Paleoanthropology of cognition: an overview on Hominins brain evolution

. Recent advances in neurobiology, paleontology, and paleogenetics allow us to associate changesinbrainsizeandorganizationwiththreemain“moments”ofincreasedbehavioralcomplexity and, more speculatively, language development. First, Australopiths display a signiﬁcant increase in brain size relative to the great apes and an incipient extension of postnatal brain development. However, their cortical organization remains essentially similar to that of apes. Second, over the last 2 My, with two notable exceptions, brain size increases dramatically, partly in relation to changes in body size. Di ﬀ erential enlargements and reorganizations of cortical areas lay the foundation for the “language-ready” brain and cumulative culture of later Homo species. Third, in Homo sapiens , brain size remains fairly stable over the last 300,000 years but an important cerebral reorganization takes place. It a ﬀ ects the frontal and temporal lobes, the parietal areas and the cerebellum and resulted in a more globular shape of the brain. These changes are associated, among others, with an increased development of long-distance—horizontal—connections. A few regulatory genetic events took place in the course of this hominization process with, in particular, enhanced neuronal proliferation and global brain connectivity.


La version française de l'article est disponible à la suite de la version anglaise 1. Introductory remarks
Hominins evolution is not a linear process but the complex ramification of many lineages that diverged, coexisted and even sometimes partly merged.It is therefore difficult to define well delimited stages, as, at almost any given period of the past, several forms of Hominins coexisted on earth, usually in different parts of the globe, but sometimes in the same continental regions.Each of these lineages displayed its own mosaic of features, from anatomy to diet and from life history to social organization.Whenever one can analyze the evolution of such lineages, it is quite obvious that they display significant changes through time.For example, the endocranial organization of the earliest forms of Homo sapiens is closer from that of Neandertals than from that of late (so called "modern") forms of Homo sapiens [1].Furthermore, when analyzing hominin evolutionary processes, homoplasy (convergent evolution without recent common origins or reversal to primitive conditions) is a major issue that should not be overlooked.These changes are expected to originate from variations at the genome level which can be examined from entire sequences of fossil DNAs in recent forms such as Neandertals [2] and Denisovans [3] and/or by interpolation between extant apes and modern Humans.Even though, the deterministic relationships of precise genetic events which define brain circuits still largely escape our understanding, preliminary attempts have been made to more closely relate genome variations with the evolution of larger integrated entities that we have referred to as global "con-nectomic fundamentals" [4] such as brain size, longrange connectivity or extended postnatal synaptogenesis.Beyond genome, the intrinsic properties of developing tissues, interactions between adjacent anatomical regions and more widely integration phenomena also contributed to the neuroanatomy of different species of Hominins [5].Finally, from a paleontological point of view, one can essentially deal with size, shape, patterns of development and cultural artefacts but brain connectivity and language are issues notoriously difficult to address.As a consequence, the present attempt to more closely relates paleoanthropology and brain evolution-even if of interest-shall remain partly speculative.

Australopiths
Australopiths represent a paraphyletic group documented since 4.2 My ago and from which the genus Homo arose ca.2.8 My ago [6].The last advances on Australopiths brain evolution demonstrate that this group of species is characterized by brain sizes in average 30% bigger than the means measured among common chimpanzees [7].However, the anatomy of the basicranial foramina shows that brain blood flow rate was lower and therefore brain metabolic rate was lower than in extant great apes and much under the values observed in extant Homo sapiens [8].For instance, between 4.4-million-yearold pre-Australopith Ardipithecus and Homo sapiens, brain volume increased 4.7-fold, but flow rate in the internal carotid arteries increased 9.3-fold, indicating an approximate doubling of metabolic intensity in brain tissue.By contrast, among haplorhine primates, flow rate in the internal carotid arteries is proportional to brain volume suggesting a constant volume-specific brain metabolic rate in these species.
In the course of Hominins evolution, the deviation from allometric relationships between brain size increases and brain blood flow rate indicates a marked increase of metabolic intensity of brain tissue in later forms [8].Along these lines [9] it was emphasized that in Hominins brain energy metabolism relies not only on the number of neurons in the neocortex but the complexity of their dendritic arborization-the number of spines and synapses-which increases "more than linearly" in the ultimate steps of hominization.In particular, the high density of synaptic contacts on dendritic spines in humans which creates an electrochemical ion-gradient disruption upon activation imposing higher energy demands [10][11][12].Given the existence of a tight metabolic coupling between neurons and glia, astrocytes in particular, the overall astrocyte/neuron ratio increases in humans.This happens together with a massive enhanced expression of genes involved in energy production in the human neocortex [13,14], particularly those involved in the electron chain transfer of ATP-producing oxidative phosphorylation [15,16].All this might be viewed as a new connectomic fundamental, that we may referred to as "brain energetics", essential in the general process of brain hominization, and which deserves further investigations along these lines.
In Australopiths, the pattern of cortical organization remains primitive and essentially ape-like (Figure 1).In particular one does not observe a reduction of the visual primary cortex and an expansion of the adjacent parietal areas, once believed to be already present in Australopiths [17].However, also notable is the prolonged period of brain development relative to chimpanzees.This change in the life history pattern is likely to be related to the energetics requirements of a larger brain in a less productive environments.It provided a basis for the later evolution on a protracted human-like growth pattern critical to an extended brain plasticity and the establishment of a long period of childhood learning through synapse selection already considered as a major connectomic fundamental [4].
From the point of view of behaviour and social organization, one cannot find extant models for Australopiths in present-days large Primates.It is important however to underline, that the strong canine reduction and the decrease of dental sexual dimorphism documented among Hominins since their emergence, are almost fully achieved in Australopiths.These features are consistent with a reduction of interindividual aggressivity (especially between males) and reproductive behaviours as well as social organisations departing from those observed among living apes.In other words, Australopiths would already show cooperative relationships reminiscent of later human societies as manifested by the first evidence of the prolonged period of childhood epigenetic imprinting and social learning unique to humans.

Early Homo
Several species of early Homo have been identified in the Early Pleistocene fossil record of Africa: Homo habilis, Homo rudolfensis and early Homo erectus ("Homo ergaster").They widely overlap chronologically and represent distinct lineages sharing a common ancestor but displaying their own evolutionary trajectories.Among them, only Homo erectus survived after 1.4 million years ago and is most generally considered ancestral to later Hominins.Before this date, absolute brain sizes are not much different between the different species of early Homo and range between 510 and 850 ml, with individuals of the three taxa plotting in the lower values of this range [18].In term of cortical organization, early forms of the genus Homo continue to retain a primitive pattern, in particular with a great ape-like organization of the frontal lobe.The cerebellum also remains relatively small compared to that of recent Homo sapiens.

Homo erectus
The oldest representatives of Homo erectus sensu lato are dated circa 1.9 million years before present in east and south Africa.This date also corresponds to major behavioural changes observed in the archaeological record, in particular regarding the increase of predation on mid-to large size mammals, access to aquatic food and related dietary habits.100,000 years later one also finds the first undisputable evidence for the presence of the genus Homo outside of Africa and  [17]).Both species display a similar organization of sulci and gyri.C, sulcus centralis; fs, frontalis superior; fm, frontalis medius; fi, frontalis inferior; fo, fronto-orbitalis; h, horizontal ramus of pci; ip, s. intraparietalis; pci, praecentralis inferior; pcs, praecentralis superior; ps, parietalis superior; pti, postcentralis inferior; ptm, postcentralis medius; pts, postcentralis superior; L, s. lunatus; ts, temporalis superior; ts-a, ramus temporalis superior; tm, temporalis medius; occi, occipitalis inferior; lc, s. calcarinus lateralis; u, s. calcarinus ramus superior; cereb, cerebellum; ld, lambdoidal suture.Scale bar is 1 cm. the development of new stone technologies including, in Africa, the production of large cutting tools such as Acheulean bifaces.
The Homo erectus fossil record covers a very long span of time, from 1.9 million years before present in Africa to maybe as late as 125,000 years ago in Java, and a huge geographical range over all Africa and the southern and middle latitudes of Eurasia.Understandably, the phenotype of the species displays large variations, probably still underestimated.Regarding brain size, the largest values are reached by terminal Indonesian specimens (1250 ml in Solo 5), in strong contrast with the oldest forms of the species (546 ml in Dmanisi 4500).Derived conditions of the cortical organization seem to develop between 1.7 and 1.5 million years ago [19], later than the first dispersals of Homo from Africa.They include the differential enlargement and reorganisation of the frontal lobe specially the inferior prefrontal cortex and Brodmann area (BA) 45/47, in tandem with the posterior parietal cortex, the temporal lobe and to some extent the occipital lobe (Figure 2).Together with BA44, BA45 make up Broca's area, known to be involved in language processing in Homo sapiens.Furthermore, in today's humans patients lesions of BA45 lead to expressive aphasia and BA47 is implicated in the processing of syntax in oral and sign languages, musical syntax, and semantic aspects of language [20] together with the facial expression of emotions [21].Such frontal lobe evolutions might have constituted the foundations of the "language-ready" brain [19] of later Homo species [22].It might also represent some intermediate-still unknown-form of communication between chimp grunts, calls or signs and properly human language.
Life history of Homo erectus is another aspect of their evolution than have been much debated.For the oldest forms (>1.5 million years old), dental microstructures suggest a timing of development still reminiscent from that of extant apes and faster than that of recent humans [23].Adulthood might have been reached by 12 years of age [24].Evidence from neurocranial individual development also indicates a proportional volumetric growth of the brain at best intermediate between that of apes and that of Homo sapiens [4,25,26].Neural maturation was therefore likely faster in Homo erectus than in later Hominins, with some implications on the length of the period of brain plasticity during growth.Later weaning period followed by faster somatic development and sexual maturation also have implications on the lesser degree of human-like cooperative breeding that early Homo erectus might have displayed.
It has been suggested that the variation in early Lower Paleolithic stone artifacts produced by Homo erectus may "reflect a heavier reliance on behaviors that reside within the zone of latent solutions than on behaviours that make use of high-fidelity social learning" [27].This would imply that genuine cumulative and transmissible culture might have been put fully established in the course of human evolution only with the late representatives of Homo erectus.

Large brained hominins
During the Middle Pleistocene, several lineages evolved large brains close in size to present-days means or even above.They include the latest representatives of Homo erectus in Indonesia (Ngandong series), but they are primarily represented by a group sharing a common ancestor circa 650,000 years ago and include Homo sapiens, Neandertals and Denisovans.Brain volume increase occurred independently and in parallel in these lineages, most likely in relation with similar pressures of selection generated by the emergence of more complex behaviours.
Changes in the endocranial organisation through time are best documented in Neandertals and in Homo sapiens.Both lineages display a broadening of the frontal lobes, although it is more marked in the latter group.However, Neandertals (and likely Denisovans) essentially evolved larger brains along the same allometric trajectory than more primitive representatives of the genus Homo such as late Homo erectus [28].In contrast, in the course of the last 300,000 years, Homo sapiens brain size have remained essentially constant [28,29], but display a Figure 3. Left: comparative form analysis (PCA) incorporating a size factor (along PC1) of the endocranial shape of H. erectus, H. neanderthalensis and H. sapiens (from [28]).Successive chronological groups of H. sapiens (numbered 1-3) are represented by purple polygons.Neandertals are represented in red and H. erectus in green.Right: comparison of the endocranial shape of the La Chapelle-aux-Saints Neandertal (in red) with that of a recent H. sapiens (in blue).significant reorganisation in shape (Figure 3) and likely connectivity [28].These changes are rather fast and affect in a concomitant manner altogether different parts of the brain and have resulted in a more globular shape of the brain in recent forms of Homo sapiens.The bulging of the parietal areas is one of the most visible changes in the endocranial shape, but does seem to be related to an increase of the parietal cortex surface.This bulging might instead be possibly related to the development of some internal structures (pre-cuneus and putamen have been discussed as possible candidates).Changes also affect the temporal lobes [30].However, most spectacular is the reorganization of the cerebellum with an increase in size of some of its parts.
Paleogenetic and genetic data further reveal that Neanderthal alleles on chromosomes 1 and 18 are associated with reduced endocranial globularity.These alleles influence expression of two nearby genes, UBR4 and PHLPP1, which are involved in neurogenesis and myelination of the cerebellum, respectively [31].The analysis of transcriptomic data in the region of human genome that have been cleared by selection of Neandertal introgressions ("Neandertal genomic deserts") indicate that structures outside de neocortex such as cerebellum, striatum and the mediodorsal nucleus of the thalamus have been under a strong positive selection [32].Also suggesting a strong positive selection, a series of non-synonymous mutations affecting coding genes are absent in Neandertals and Denisovans but present in virtually all extant humans [2].To date few of these changes have been deciphered, but several are involved in brain development and connectivity.The modern variants of KIF18a and KNL1 are proven to cause metaphase prolongation and fewer chromosome segregation errors during cell divisions in apical progenitors of developing neocortex [33].The modern version of TKTL1 results in an increase in cortical neuron production and increases the ability of the basal radial glia to self-amplify [34].Although rather fast, at the geological scale, these evolutionary changes seem to be gradual through time and it is only between 100,000 and 50,000 BP that endocranial shapes of fossil Homo sapiens start entering the present-day range of variation [28].
A significant inverse relationship of globularity and grey matter was noted in large parts of the frontal, temporal, and occipital gyri; parts of the cerebellum; and several subcortical regions (thalamus, putamen, hippocampus) [31] to the unmentioned benefit of the white matter (see also [35]).A plausible connectomic backbone which would integrate-"link together"-these widely distributed territories of the brain (and make the brain "more globular" or compact) might be hypothesized as an enhanced development of the long-range horizontal connectivity of the brain [4].This high-order core network has been postulated to contribute to the global neuronal workspace (GNW) [36,37] and suggested to be mobilized by conscious processing, higher cognitive functions and global integration [37,38].Interestingly the cerebellum-which was not mentioned in the original version of the GNW-might according to the above mentioned paleontological data [31] be viewed as one of its basic components (see also [39] in the case of the genetic predispositions to autism).This possibility was evoked quite a while ago [40] but still remains to be further demonstrated.
It is highly significant that the globular brain shape differenciating recent adult Homo sapiens from other Hominins results from the appearance of a new phase of development taking place during the first year of post-natal development [41].During this critical period, the dendritic arborizations of cortical pyramidal neurons and the long-range connectivity actively develop together for up to 15 years and even later [42].The development of the baby brain progresses as a multistep nested foliation resulting from successive waves of synapse outgrowth, stabilization, and elimination processes, which formally resemble an evolutionary "Darwinian" process [43][44][45][46][47].In developing monolingual children, one-word utterances are produced between 12 and 18 months, and vocabulary development "explodes" at 18 months of age [48], then sentences develop from 2-3 years until complete sentences of 4-5 words at age 4. On the other hand, neither monkeys nor chimpanzees ever learn to combine words into a multiword "utterance" [49,50].The ability to process hierarchically structured sequences resulting in a new higher-order element been assigned to a subpart of Broca's area, BA44 and a fronto-temporal language network connecting in the left hemisphere language-relevant regions via dorsally located white matter fiber tracts [4,50].Intriguingly, around 2 years, characteristic changes in the postnatal maturation of pyramidal projection neurons which originate from upper layers-specifically layer IIIc-in the human prefrontal cortex take place [4,[51][52][53].These neurons exhibit a unique differential increase and epigenetic pruning in the number of segments and length of their basal dendrites between 16 months and 2.5 years especially in the prefrontal cortex (ref in [4,51]).They are a major integrative element of cortical processing and might possibly contribute to the more globular shape of Homo sapiens brain.
One may speculate that advanced theory of mind and high order language processing depend on the fine tuning of these long-range connections and thus possibly contribute to the observed difference with archaic hominin brains.
Brain asymmetry that underlies behavioral and anatomical hemispheric specialization for complex cognitive abilities such as speech production is commonly interpreted as "unique" to the human brain.Examination of endocranial shape in humans, chimpanzees, gorillas, and orangutans, respectively all reveal a mean deviation from symmetry.These findings also indicate a decreased correlation and thus higher variability of frontal, occipital, and cerebellar asymmetries in humans as compared to nonhuman apes [54].This might possibly result from the differential postnatal out growth and synaptic epigenesis of some cortical territories typically illustrated with Broca's area and language learning.
From a behavioral point of view, the second the Middle Pleistocene witnessed numerous changes.Although pre-existing, the use of fire by Hominins become gradually more common [55] and is well documented in the archaeological record after 400,000 BP.Gathering around a common fire may have enhanced oral communication and its diversification within members of the social groups.In Acheulean assemblages produced by groups ancestral to Neandertals and early Homo sapiens, the occurrence of highly symmetrical bifaces indicates the development of an esthetic sense and the high investment in some artifacts production beyond the functional needs.Late Middle and Upper Pleistocene Hominins were capable of advanced technical skills and, to variable extents, of "symbolic" comportments.However, the frequency of "non-utilitarian" objects in the archeological record of African Homo sapiens after 140,000 BP [56] remains, to date, unparalleled in other groups of large brained hominins of the same time period.After the last expansion of terminal ("modern") Homo sapiens outside of Africa and south east Asia starting 50,000 years ago and the replacement of all other forms of Hominins, another major step was taken with the massive production of figurative "art" representing animals, humans and imaginary creatures [57].This peculiar cultural trajectory proper to our species may have contributed to the selection of late features of human brain connectivity, eventually referred to as belonging to a "cultural brain" [4].

Late small brained Hominins
The increase in encephalization level is a major trend in hominin evolution since the Australopiths, with an acceleration in the course of the last million years.This increase has impacted many other aspects of Hominin adaptations ranging from life history to metabolic rate, and from technological development to social networking.However, in the past two decades, at least two lineages of fossil hominins escaping to this general pattern have been identified.One is represented by Homo floresiensis (from Flores Island, Indonesia), a minute creature with a stature circa 1.0 m and an endocranial capacity of 425 ml [58].Homo floresiensis survived in isolation until the arrival of modern Homo sapiens of African origin in the region, 50,000 years ago [59].It is generally interpreted as an example of insular dwarfism.The second example is represented by Homo naledi, discovered in two nearby caves in the vicinity of Johannesburg, South Africa.Its geological age is estimated around 300,000 years.Homo naledi was somewhat (ca.1.45 m) taller than H. floresiensis, but, considering its age, it also displays a remarkably small cranial capacity ranging between 460-560 ml [60].
The presence of derived frontal lobe morphologies in both species, and the mosaic of skeletal features they display suggest that they possibly evolved from some earlier Homo erectus local populations.Alternatively, if the emphasis is put on some of their primitive features, these derived conditions may result from homoplasy.The cranial capacities of the two forms lay in the range of Australopiths and early Homo, indicating at best a stagnation (H.naledi) or even a regression (H.floresiensis) of their level of encephalization.Such trends might result from limitations in the capabilities of these species to extract high amounts of energy from their environments.Although this is a rather common situation in insular environments, this pattern is more puzzling in the case of Homo naledi.The latter species did not evolve in the complete geographical isolation of an island and must have been exposed to the competition of large brained primitive Homo sapiens documented in the southern part of the African continent.A peculiar ecological niche and/or dietary adaption [61] likely explain the oddity of this Hominin, which, so far, has not been found in direct association to lithic artefacts or any kind.

Conclusions
If most speculatively one can tentatively index the evolution of the human brain upon plausible neural signatures of language evolution one may distinguish there main "moments" within a highly complex ramification of many lineages.Australopith are characterized by brain sizes in average 30% bigger than the means measured among common chimpanzees with a pattern of cortical organization which remains primitive and essentially ape-like.Also notable is the first evidence for a prolonged period of brain development relative to chimpanzees anticipating the postnatal period of childhood learning unique to humans.Australopiths would also show early signs of cooperative relationships proper to human society.In other words, the connectomic fundamentals related to brain size increase and extended postnatal development look already at work.
During the following 2 My, the brain size increases dramatically, partly in relation to changes in the body format.This increase includes the differential enlargement and reorganization of the frontal lobe specially the inferior prefrontal cortex and Broca language areas BA 45/47, in tandem with the posterior parietal cortex, the temporal lobe and to some extent the occipital lobe.They produce the foundations of the "language-ready" brain of late H. erectus and following Homo species which developed a rich social life and a diversity of techno-cultures.The connectomic fundamental concerned by the increase of brain size and especially the increase of neurons numbers and cortical areas seem to dominate the picture.
In recent forms of Homo sapiens, there is a modest increase in brain size, however with a significant reorganisation resulting in a more globular shape of the brain affecting concomitantly altogether the frontal and temporal lobes, the parietal areas and the cerebellum.A plausible connectomic backbone might "link together" these widely distributed territories of the brain possibly due to an enhanced development of the long range-horizontal-connectivity of the brain.In Homo sapiens, the postnatal development of the baby brain maturation extends considerably up to 15 years and in particular between 1.5 and 2.5 years the dendritic arborization of layer IIIc pyramidal cells in the prefrontal cortex blossoms as an important connectomic fundamental plausibly related to the development of high order language processing and higher learning functions.
The genetic events related to the evolution of these basic connectomic fundamentals have been already discussed [4] yet without the precise timing and species attributions provided by the paleontological approach.The evidence that Neanderthal alleles on chromosomes 1 and 18 are associated with reduced endocranial globularity deserves further consideration.Indeed, it is claimed that these alleles modulate the expression of two nearby genes, UBR4 and PHLPP1.UBR4 encodes a ubiquitin ligase that regulates neurogenesis in the developing neocortex and, among other roles, promotes neuronal migration and PHLPP1 encodes a negative regulator of the PI3K/Akt growth-factor signaling pathway that drives myelination (14).Such "single structural genes" approach may indeed reveal the contribution of protein-coding genes in the hominization process.Yet, as discussed (4), such contribution may actually be part of a wider regulation of larger terns of genes under the control of a small number of genetic regulatory events.Accordingly, by analogy with genetic predispositions of ASD (39), "gene networks would encode neuronal networks" that determine behavioral phenotypes.This perspective advocates the heritability of complex traits of the connectome (and of many diseases) is spread broadly across the genome, which implies that a substantial fraction of all genes contributes to it (4).
Sill much remains to be explored to the benefit of this joint neurobiological-paleontological approach.In any case, the evolutionary trajectory followed by Homo sapiens over the last 300,000 years contrasts sharply with that followed by other largebrained Pleistocene Hominins.In our species, the improvement in cognitive abilities did not ultimately result from the continuous volumetric increase observed previously.Its evolutionary divergence represented a truly novel adaptive response to the everincreasing metabolic costs of a larger brain.