What Happen When Our Brain Gets Bigger?

Not only are the ‘laws of brain-size evolution’ consistent with massive modularity,but they also give the latter some independent support. Striedter (2004)reviews a wide range of comparative and evolutionary data, and is able to extract a number of robust generalizations about what happens to brains when their absolute size increases, within a certain lineage (see also Geary, 2005). One is that larger brains become significantly more modular in their organization. This is because, as the total number of neurons increases, the density of dendrite connectivity amongst neurons significantly decreases. (This is, no doubt, partly because of the energetic costs associated with building and maintaining neural connections—Aiello and Wheeler, 1995; and partly because of the constraints imposed by slow speeds of signal propagation within and between neurons.) The result is that neurons tend to maintain more of their local connections, while giving up a greater proportion of their long-distance ones, resulting in an architecture that appears significantly more modular.

Another (related) generalization described by Striedter (2004) is that an increase in the size in a given brain region tends to be accompanied by increases in functional differentiation amongst its sub-regions. And this generalization holds good, not only for local brain regions, but also for the brain as a whole.

In particular, the lateralization of brain functions increases as the overall size of the brain goes up, no doubt because of the costs of maintaining neural connections between the two hemispheres. What we should predict, then, as a result of the fourfold increase in brain size that occurred through the evolution of hominids, is that the brains (and presumably the minds) of humans became much more modular—containing many more functionally distinct processing systems—than the brains of our great-ape ancestors. Since I have previously argued the minds of apes and other non-human animals are already massively modular in organization, this is likely to mean that the transition to Homo sapiens will have seen the addition of many new mind / brain adaptations.

Should the resulting systems count as adaptations, however? Aren’t they rather side-effects of some general Bauplan governing increases in brain size? Initially, perhaps, some of them may have been. But exaptation is actually a form of adaptation. Where systems are shaped and maintained by natural selection they can count as adaptations, even if the processes through which they initially arose weren’t targeted on them in particular. (The penguin’s ‘wings’ are an adaptation for swimming, although they are exapted from limbs once designed for flight.) And such shaping and maintaining is likely to apply to most of the brain systems underlying the human mind, since cognition is just too important for survival and reproduction for it to be otherwise.

There is one other ‘law of brain evolution’ described by Striedter (2004) that is worth mentioning here. This is that as the size of one brain region increases relative to another, the extent of the neural projections from the larger to the smaller also increases. The result is that the relatively enlarged human neocortex has culminated in unusually extensive projections from the neocortex to the motor neurons in the medulla and spinal chord. This may give us a partial explanation of our species’ impressive abilities for fine-grained motor control, not only involving movements of the hands, but also the lips, tongue, and respiratory system necessary for the control of speech.

Neural Proliferation and Pruning

I have replied to the argument that the ‘laws of brain evolution’ rule out massive modularity. Buller (2005) claims that massive mental modularity is inconsistent with what is known about the processes of brain development, however. He suggests that we can, on the contrary, see just a single novel adaptation underlying human uniqueness. This is the process of neural overproliferation, and subsequent pruning in response to experience, that sculpts our brain organization during the course of development. Let me review some of the pertinent facts, following Webb et al. (2001).

Almost all of the neurons that the brain will ever contain are produced before birth, during the first seven months of development. These continue to grow connections and build synapses with one another for the first two years after birth, with the maximum total number of brain synapses occurring at around two years of age. Thereafter there are two major periods of ‘pruning’, during which both neurons and their synapses are lost in response to patterned experience (or rather, caused by the lack of it). One occurs in the third and fourth years of life, followed by a ‘plateau’ through the childhood years, after which there is another period of pruning in early adolescence.

Around 40% of the neurons created in the first seven months of life die by the time that children reach adulthood, and likewise the same percentage of the total synapses present at two years of age are lost by the age of sixteen. The process of pruning itself appears to obey the old adage, ‘use it or lose it.’ Neural connections that aren’t functional, or that are rarely used, are lost, whereas those that are frequently employed are retained. Are these facts inconsistent with, or otherwise problematic for, the thesis of massive mental modularity, as Buller (2005) alleges? By no means. First of all, it is important to stress that the initial placement of neuronal connections in the brain isn’t at all random. On the contrary, neurons migrate to their eventual positions within the brain, and send out their axons to quite specific locations,guided by a range of chemical gradients and other forms of signaling (Korsching,1993; Thoenen, 1995).

So the initial organization of the cortex certainly isn’t some sort of equipotent set of neural connections that is then sculpted into shape entirely by experience, as Buller contrives to suggest. On the contrary, a number of recent interventionist studies have found that the brains of animals will develop normally up to the time of birth, at least, in the absence of any relevant experience whatsoever (relevant to the formation of the brain regions under study, that is) (Miyashita-Lin et al., 1999; Verhage et al., 2000; Molnar et al., 2002). In fact it seems likely that the subsequent role of experience in brain development (in the first instance, at least) is to add precision to the inherently vague and noisy chemical signaling systems that guide the placement of initial synaptic connections, on a highly local basis.

Since neurons that perform the same function are generally arranged together in banks, by making the eventual placement of synaptic connections both activity-dependent and competitive, the brain can ensure that boundaries between systems are drawn much more precisely than would otherwise happen. A second point is that a significant amount of the neural sculpting that takes place, at least within some systems, is produced by endogenously produced signals, rather than by experience of the environment. In early visual cortex in mammals, for example, neurons are organized into sequences of columns corresponding to each of the two eyes. It has long been known that depriving cats of post-natal visual experience disrupts the formation of these columns (Hubel and Wiesel, 1970). But it is now known that these ocular dominance columns develop in stages, the first of which is guided by spontaneous neural activity produced by the two retinas prior to birth.

The basic structure of the cortical circuits is laid down in advance of any externally produced visual experience (Katz et al., 2000). And we now know that something similar is true in connection with the face-recognition system, likewise (Bednar, 2003). Moreover, a great deal of the role of experience in sculpting brain circuits isn’t well characterized as a form of learning on the basis of experience, as Buller (2005) himself acknowledges. Indeed, Katz and Shatz (1996) argue that much of the later role of experience is best seen as an extension of the work of the earlier endogenously produced stimuli, namely preparing cortical circuits to engage in learning. For in many cases all that is required is some experience or other, or experience of some very general (module-specific) type, such as face-like shapes, or object boundaries.

What a cortical circuit requires, in order for its neural components to be ‘pruned’ appropriately, is experience of the general type that falls within its intended domain. At this early stage the system isn’t so much learning in response to experience, as using that experience to distinguish itself functionally from surrounding modules. None of this is to deny, of course, that experience and learning can cause neuronal changes of various sorts to occur, some of which involve alterations in the local wiring patterns within cortical circuits. Indeed, we are beginning to discover how learning can trigger cascades of gene expression within particular groups of neurons, causing new dendrites to grow and new synapses to be formed (Ridley, 2003; Marcus, 2004). But even where this is appropriately thought of as the formation of a new brain system or the reconfiguring of an old one (as opposed to the laying down a specific memory, for example), it turns out that it isn’t inconsistent with massive modularity.

For remember that the thesis of massive mental modularity, as I am understanding it, doesn’t require modules to be genetically pre-specified. And it is still possible for the resulting system to count as an adaptation. For natural selection can rely on the presence of reliably recurring features of the environment (such as the presence of faceshaped stimuli) when selecting for particular developmental programs (Griffiths and Gray, 2001;West-Eberhard, 2003; Barrett, 2006). Moreover, I have argued in any case that some modules are a product of experience and learning, without their modular status thereby being compromised. I conclude, then, that the processes by means of which brains develop and form their cortical circuits are fully consistent with massive mental modularity, just as we earlier found for the processes through which brains evolve.

NB:
This post is taken from The Architecture of the Mind by Peter Carruthers (Professor of Philosophy at the University of Maryland, College Park). If you get benefits from reading this, please consider to buy his original work.