Correlating Brains, Bodies and Behaviors

It seem reasonable that evolution might select for adaptive behaviors by increasing the relative size of particular brain regions that support those behaviors; for example, bats might have an enlarged auditory cortex since they navigate with echolocation. To some extent this does happen, but such differences are often apparent only after controlling for a much larger source of variance: changes in brain size that correlate with changes in body size - and the implications of this fact are wide-reaching.

As Barbara Finlay and coauthors wrote in this 2001 Behavioral and Brain Sciences article, the order of neurogenesis and ultimate size of various brain regions is highly conserved across all mammals, suggesting natural selection largely works not by selecting for specific "modules" or components in the nervous system as appropriate to a particular organism, but instead "by adjusting the parameters of a 'standard' developmental program."

Finlay et al. first review the history of comparative brain research, citing the well-established fact that brain size scales exponentially with body size. The brain-to-body-mass ratio (also known as the Encephalization Quotient, or EQ) shows some interesting correlations; for example, predators show larger EQs than their prey. Individual brain regions each have a characteristic size "scaling factor" with respect to the size of the rest of the brain; neocortex has the highest scaling factor (also known as "cortical hyperallometry") and the medulla one of the smallest. These relationships appear to explain 95% of the variance in sizes of particular brain regions.

It is thus not surprising that various brain structures tend not to show much variance in size that is related to an animal's behavior. Finlay et al report that most factor analyses reveal two major sources of variance: a general brain size factor and an "olfactory bulb" factor (which is perhaps the only brain region to reliably show behavior-appropriate size variance). On the other hand, it is possible that strong relationships between brain region size and behavior do exist, but are statistically hard to detect.

To investigate this possibility, Finlay et al. used more specific criteria for parcelling up various brain regions (for example, by separating motor, visual and auditory structures). They used measures of 11 brain regions (including cerebellum, hippocampus, neocortex, striatum, olfactory bulb and others) in over 130 mammals, falling into four different taxonomonic groups: 40 insectivores, 43 bats, 21 prosimians, and 27 simians. For every region except for olfactory bulb and medulla, region size was best predicted by a model that included the size of each other region, as compared to models which measures body size, taxonomic group, and all of their interactions.

Finlay et al. also analyzed the duration of neurogenesis and the order of development of each of 51 regions in four species, and discovered that structures with prolonged developmental timelines were the same that showed a disproportionately large scaling factor with respect to other brain regions. Thus not only brain region size, but also developmental timing, appears highly conserved across species.

Although one would expect evolution to select for brain regions that are most directly responsible for adaptive behaviors, it appears that selection operates only at the most gross level in terms of brain size. Later in the article, Finlay et al even argue for "egg before chicken": they suggest that "additional structure preceded enhanced function in hominid brains." Specifically, the authors view neocortex as a "spandrel" which may have evolved merely as a side-effect of other neural adaptations, and was then exapted for its current use.

Stretching a bit, the authors argue further that this could even explain the tens-of-thousands of years that separate the "appearance of anatomically modern people some 100,000 years ago and the advent of unequivocally modern behavior." Sidestepping the somewhat thorny issue of how to define "modern," this work by Finlay et al. goes a long ways towards, this account clearly falls under the "domain-general" perspective of the evolutionary development of higher cognition, including working memory and executive function. How this perspective relates to those advocated by Coolidge & Wynn, for example, will be the subject of future posts.