This PLoS Biology review, "Molecular insights into human brain evolution," is from 2005, but it's well worth reading. Here's the abstract:
As a species, we pride ourselves on the uniqueness of our brain. Relative to our body size, the human brain is bigger than that of any other animal. It may also contain unique structures and patterns of organisation that presumably underlie our intelligence and ability to manipulate our environment. But how did our unique brain originate, and under what selective pressures did it evolve? Some of the answers may lie in the genetic differences that researchers are now uncovering between us and our closest relatives.
I'm mainly blogging this so that I won't forget this section:
For natural selection to work, the costs of brain evolution must be outweighed by the advantages gained in terms of fitness. For many years, explains ecological psychologist Robin Dunbar (University of Liverpool, United Kingdom), "people thought that the ability to hunt or forage better was what drove the evolution of our brains. But a better diet had to come before we could grow a bigger brain." Dunbar believes instead that brain evolution in primates and more generally in mammals "has been driven by the need to manage social relationships, and in primates, in particular, to coordinate coherence in social groups through time and space". More complex social interactions, he says, mean that individuals are better able to pool resources to solve problems like finding food, and so they survive better.
This theory, says Dunbar, is supported by a correlation between social group size and neocortex size across primates and modern humans. Furthermore, during primate brain evolution, the trend has been to add more material to the front than the back of the brain. The front of the brain is where information from the rest of the brain is interpreted, and the capacity to interpret information underlies social interactions, says Dunbar. The number of problem-solving cognitive tasks you can do may well depend on how much frontal lobe volume you have and how it is organised. Just think of how few moves you can run a chess game into the future with a 1980s personal computer compared to a 21st century mainframe machine, he suggests.
I get so irritated by this idea that "a better diet had to come before we could grow a bigger brain". The published research generally doesn't include this little ditty, including Dunbar's work. I've heard it a lot, though, usually in the form of "the evolution of larger brains was constrained by low diet quality in such-and-such species."
Especially applied to the Australopithecus-Homo transition. Supposedly, australopithecines ate such low-quality diets that they couldn't grow large brains. Never mind that australopithecines somehow managed to have larger brains than any other primate besides humans. It's total nonsense. By the same logic, if hunters leave piles of sugar beets out for the deer for a couple thousand years, then deer should evolve bigger brains.
The fallacy here is a confusion about fitness versus selection. It is certainly true that natural selection increases a population's mean fitness. But like any other phenotype, fitness may be increased by changes in the environment. A constantly replenished pile of sugar beets increases the fitness of the local deer. They leave more offspring, and their population grows.
But that is just what fitness is: an increase in the rate of intrinsic population growth. Under the sugar beet diet, large, expensive-brained deer survive better. But so do small-brained, stupid deer. No correlation between brains and sugar beets means no selection. Hence, a never-ending supply of sugar beets will certainly get you more deer (which is the hunters' goal) but not smarter deer.
Add any significant amount of meat to the diet of an early hominid, and you will increase its fitness. Higher fitness means faster intrinsic growth rate, means more genes of the meat-eaters in the next generation. Nice work if you can get it. But the trick is getting the meat in the first place.
You can be sure the lions weren't leaving never-ending meat piles to increase the local hominid population. No, the hominids just had to pull themselves up by their bootstraps and find a way to hunt, scavenge, or steal it.
Claims about meat eating in human evolution are claims not about fitness, but about selection. Brains are expensive. They don't just expand for no reason: there must be some increase in the fitness value of large brains to outweigh their energetic and developmental disadvantages. Over the short run, these disadvantages are slight, but they are substantial enough to constrain brain size across evolutionary time spans. If hominids increased their meat intake, some cleverness allowed them to do it.
Does diet constrain brain size? Not by itself. A species that kept the same diet might have an increase in brain size if larger brains increased fertility. This is a plausible mechanism for selection based on social interactions -- if an individual's chance of mating was influenced by his cleverness. But using this as an explanation for long-term increases in brain size in a lineage just doesn't work for me. It supposes a feedback selection on intelligence that has little justification. If such feedbacks occur, I'd like an explanation for why brain size doesn't increase in most lineages.
Anyway, more on this later.
Bradbury J. 2005. Molecular insights into human brain evolution. PLoS Biol 3:e50. doi:10.1371/journal.pbio.0030050