Last week I wrote about the British neurobiologist Colin Blakemore’s assertion that human brain size suddenly increased by 30% about 200,000 years ago, and his idea that this increase was due to the fixation of a single macromutation. While I questioned this mechanism, John Hawks also questioned the data, claiming that brain size increase over time had been pretty steady, with no leap at the proposed time.
Well, an alert reader, anthropologist Ciarán Brewster from University College Cork, has done the proper statistical analysis of the brain-size data. Over at his website, Ad Hominin, Brewster shows that there is indeed an inhomogeneity in brain evolution, and that the pace of brain-size increase did indeed pick up about 200,000 years ago.
So Blakemore’s assertion about the acceleration was correct. But of course this says nothing about whether that acceleration reflected the fixation of a macromutation, which I consider very unlikely for reasons I’ve already mentioned. And Blakemore’s theory that any brain-size “macromutation” was initially neutral, and became advantageous only after it had already been fixed in our lineage by random processes (e.g., genetic drift), is pretty outlandish.
One thing that Brewster doesn’t discuss is whether that accelerated brain-size evolution was correlated with accelerated body-size evolution. If brains got bigger simply because bigger bodies produce bigger brains as a byproduct (this is what we geneticists call a “pleiotropic effect”), and if for some reason bodies were selected to get bigger around 200,000 years ago, then it’s not necessary to posit that there was something selecting for bigger brains per se. (Of course, increased body size could also be a byproduct of increased brain size!) I don’t know if this is likely given the data, but it’s something to consider.
How in hell does something neutral get fixed before an advantage is arrived at – this simply makes no sense.
If you have ever bought a novel (to you) tool, you know the wonder of how you ever got along without it. Thus, short any factual explanation of how it might occur, I cannot imagine having a batch of synapses that my parents didn’t have, just lying dormant.
“How in hell does something neutral get fixed before an advantage is arrived at – this simply makes no sense.”
It happens all the time, especially in small populations. It’s called genetic drift, the other dominant force of evolution (aside from natural selection).
I am here to learn. Comment just to get notified of follow-up comments.
I suggest to read the work of Bruce Lahn and Cris Walsh on human brain genes; Bernard Baars, global workspace and Olaf Sporns, (Dr Lindenfors work is also illuminating) of connectome (?) fame for a panoramic of brain cell architecture questions, constraints and required evolutionary steps leading, or not, to us (as compared to say, chimp brain size and architecture) Im sure Bruce and Dr Walsh are in good health, cant say about th others. Of course you can do something else with my sugestions.
This is a lot of reading into a figure with sparse data, which was why I was reluctant to go further than to say there’s no sudden jump. I thought it obvious from the chart, but Jerry seems intent on finding some inflection point at 200,000 years ago.
Of course one way to answer that question is to examine it directly. Wolpoff and Lee (2003) tested explicitly the hypothesis of a single geometric increase (which they viewed as the null), compared to higher-parameter models with two distinct stages of size increase. They found that no two-stage model gave a significantly better fit than the single geometric fit. That doesn’t prove that the rate didn’t increase, but as others have pointed out in this example, perhaps a log linear fit should be examined before we discuss a change in the *rate* of growth.
Now, as to the question of what happens at or around 200,000 years, I think we need to consider two important pieces of information that *aren’t* in the simple chart:
1. More than 2/3 of the post-200k specimens are male.
2. *Present* human populations have endocranial volumes between 1200 and 1500 ml.
Both these suggest (although do not prove) that the apparent trend in the later part of the graph is an overestimate of the actual change in endocranial volume.
I’m extremely surprised that Jerry writes that “Blakemore’s assertion about the acceleration is correct.” How can that be? He posited a “sudden jump” in brain size.
There’s no sudden jump. Period.
What’s the deal?
I’m not “intent” on proving anything; I just referred to a new analysis suggesting inhomogeneity in the evolutionary rate. I have no investment in the results one way or another; why would I?
Yes, I agree that there was no instantaneous “leap” at 200,000 years ago, but there did seem to be an acceleration. That’s all I meant. Of course I have no truck with the macromutation idea.
I’m not “intent” on proving anything; I just referred to a new analysis suggesting inhomogeneity in the evolutionary rate. I have no investment in the results one way or another; why would I?
I was wondering about that, and glad to hear that my impression was wrong, because it seemed bizarre.
I think Ciaran did a nice job of slicing the data, but tested the wrong fit — a linear model of increase rather than geometric. The hypothesis of a constant rate entails a geometric increase.
The thing about Blakemore is, he’s asking precisely the opposite question from the interesting one. It’s not, “Why did brain size suddenly start increasing faster?”
Fast evolution is exactly what you’d expect from directional selection.
The interesting question is, “How did brain size change so slowly for such a long time?”
Whether a constant rate entails a linear increase depends on what one means by “rate”. If you mean rate in terms of cc/year, that predicts a linear increase by time. If you mean rate in terms of, say, x per cent increase per year, then, yes, you expect a geometric increase.
Just to pipe in here, from a statisticians viewpoint …
The data, as is, do support the possibility of a two phase segmented model, as postulated by Ad Hominim. What’s more, the model can estimate the phase shift time (join point), which I find at about 252 Kya. That said, however, there is little statistical advantage in this segmented model (lack of parsimony, for one) over a non-linear power model (my preference) or the model proposed by Lee and Wolpoff. Their bootstrap method of assessing a phase change was original, if nothing else, but questionable and convoluted, IMO. All these show unquestionably better fits than a straight out linear model.
As has been pointed out, while the data may support this increasing rate pattern, it does not say that it is real (eliminating a few key data points could easily change the pattern). Whether it is supported theoretically is definitely more in your professional venue than mine 🙂
My results can be found at:
http://pdiff.com/Brain/brain.html
Pdiff
Your axis labeling is much nicer, thanks!
Just a couple of observations from a dummy:
1) It looks like a line can be set against the largest capacities that would eliminate (except for the two highest points at 50k years) the abrupt increase. True, its not the average but, the data obviously isn’t complete either.
2) Blakemore states:
“The question is: why is [our brain] so big compared to the brains of our predecessors, such as Homo erectus? Until 200,000 years ago, there had been a gradual increase in brain size among hominins, starting three million years ago. Then, abruptly, there was a remarkable increase of about 30% or so.”
However, as Jerry noted it looks (from the data) more like an increase in the rate of change. Blakemore speculates that it was a one time event and if that is the case, what appears (from the data) to have been passed on was a rate of change not an abrupt one time increase in capacity.
3) The completeness of the data is obviously lacking. Use your eye to find the midpoint of each of the time sets – see how abruptly the midpoints cross the average trend line? That makes me think that the data to support the average isn’t really complete but, from an abrupt change point of view, why is Blakemore ignoring all the other “one time changes” both increases and decreases?
4) If the rate of change line is extended it looks like an average capacity of 1600cc would be expected of our present brain cases. However, as measured the average is somewhere around 1400cc. So our current average is significantly less than the average (according to the data)
brain capacity of 50k years ago. Where did that supposed “abrupt increase” go?
5) A “good” christian might have faith that their species abruptly began 6,000 years ago. If so, that might be a christianized explanation for our apparent dwindling intellectual capacity since they arrived.
6) If the straight line is extended (without the increase) to the present it looks like it would give about 1300cc for the current capacity, which is closer to what is reported to be the case than the 30% increase would be.
Like the OP, I too wonder as to the why of the event. Was there a mutation that increased the size of the human brain? Was it connected to an increase in overall body size? Did a larger body precede a larger brain? Was the increase in brain size proportionate to the increase in body size, or out of proportion to it? Where did the brain size increase take place originally, and where there any changes in the environment that may have given some advantage to the increase in cognitive ability?
Based on what I know, I’d have to say that our new, larger brains, even if accompanied by larger bodies, were disproportionate compared with equivalent changes in other species. That is, our brains grew larger compared to our bodies. With this larger size came increased cognitive ability and increased capability.
There is one thing we need to remember, this increase in brain size was not universal across hominin species. Erectus for doesn’t show the sudden increase, but a more gradual enlargement between 200 thousand years ago and their subsequent extinction. The same, I believe, holds true for neanderthals. The hobbit on the other hand shows a brain proportionate to his size, and his ancestors likely did as well.
Still more to learn here.
The drift idea is indeed an odd thing to propose in concert with the “instantaneous” appearance idea. How long would a drifting allele take to fix in the human population? On the order of 100,000 years. That the fixation time is so long would speak against “instantaneous” appearance.
Could neoteny have played a role in this? Birth at an earlier stage of development would have made it physiologically possible for further brain growth outside the womb. A reduced pregnancy period (if that indeed was the case) could have had survival advantages for the mother, primary to any immediate advantage for the bigger brain.
I’m not a biologist, and that is an honest question. I have no idea if it’s an utter absurdity, a truism, or a vague possibility. Am I totally on the wrong track with that line of thinking?
No, it’s not neoteny; more likely modular hypermorphosis (which may have had pleitropic effects on other systems). Maybe due to deconstraint on brain energetic/nutrient allocation allowed by more meat in the diet (more fat for brain, less metabolism required by digestion of plants) thanks to improved hunting techniques.
The fossil data is so choppy that 200 k is probably a modeling artifact. About half a million BP, including all archaic and modern hs, is a better pivot.
To get from a chimp-like LCA brain to a modern human brain in terms of intra-brain structural relationships is mostly just allometric scaling. Differences in shape between archaics, neandertals and amhs brains are probably not terribly relevant. Mainly changes in cranial dome shape associated with facial reduction.
The really big change, of course, was from australopiths/habilines to erectus. May be a handful of mutations of large effect there.
And after 200,000 years not a hell of a lot of lot happened in terms of brain size and organization (sure, some size reduction in the last tens of thousands). (Never mind the ’10 K Explosion’ silliness.)
Thanks, Colugo. I’ve got some reading to do!
Thanks for the mention, Jerry! I think that much of the differences in brain size are indeed related to allometric scaling going on in the rest of the body. The very heavily built Homo heidelbergensis has a brain volume of around 1200 cc (compared to around 900 cc for Homo erectus) while the equally well built Neandertals had a brain volume of around 1500 cc. The modern human mean is around 1350 cc. I think these disparities can be accounted for by body size differences. For me, a more interesting question is why our species in particular became smaller and more gracile.
Hey,
I’ve been playing around with this data a bit as well, and Ad Hominin’s post inspired me to do a proper analysis.
In short: the increase in rate at ~250 kya stays even if you fit an exponential model, and the relationship stays even if you consider sexes separately (but my sex-classification is distinctly amateur, so if anyone wants to correct it I would be joyous).
Anyway, my results are written up as a blog post here:
http://www.genetic-inference.co.uk/blog/?p=871
I use the same approach as Pdiff, but generalising beyond just linear modeling, and including sex.
I did an analysis using local linear regression, which shows a clear “kink” in the smooth of log-brain-size at 300 thousand years ago.
Sorry, forgot the link!
http://ecstathy.blogspot.com/2010/04/human-brain-evolution-clear-kink.html
hi
as far as i know the introduction of meat in diet led to a decrease in the digestive system size,whose energy flew to the brain,which increased 30% in a short time.Cooked meat is easier to digest than plants