(Figure below modified from MacLean (1990)
Similarly, the density of nerve cells in the human neocortex is as predicted for a brain of our size. (HP., p.89).
Neurons are greedy cells. In resting humans, 18% of circulating oxygen is consumed by brain tissue even though brain makes up only 2% of the body mass (Kety & Schmidt 1948). In infancy the brain is proportionally much larger (Blinkov & Glezer 1968), a situation that is only partially offset by lower metabolic requirements (Kreisman et al 1986). Even during hibernation and estivation, brain metabolism is not reduced appreciably (Meyer & Morrison 1960). Thus neurons represent a high and fixed metabolic expense and the cost of feeding neurons and associated glial cells is undoubtedly one of the principal factors that constrains the total neuron population. (Footnote 3) The relative cost of neurons depends on an animal's size, its functional metabolic rate (Martin 1981, Armstrong 1983, Armstrong & Bergeron 1985), and the specific metabolic rate of its brain tissue (Krebs 1950). This cost is proportionally highest in small carnivores, birds, and primates, (e.g., Cebus albifrons ) in which the brain may represent up to 10% of the adult body weight (Spitzka 1903).
In contrast, large species may be able to afford a superabundance of neurons, even though they may have no pressing need for these cells. The metabolic demands of neurons may represent a minor addition to the total energy consumption of a whale. Alleles that contribute to the production of surplus neurons may be retained in the population because the behavioral and metabolic consequences of doing so are negligible. Although this extra baggage does have a small cost, the capacity to retain more neurons than needed increases the range of variation both in neuron number and in patterns of deployment in different individuals. Maximizing variation increases the evolutionary plasticity of the lineage, an idea analogous to the concept of genetic load-- the preservation of alleles in the population that reduce mean fitness but that compensate by providing a reservoir of heritable variation (Mayr 1963). In contrast, an equivalent fractional increase in total neuron number in a small animal will represent a greater increase in the metabolic load and may consequently be strongly selected against (Ricklefs & Marks 1984). Thus, in a mole, canary, or nematode in which the fraction of metabolism devoted to neurons is great, individuals with superfluous neurons will be weeded out quickly.
Neuron loss can be interpreted as contributing to an individual's fitness either by improving the efficiency of information processing or the efficiency of metabolism. At a higher level of analysis the combination ofoverproduction and loss of neurons may also buffer and conserve heritable variation and thereby increase the evolutionary plasticity of a species (Katz& Lasek 1978, Williams et al 1986, Finlay et al 1987).
[Taken from http://mickey.utmem.edu/personnel/PAPERS/NUMBER_REV_1988.html]
This is true for all human groups, regardless of contemporary diets. Hence it must derive from a common ancestor to contemporary humans.
This and other features of the human digestive system --including teeth--suggest our ancestors ate an unusually high quality diet compared with existing apes. Unlike the large tropical apes who spend their day foraging and eating, our human ancestors may have had to devote less time to the actual eating in order to obtain equivalent nutritional value. (Recall that gorillas and orangs may eat eight to ten hours per day.)
On the other hand, procuring food through hunting and gathering promotes a range of brain-mediated skills. Milton (1988) writes: "one of the strongest selctive pressures on early humans would have to develop the ability for co-operative behavior, delayed gratification, and the sharing of highly desirable and essential goods. p. 304"