evolution of the brain in primates

The earliest placental mammals were small, nocturnal, insectivorous animals that lived in the Cretaceous period more than 100 million years ago (Fig. 1). They possessed acute senses of smell and hearing and long, sensitive snouts bearing vibrissae. Their eyes were small, laterally directed, and possessed very limited acuity. Their brains were somewhat larger than those of similarly sized reptiles, but the neocortex, which was to become the great focus of mammalian brain evolution, had undergone only very limited development (Fig. 2). The telencephalon was largely devoted to the olfactory bulbs and olfactory cortex. The early placental mammals lived in a landscape dominated by reptiles, but this was to change radically with the massive extinctions that devastated the ranks of the reptiles at the end of the Cretaceous period. A host of theories has been advanced to explain the sudden extinctions. Perhaps the most compelling, because of its strong support from geophysical data, is the theory advanced by L. W. Alvarez and his collaborators that a large asteroid struck the earth, resulting in the release of enormous quantities of dust into the atmosphere and thus a great reduction in sunlight reaching the earth's surface. The nocturnal, warm-blooded mammals would have been much better equipped for survival during a period of sudden climatic cooling than were their reptilian contemporaries. Whatever the cause, 65 million years ago at the beginning of the Cainozoic era, the 'age of mammals', a large number of ecological niches lay vacant that formerly had been occupied by reptiles.


Fig. 1. Skull and restored head of the Cretaceous placental mammal Zalambdalestes lechei.



Fig. 2. Dorsal and lateral views of the brain of the European hedgehog (Erinaceus europaeus), a modern insectivore that has retained many features characteristic of the primitive placental mammals. The neocortex lies dorsal and medial to the rhinal sulcus. V-I, first visual area; V-II, second visual area. The circles indicate the representation of the vertical meridian (mid-line) of the visual field; the triangles indicate the representation of the far periphery of the contralateral half of the visual field. Note the large olfactory bulbs (OB). The cortex ventral to the rhinal sulcus is largely devoted to the processing of olfactory input.
During the next 10 million years the basal stock of placental mammals began to differentiate into the various orders of mammals. The fossil remains of the earliest primates that bear a close resemblance to living primates have been recovered from early Eocene deposits approximately 55 million years old. The skull and cranial endocast of one of these early primates, Tetonius homunculus, is illustrated in Fig. 3. Tetonius possessed large bony orbits that completely encircled its eyes, and a cranium containing a large brain compared with its similarly sized contemporaries. The relative size and position of the orbits in Tetonius closely resemble living nocturnal prosimians such as Galago (Fig. 4). As can be seen in Fig. 5, Galago possesses large, frontally directed eyes with virtually as much binocular overlap, in the order of 120–140 degrees, as is present in monkeys, apes, and man. The great similarity in the size and position of the orbits in Galago and Tetonius suggests that Tetonius also possessed large, frontally directed eyes and was crepuscular or nocturnal. The cranial endocasts of Tetonius and other Eocene primates show that their brains possessed a conspicuous enlargement of the neocortex in the occipital and temporal lobes, which in modern primates are known to be devoted mainly to the cortical processing of visual information (Fig. 4). Thus, by the early Eocene, there appeared in primates the concomitant development of large, frontally directed eyes with improved acuity and an expanded visual cortex. The sensitive snout of the primitive placental mammals was greatly reduced in the early primates. The functions of the snout as a tactile probe and an apprehender of insect prey were taken over by the hands. The olfactory system in the early primates retained a comparable degree of development to that present in the early placental mammals. The size of the olfactory bulbs did not diminish relative to body size, but the occipital and temporal neocortex expanded so greatly that the olfactory bulbs became small by comparison (see Figs. 2, 3, and 4).


Fig. 3. Left. Dorsal view of the skull of Tetonius homunculus. (A.M.N.H. No. 4194.) Right. Dorsal view of L. B. Radinsky's cranial endocast of Tetonius. OB: olfactory bulbs. S: sylvian sulcus.



Fig. 4. Left. Dorsal view of the skull of Galago senegalensis. Right. Dorsal view of the brain of Galago senegalensis. The visual cortex corresponds to approximately the posterior half of the neocortex. The Vs demarcate the anterior border of visual cortex. OB: olfactory bulbs. S: sylvian sulcus.



Fig. 5. Close-up of the face of Galago senegalensis. Note the mid-line cleft in the upper lip, which is a feature present in all strepsirhine primates but absent in haplorhine primates.
These developments must reflect fundamental changes in ecological specialization that occurred in the evolutionary progression from the earliest placental mammals to the early primates. Two theories have recently been advanced to explain the basic adaptations that served to differentiate primates from the early placental mammals. In the first, R. D. Martin has suggested that the early primates, like the smaller living prosimians they closely resemble and the small arboreal marsupials of Australia and South America, adapted to a 'fine branch niche'; the prehensile hands and feet found in these animals developed to grasp the fine terminal branches of trees. Most arboreal mammals, such as squirrels, run on the trunk and larger branches but are unable to grasp the finer branches. The second theory is based on the observation that, outside the order Primates, animals with large frontally directed eyes (owls and felids) are nocturnal, visually directed predators, which has led M. Cartmill to propose that visually directed predation was the ecological specialization responsible for the developments in the early primates. Cartmill's visual predation hypothesis is supported by the fact that the tarsier, the living primate that most resembles the early primates of the Eocene, is exclusively a predator (Fig. 6). It appears that both hypotheses have considerable merit. It seems probable that the early primates did invade the 'fine branch niche', where they gained access to a rich array of insect and small vertebrate prey.

What are the advantages that frontally directed eyes afford nocturnal, visually directed predators? Predators generally orient so that the prey is located in front of them so that they can propel themselves forward rapidly and carry out a coordinated attack with forelimbs and jaws, and it is likely that frontally directed eyes provide maximal retinal image quality, for the central part of the visual field where the prey is located, in the crucial moments before the final strike is made. Image distortion tends to increase the further an object is located off the optical axis of the lens system, and thus it is advantageous to a visually directed predator to have frontally directed eyes in which the optical axes are directed toward the central part of the visual field, so that the predator can utilize the maximum-quality retinal image in the crucial moments before the strike, when it is evaluating the prey's movements, the prey's suitability as food, and the prey's ability to defend itself. The dimly illuminated nocturnal environment makes these optical factors particularly important, and rules out other mechanisms for improving retinal image quality such as stopping down lens aperture. Frontally directed eyes also provide a large binocular field over which binocular summation can be achieved, which may be of particular value in conditions of low illumination. The binocular input is also used to reconstruct a stereoscopic view of a large portion of the visual field. The advantages of stereoscopy to a predator are that it provides information about the distance of prey, and, as Bela Julesz has pointed out, it helps the predator to discriminate camouflaged prey from background.


Fig. 6. Tarsier seizing a lizard. The tarsier is exclusively a predator; it eats no fruit or vegetable matter. Of the living primates, the tarsier most closely resembles the early primates living in the Eocene
Almost all of the distinctive features of the primate visual system relate to the frontal direction of the eyes and binocular integration and, since these features are present in all primates, it is likely that they developed in the early primates. These features include:

(i) a high concentration of retinal ganglion cells in the central retina, and the greatly expanded representation of the central retina in the neural maps of the visual field in the brain;
(ii) the representation restricted to the contralateral half of the visual field in each side of the optic tectum, differing from complete representation of the field of view of the contralateral retina found in each side of the optic tectum in all other vertebrates that have been investigated (Fig. 7);
(iii) a relatively large retinotectal projection, which, together with the unique visuotopic organization found in the primate tectum, suggests that the optic tectum in the early primates developed capacities related to the integration of binocular input;
(iv) a distinctly laminated dorsolateral geniculate nucleus in which inputs from the two retinae are brought into precise visuotopic register before being relayed to the primary visual cortex (VI); and
(v) a greatly expanded visual cortex containing a number of neural maps of the visual field (see localization of brain function and cortical maps).




Fig. 7. Schematic plan of the representation of the visual field in the optic tectum of primates and non-primates. Circles indicate the vertical meridian (mid-line) dividing the two halves of the visual field; squares, the horizontal mid-line; triangles, the extreme periphery of the visual field; stars, the division between binocular and monocular portions of the visual field.
Why does the visual cortex contain a series of separate representations of the visual field rather than a single map? In attempting to develop computer analogues of visual perception, David Marr elaborated the principle of modular design. Marr stated that any large computation should be broken into a collection of smaller modules as independent as possible from one another, otherwise, 'the process as a whole becomes extremely difficult to debug or improve, whether by a human designer or in the course of natural evolution, because a small change to improve one part has to be accompanied by many simultaneous changes elsewhere'. This modular principle has many counterparts in other biological systems. The palaeontologist W. K. Gregory noted that a common mechanism of evolution is the replication of body parts due to genetic mutation in a single generation which is then followed in subsequent generations by the gradual divergence of structure and functions of the duplicated parts. An analogous idea has been advanced by a number of geneticists. They have theorized that replicated genes escape the pressures of natural selection operating on the original gene, and thereby can accumulate mutations which enable the new gene, through changes in its DNA sequence, to encode for a novel protein capable of assuming new functions. Many clear-cut examples of gene replication have been discovered, and DNA sequence homologies in replicated genes have recently been established. Using this analogy, the author and J. H. Kaas have proposed that the replication of cortical sensory representations has provided the structures upon which new information-processing capabilities have developed in the course of evolution. Specifically, it has been argued that existing cortical areas, like genes, can undergo only limited changes and still perform the functions necessary for the animal's survival, but if a mutation occurs that results in the replication of a cortical area, then in subsequent generations the new area can eventually assume new functions through the mechanisms of natural selection while the original area continues to perform its functions.

The primary visual cortex (VI) and the adjacent second visual area (VII) are present in all mammalian species that have been investigated. Thus, VI and VII were probably present in the early placental mammals that were the common ancestors of the various living mammalian orders. The full complement of cortical visual areas varies in different mammals from a minimum of two in a basal insectivore, the hedgehog (see Fig. 2), to a maximum of twelve found in cats by Palmer, Tusa, and Rosenquist. At least nine cortical visual areas are present in the owl monkey, the primate that has been mapped most completely (Fig. 8). Beyond VI and VII it is very difficult to establish homologies among the visual areas present in mammals belonging to different orders. The last common ancestor of the different mammalian orders lived no more recently than the late Cretaceous period more than 65 million years ago, and this ancestral mammal had only a very limited development of its neocortex. In addition, the adaptive radiation of mammals into different ecological niches with widely divergent behavioural specializations serves to make very difficult the discovery of diagnostic similarities among potentially homologous cortical areas in different mammalian orders.

Within the order Primates, it is easier to determine homologous areas in different species. The highly distinctive middle temporal visual area (MT) is present in prosimians and in both New and Old World monkeys and thus probably existed in the early primates. MT neurons are selective for the direction of movement of visual stimuli and often are particularly responsive to moving fields of visual texture. MT may have developed as a specialized mechanism for detecting and tracking moving prey. MT may also participate in visuomotor coordination by analysing the visual flow patterns that the animal sees as it moves through its environment; MT projects via the pontine nuclei to the cerebellum, a major centre for the control of body and eye movements. The evolutionary development of this system may be related to the special demands of locomotion in the 'fine branch niche'.

The dorsolateral visual area (DL) lies adjacent to MT and also appears to be part of the basic complement of visual areas common to all living primates and thus probably existing in the early primates. In DL about 70 per cent of the neurons are selective for the spatial dimensions (length and width) of visual stimuli within excitatory receptive fields that generally are much larger than the preferred stimulus dimensions. The dimensional selectivity of DL neurons is independent of the sign of contrast in the receptive field, as they are equally selective to both light-on-dark and dark-on-light stimuli, the amount of contrast, and the position of the stimulus within the excitatory receptive field. It suggests that DL contributes to form perception. This hypothesis is consistent with the fact that DL, of all the visual areas, has the most expanded representation of the central visual field where the most acute recognition of form takes place, and with the discovery (by Weller and Kaas) that DL is the main source of input to the inferotemporal cortex (IT) has been strongly implicated in the analysis of complex visual stimuli and the learning of visual form discriminations.

The small amount of neocortex possessed by the early placental mammals and the great variation in the number of corticovisual areas reported for different mammalian species suggest that some of the areas beyond VI and VII developed at different stages in evolution and independently in different lines of descent. One clear example of variation within the order Primates is among the areas immediately anterior to VII. In the prosimian Galago, only one map of the visual field, the dorsal area (D), is located in the position occupied by three separate maps, M, DM, and DI, in the owl monkey (see Fig. 8). The existing data suggests that there exists a core of areas including VI, VII, MT, and DL and possibly one or two others that are common to all primates, but that there also exist areas present in some species but not in others. It is probable that each area performs a distinct set of functions in visual perception and visuomotor coordination, and that an area possessed by one species (or larger taxon) and not by another will endow its possessor with behavioural capacities not present in the other. A major task for the future will be to determine what are the distinctive functions of these cortical areas and how they relate to the behavioural and ecological specializations of their possessors.

The early primates probably were small nocturnal predators living in the fine branches; some primates have retained this mode of life, but most have become larger, diurnal folivores or frugivores. Frugivorous diet is correlated positively with brain size and the amount of neocortex relative to body size in primates. This association between frugivorous diet and enlarged brain and neocortex may be related to the special demands imposed because a fruit eater's food supply is not constant, since different plants bear fruit at different times and at different locations in the complex matrix of the tropical forest. It is clear that an animal guided by memory of the locations of fruit-bearing trees can more efficiently exploit the available fruit resources than would be possible otherwise. Thus natural selection would have favoured the development in frugivorous primates of capacities for visuospatial memory, which may be localized in a particular area or set of areas.


Fig. 8. The representations of the sensory domains in the cerebral cortex of the owl monkey. Above is a ventromedial view of the right hemisphere, below is a dorsolateral view of both hemispheres. On the left is a perimeter chart of the visual field. The symbols in this chart are superimposed on the surface of the visual cortex. Pluses indicate upper quadrant representations; minuses, lower quadrants. The row of Vs indicates the approximate border of visually responsive cortex. AI, first auditory area; AL, anterolateral auditory area; CC, corpus callosum; DI, dorsointermediate visual area; DL, dorsolateral crescent visual area; DM, dorsomedial visual area; IT, inferotemporal cortex; M, medial visual area; MT, middle temporal visual area; ON, optic nerve; OT, optic tectum; PL, posterolateral auditory area; PP, posterior parietal cortex; R, rostral auditory area; VI, first visual area; VII, second visual area; VA, ventral anterior visual area; VP, ventral posterior visual area; X, optic chiasm. The main connections of the visual areas are as follows. VI projects to VII and MT. VII projects to DL, which in turn projects to IT. MT projects to all visual areas except VII and IT. MT and PP project to the frontal eye fields.
Another, even more significant behavioural specialization is the development of complex systems of social organization in many primate species. The neural substrate for the mediation of social communication is bound to be an important focus of evolutionary change in the brains of primates. The order Primates is divided into the strepsirhines (lorises, lemurs, galagos), which tend to have relatively simple forms of social organization, and the haplorhines (tarsiers, monkeys, apes, and humans), in which social organization tends to be much more complex. In strepsirhines, as in most mammals, the rhinarium, the space between the upper lip and the nostrils, is furless, moist mucosal tissue that is tightly bound to the underlying maxillary bone and is divided along the midline by a deep cleft (see Fig. 5). Since strepsirhines share this type of rhinarium with most other mammals, it is very likely to have been the primitive condition in primates. By contrast, haplorhines possess a furry rhinarium and a mobile upper lip that is capable of participating in facial expression. Strepsirhines, like most primitive mammals, have scent glands and scent-marking behaviours that play a very important role in their social communication, and, while haplorhines also use olfactory cues to some extent, they rely much more heavily on the use of visually perceived facial expressions and gestures, which allow much more rapid and subtle communication. Strepsirhines also tend to have much larger olfactory bulbs than do haplorhines. Thus it appears that, as complex systems of social organization evolved in haplorhine primates, social communication was increasingly mediated by the visual channel at the expense of the olfactory. One expression of this evolutionary development is the sensory input to the amygdala, which controls the neuroendocrine functions of the hypothalamus and thus emotion. Primitively, the main input to the amygdala was from the olfactory bulb, but in haplorhines the main input is from the temporal lobe and particularly from the inferotemporal cortex, which is a high-level processor of visual information. Neurons responsive to the specific configurations of faces have been recorded in the amygdala and temporal cortex. The clinical condition prosopagnosia, the inability to recognize familiar faces with relatively little impairment of other visual functions, which is usually associated with lesions located near the occipital temporal junction (see split-brain and the mind), suggests the development of a specialized system for processing the information in faces. Finally, in man, another system of social communication, language, has developed along with specialized cortical regions in the temporal and frontal lobes (see language areas in the brain).

(Published 1987)

— John M. Allman

Bibliography
• Allman, J. M. (1977). 'Evolution of the visual system in the early primates'. In Sprague, J. M., and Epstein, A. M. (eds.), Progress in Psychobiology and Physiological Psychology, vol. vii.
• — —  (1998). Evolving Brains.
• Cartmill, M. (1974). 'Rethinking primate origins'. Science, 184.
• Clutton-Brock, T. H., and Harvey, P. H. (1980). 'Primates, brains and ecology'. Journal of Zoology, 190.
• Gregory, W. K. (1951). Evolution Emerging.
• Jerison, H. J. (1973). Evolution of the Brain and Intelligence.
• Julesz, B. (1971). Foundations of Cyclopean Perception.
• Martin, R. D. (1979). 'Phylogenetic aspects of prosimian behavior'. In Doyle, G. A., and Martin, R. D. (eds.), The Study of Prosimian Behavior.
• Penfield, W., and Roberts, L. A. (1959). Speech and Brain Mechanisms.
• Polyak, S. (1957). Vertebrate Visual System.
• Radinsky, L. (1979). The Fossil Record of Primate Brain Evolution.

---