposed a mechanism to explain the relationship between neuron and glial cell numbers (Herculano-Houzel et al., 2006). Gliogenesis is known to be density-dependent (Zhang and Miller, 1996), glial progenitor proliferation stops at confluence probably by cell contact inhibition. As glial cell size does not scale with brain size, the glial/neuron ratio will depend only on the neuronal density, thus on neuronal size. It seems that glial cells properties are much more conserved among structures and clades than neuron cell properties indicating a stronger selection pressure and a fundamental role in the brain.
size as rodents (Figure 1.5).
It has long been thought that the human brain was exceptional in the mammalian order by its composition and cognitive abilities (Deacon, 1997). However, Azevedo et al. found using the isotropic fractionator technique that the human brain is not an exception to the primate scaling rule (Azevedo et al., 2009). With its 86 billion neurons and 85 billion glial cells for 1.5kg, the human brain deviates only by 10% from the the primate scaling rule predictions. The relative size of the cerebral cortex is increased (82%) but it takes up only 19% of the brain’s neurons. The human brain possesses two advantages compared to other mammalian brains. It is built according to the economical, space-saving primate scaling rules and it is the largest among those hence the one containing the most neurons. The number of neurons of the cetaceans and the elephant has not been assessed by the isotropic fractionator method. Based on previous estimates the false killer whale and the African elephant would have 11 billion neurons in the cerebral cortex, only slightly fewer than the 11.5 billions found in the human cerebral cortex using the same method (Roth and Dicke, 2005). They seem to obey rather the rodent than the primate scaling rules, showing a very low neuronal density. Even if these estimates of number of neurons are true, the di↵erence between humans and elephants or whales is small, if any. The higher cognitive function achieved in the human brain might lie beyond the number of neurons.
Table of contents :
1 Introduction
1.1 The primate cerebral cortex: organization and evolution
1.1.1 The cerebral cortex: general organization and composition
1.1.2 Radial organization of the cerebral cortex
1.1.3 Tangential organization of the cerebral cortex
1.1.4 Evolution and specificities of the primate cerebral cortex
1.2 Corticogenesis: lessons from rodent models
1.2.1 Overview of mouse cortical development
1.2.2 Specification of the cerebral cortex territories
1.2.3 Inside-out generation of the cortical layers
1.2.4 Origin of excitatory neurons
1.3 Toward a better understanding of primate corticogenesis
1.3.1 The primate cortical lamination is also inside-out
1.3.2 A more complex early cortical development in primates
1.3.3 Multiple origins of primate interneurons?
1.3.4 An extra proliferative zone: the OSVZ
1.3.5 Towards the identification of the primate cortical precursors heterogeneity and diversity
1.3.6 Mechanisms of cortical gyrification
1.3.7 Hypotheses for primate cortical expansion
2 Results
2.1 Paper in press at Neuron: Precursor diversity and complexity of lineage relationships in the outer subventricular zone of the primate
2.1.1 Abstract
2.1.2 Introduction
2.1.3 Results
2.1.4 Discussion
2.1.5 Material and methods
2.1.6 Acknowledgements
2.1.7 References
2.1.8 Figures
2.1.9 Supplemental information
2.2 Cell lineage analysis in the primate germinal zones based on a Hidden Markov Tree model
2.2.1 Introduction
2.2.2 Methods
2.2.3 Results
2.2.4 Discussion
3 General discussion
3.1 Mechanisms for neuronal output amplification in primate cortical development
3.1.1 Candidate proteins for primate cortical expansion : proteins involved in rodent BP pool amplification
3.1.2 A time-dependent cell cycle duration / mode of division regulation allowing an upsurge of proliferation
3.2 Diversity of precursor types and complex lineage relationships in the primate developing cortex
3.2.1 5 types of precursors in the primate OSVZ
3.2.2 An emerging neuronal lineage complexity during evolution?
3.3 Hypothesis for the origin of precursor type diversification and lineage relationships complexification during evolution
3.3.1 The metaphor of Waddington epigenetic landscape
3.3.2 Di↵erent landscapes in mouse and primate cortical developments?
3.4 General conclusion
