Monday, January 08, 2007

Brain size review   posted by amnestic @ 1/08/2007 06:34:00 PM

So Chris at DevIntel had a post a few days ago about a Barbara Finlay paper from 2001 that examined the ordering and relative length of phases of neurogenesis in mammals. I haven't got a copy of the paper, but it appears that she's arguing that the pattern of brain development is highly constrained and that changes through evolution will be quantitative rather than qualitative. This is evidence for a domain-general view of human evolution wherein first we get extra tissue and (i'm not sure of my usage here) we exapt it into human-specific brain parts.

I remembered a study from a while back suggesting that humans had a disproportionate amount of white matter in the prefrontal cortex. The author of that study (Schoenemann) has a thick 2006 review (pdf) covering many areas of human brain evolution available on his website. There are sections on cranial capacity, the fossil record, and behavioral evolution, but the bulk of the most interesting data are in the comparative internal allometry section and the genetic correlations between brain size and behavior. Neither are my specialty, but I'm really out of my league in the genetic correlations area, so I'll give you the brain size business and you can study the other part on your own.

Brain size scales with body size in a log-log relationship. You can determine the normal parameters of this relationship by sampling brain and body sizes across primates, for instance. You can get an idea of the relative increase in brain size controlling for body size by taking the ratio of observed brain size to that expected from this line of best fit. Compared to the rest of the primates, human brains are 3.1 times too big.

Brain regions evolve at different rates. For instance, in the boring case, the motor and somatosensory cortices could grow to accommodate larger bodies leaving other regions behind. We can compare the relative region sizes across primates and control for brain size, so we know how big the region should be "for a primate brain of our size". I like this measure more than the body size comparison because it seems to show how the brain has specialized. Thus, the human olfactory bulb, visual cortex, primary motor cortex, and premotor cortex are smaller than expected. As an aside, Michael Graziano is scrambling all of our distinctions about primary and pre- motor areas and moving away from the homunculus model that you see in your intro to neuroscience books. All these regions scale with body size reasonably, and we know we can see and move pretty well. Not only that, but the frontal lobe (which contains the motor cortices) scales properly while the motor parts lag. What is making our brains so big and raising the expectations for all these poor workhorse regions? The prefrontal cortex is what.

It turns out that measuring the prefrontal cortex (PFC) is messy because there aren't clear markers for the boundaries. Some MRI studies have used the area in front of the corpus callosum. This underestimates the size, but does so more for humans than for other primates. Using measures such as this, we find that the human PFC is larger than expected compared to brain and body size. There are some more detailed techniques outlined in the review. Three studies in 2005 morphed common chimp, bonobo, and macaque brains into human brains and came to the conclusion that human PFCs were expanded. Also, based on traditional cytoarchitectural markers of PFC areas, specific subregions within the PFC scale at different rates. He didn't mention any area that lags behind though, so it can't really help us focus yet. One interesting point that runs through the analysis is that you shouldn't write off a change in absolute size just because you can account for it using these scaling relationships. We get two examples (abstract rule learning and object-discrimination) that are better explained by absolute brain size than any relative measure.

In a semi-related note, Buzsaki suggested that a constraint on overall brain size may be axonal conduction delays. Certain cognitive feats may require synchronized oscillations in different cortical areas. While the networks can adjust to some out-of-phase inputs, there must be a certain degree of incoherence that they just can't handle. The idea is that Neandertal brains could possibly've gotten too big for their bridges. If this interests you, look into gamma oscillations and binding. I don't understand it yet well enough to explain it.

Update:Mindblog discussed a paper about long-range synchrony last week.