Figure 5. The problem of network allometry is represented by two neural circuits that exhibit local and global connectivity, respectively. Once the brain has grown to a point where the bulk of its mass is in the form of connections, then further increases as long as the same ratio in interconnectivity is maintained will be unproductive. Increases in number of units will be balanced by decreased performance of those units due to the increased conduction time.
This implies that large brains may tend to show more specialization in order to maintain processing capacity. Indeed, an increase in the number of distinct cortical areas with increasing brain size has been reported Kaas, , ; Striedter, It may even explain why large-brained species may develop some degree of brain lateralization as a direct consequence of size. If there is evolutionary pressure on certain functions that require a high degree of local processing and sequential control, such as linguistic communication in human brains, these will have a strong tendency to develop in one hemisphere.
Although the cerebral cortex is not the only brain structure which was selected for in evolution to expand, as a result of growing environmental pressure for more sophisticated cognitive abilities, it has played a key role in the evolution of information processing in the mammalian brain. The primate cortex, as we have seen, has evolved from a set of underlying structures that constrain its size, and the amount of information it can store and process. If the ability of an organism to process information about its environment is a driving force behind evolution, then the more information a system, such as the brain, receives, and the faster it can process this information, the more adequately it will be able to respond to environmental challenges and the better will be its chances of survival Hofman, The processing or transfer of information across cortical regions, rather than within regions, in large-brained primates can only be achieved by reducing the length and number of the interconnective axons in order to set limits to the axonal mass.
The number of interconnective fibers can be reduced, as we have seen, by compartmentalization of neurons into modular circuits in which each module, containing a large number of neurons, is connected to its neural environment by a small number of axons.
The length of the interconnective fibers can be reduced by folding the cortical surface and thus shortening the radial and tangential distances between brain regions.
Local wiring—preferential connectivity between nearby areas of the cortex—is a simple strategy that helps keep cortical connections short. In principle, efficient cortical folding could further reduce connection length, in turn reducing white matter volume and conduction times Young, ; Scannell et al. Thus the development of the cortex does seem to coordinate folding with connectivity in a way that could produce smaller and faster brains.
Wang et al. They found that the composition of white matter shifts from compact, slow-conducting, and energetically expensive unmyelinated axons to large, fast-conducting, and energetically inexpensive myelinated axons. Delays and potential imprecision in cross-brain conduction times are especially great in unmyelinated axons, which may transmit information via firing rate rather than precise spike timing.
Another way to keep the aggregate length of axonal and dendritic wiring low, and with that the conduction time and metabollic costs, is to increase the degree of cortical folding. A major disadvantage of this evolutionary strategy, however, is that an increase in the relative number of gyri can only be achieved by reducing the gyral width.
At the limit, the neurons in the gyri would be isolated from the remainder of the nervous system, since there would no longer be any opening for direct contact with the underlying white matter. A further increase in the size of the brain beyond that point, i. The remarkably high correlation between gray matter, white matter and brain size in anthropoid primates ensures that the proposed model can be used for predictive purposes to estimate the volume of white matter relative to brain volume for a hypothetical primate Hofman, b.
Model studies of the growth of the neocortex at different brain sizes, using a conservative scenario, revealed that with a brain size of about cm 3 the total volume of the subcortical areas i. Increasing the size of the brain beyond that point, following the same design principle, would lead to a further increase in the size of the neocortex, but to a reduction of the subcortical volume. Consequently, primates with very large brains e.
Figure 6. The number of connections C , cortical processing units U and level of interconnectivity I in the primate neocortex as a function of brain size. Semi-logarithmic scale.
Values are normalized to one at a brain volume of cm 3 , the size of a monkey brain. Note that the number of myelinated axons increases much faster than the number of cortical processing units see also Figure 5. The human cerebrum, for example, contains 6 times more myelinated axons than that of a rhesus monkey, whereas the number of cortical processing units is only 3 times larger.
Dashed lines show the potential evolutionary pathway of these neural network elements in primates with very large brains, i. Note that a further exponential growth in the number of cortical processing units, without an increase in the number of connections, will lead to a decrease in connectivity between these units and thus to more local wiring.
A progressive enlargement of the hominid brain started about 2. Since then, a threefold increase in endocranial volume has taken place, leading to one of the most complex and efficient structures in the animated universe, the human brain. Explanations for the evolution of the human brain are mainly focusing on selection pressures of the physical environment climate, diet, food availabilty and those of the social environment group size, coalition formation, parental care.
Although attempts have been made to discriminate between ecological and social theories see e. Obviously, they all play a role in explaining the marked differences in brain size between humans and apes, but in which way and to what extent is far from clear at this moment. We will need better studies of cognition and behavior, along with comparative brain studies, to answer these questions.
Despite these difficulties in explaining the selection pressures of the evolution of the human brain, comparative neuroimaging studies in primates have identified the underlying neural substrate and unique features of the human brain for a review see, Rilling, These studies, for example, have clarified how the dramatic differences in brain size between humans and chimpanzees develop.
First, human brains are already twice as large as chimpanzee brains from an early point in gestation 16 weeks. Although both show an increase in growth velocity at this time, they diverge sharply at 22 weeks of gestation, when human brain growth continues to accelerate, whereas chimpanzee brain growth decelerates.
Finally, during early infancy, humans experience a very rapid increase in white matter volume that significantly exceeds that found in chimpanzees see e. In view of the central importance placed on brain evolution in explaining the success of our species, one may wonder whether there are physical limits that constrain its processing power and evolutionary potential. The human brain has evolved from a set of underlying structures that constrain its size, and the amount of information it can store and process.
In fact, there are a number of related factors that interact to limit brain size, factors that can be divided into two categories: 1 energetic constraints and 2 neural processing constraints see e.
The human brain generates about 15 watts W in a well insulated cavity of about cm 3. From an engineering point of view, removal of sufficient heat to prevent thermal overload could be a significant problem.
But the brain is actively cooled by blood and not simply by heat conduction from the surface of the head. So the limiting factor is how fast the heat can be removed from the brain by blood flow. So, to increase cooling efficiency in a larger brain, either the blood must be cooler when it first enters the structure, or the flow-rate must be increased above current levels.
Another factor related to blood flow has to do with the increasing energy requirements of a larger brain, a problem that is exacerbated by the high metabolic cost of this organ. A bigger brain is metabolically possible because our cardiovascular system could evolve to transport more blood at greater pressure to meet the increased demand. This should not be taken to imply that thermal and metabolic mechanisms play no role at all in setting limits to brain size.
Ultimately, energetic considerations will dictate and restrict the size of any neuron-based system, but as theoretical analyses indicate, thermal and metabolic factors alone are unlikely to constrain the potential size of our brain until it has increased to at least ten times its present size Cochrane et al. The same holds for extrinsic developmental constraints that have to do with pelvic anatomy related to bipedalism , parturition, and maternal and fetal mortality. Although these factors are relevant in human evolution it is unlikely that they are setting limits to human brain growth.
It means that natural selection operates on brain size at the expense of growth and reproduction, which could explain its correlation with life span Hofman, This evolutionary strategy is most obvious when considering the evolution of our own species, where there has been a presumed twofold increase in life span associated with a more than threefold increase in brain size in a mere 2.
The limit to any neural system lies in its ability to process and integrate large amounts of information in a minimum of time and therefore its functional capacity is inherently limited by its neural architecture and signal processing time. The scaling model of the geometry of the neocortex, for example, predicts an absolute upper limit to primate brain size Hofman, b ; Figure 7.
Without a radical change in the macroscopic organization of the brain, however, this hypothetical limit will never be approached, since at that point ca. Figure 7. Relative subcortical volume as a function of brain volume. The predicted subcortical volume i. Likewise, the subcortical volume will be zero when the brain is exclusively composed of cortical gray and white matter.
At a brain size of cm 3 the subcortical volume has a maximum see also Figure 6. The larger the brain grows beyond this critical size, the less efficient it will become. Assuming constant design, it follows that this model predicts an upper limit to the brain's processing power. Modern humans are at about two-third of that maximum. Modified with permission from Hofman b. Cochrane et al. They argue that the human brain has almost reached the limits of information processing that a neuron-based system allows and that our evolutionary potential is constrained by the delicate balance maintained between conduction speed, duration of the electrical pulse pulse width , synaptic processing time, and neuron density.
Any further enhancement of human brain power would require a simultaneous improvement of neural organization, signal processing and thermodynamics. Such a scenario, however, is an unrealistic biological option and must be discarded because of the trade-off that exists between these factors.
Of course, extrapolations based on brain models, such as the ones presented here, implicitly assume a continuation of brain developments that are on a par with growth rates in the past. At a brain size of about cm 3 , corresponding to a brain volume two to three times that of modern man, the brain seems to reach its maximum processing capacity.
The evolution of the neocortex in primates is mainly characterized by the development and multiplication of clusters of neurons which are strongly interconnected and in physical proximity. Since these clusters of neurons are organized in vertical columns, an increase in the number and complexity of the neuronal networks will be reflected by an expansion of the cortical surface area beyond that expected for geometric similar brains.
As a result the cortical surface area fractally evolves into a volume with increasing brain size. It is evident that the potential for brain evolution results not from the unorganized aggregation of neurons but from cooperative association by the self-similar compartmentalization and hierarchical organization of neural circuits and the invention of cortical folding, which reduces the interconnective axonal distances.
The competing requirements for high connectivity and short conduction delay may lead naturally to the observed architecture of the primate neocortex.
The similarity in brain design among primates, including humans, indicates that brain systems among related species are internally constrained and that the primate brain could only evolve within the context of a limited number of potential forms. It means that internal factors of brain design may be the primary determinants constraining the evolution of the brain and that geometric similarity among species in the functional organization of the brain may be derived from a common ancestor rather than being immediately evolved in response to specific environmental conditions.
The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Aboitiz, F. From tetrapods to primates: conserved developmental mechanisms in diverging ecological adaptations. Brain Res. Allen, J. Cambridge, MA: Belknap Press. Azevedo, F. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. Bayly, P. Mechanical forces in the cerebral cortical folding: a review of measurements and models.
Bohland, J. A proposal for a coordinated effort for the determination of brainwide neuroanatomical connectivity in model organisms at a mesoscopic scale.
PLoS Comput. Budd, J. How do you wire a brain? Communication and wiring in the cortical connectome. Bullmore, E. The economy of brain network organization. Buxhoeveden, D. Minicolumn size and human cortex. The minicolumn hypothesis in neuroscience. Brain , — Neuronal oscillations in cortical networks. Science , — Scaling brain size, keeping time: evolutionary preservation of brain rhythms.
Neuron 80, — Changizi, M. Principles underlying mammalian neocortical scaling. CrossRef Full Text. Parcellation and area-area connectivity as a function of neocortex size. Brain Behav. Charvet, C. The third idea is that the gray matter grows more than the white matter, leading to a "buckling" that gives the cortex its shape, the researchers said. But earlier attempts to model this buckling were not successful, the researchers said.
In previous studies, researchers assumed that the gray matter is a thin, stiff layer growing atop of a thick, soft base of white matter, but this assumption yielded wrinkles that aren't like the ones in real human brains. In the new study, the researchers assumed that the gray and white matter have similar stiffness, but different growth rates. Using mathematical simulations, they showed that depending on the size of the brain, their model results in different shapes of brain surfaces.
For example, for a small brain with a diameter of less than half an inch, the brain surface is predicted to be smooth. In essence, this expansion causes pressure to increase in that outer surface, which is then mitigated by folding, Ronan, told Live Science.
Basically, imagine pushing at either end of a piece of rubber — at some point, the surface will bend in response to the increasing pressure. Or, if you're into geology, think of it like two tectonic plates crashing into each other: The pressure during the collision eventually becomes so great that those plates experience a geological fold.
These countless folds allow humans to pack in more neurons which, in turn, can mean more advanced brains with increased cognitive abilities , Ronan said. The alternative would be a cortex that did not fold but expanded like a balloon, an inefficient use of the cramped quarters of the head.
Again, paradoxes abounded. Human cortices have three times the number of neurons than elephant cortices, yet human brains are half the mass and far less folded. Baboon and pig cortices display equivalent amounts of folding, yet baboons have 10 times as many neurons as pigs do. These outliers seemed to suggest that different species possess different mechanisms to control cortical folding—that is, each species has its own way of growing a brain.
The Science study both disproves those hypotheses and reconciles anomalies like the manatee with a simple physical law. Whether or not a brain folds, Herculano-Houzel says, is pure physics. To model brain folding Herculano-Houzel had graphed a power function derived from the product of cortical surface area and the square root of cortical thickness. Mota had recognized that the same model that predicted the degree of brain convolution also explained the crumpling of paper balls.
To see their results you can run essentially the same experiment as the experimenters conducted using four sheets of paper. First, take one sheet, crumple it hard, and set it down.
The paper should expand slightly as it releases some energy but eventually it will settle in a crumpled form. To get a feel for this concept, Mota suggests imagining dropping a ball in a bowl. The ball will settle at the bottom of the bowl, the position of least effective free energy where the ball can remain with least effort. Like crumpled paper, a folded brain will not stray from this new stable state.
0コメント