Relative Size versus Absolute Size

The most obvious difference between species is that they differ enormously in size. Because life began with tiny organisms, evolutionary increases in body size must have outnumbered or outpaced the decreases. This is true of organisms generally, but it also holds for several individual lineages, including mammals and, within mammals, primates (Stanley, 1973; Alroy, 1998). The most fascinating aspect of those changes in body size is that they involved much more than the isometric scaling up or down of the ancestral condition; they involved allometric changes in the proportions of body parts and physiologic processes. For example, skeletal mass increases disproportionately with increasing body size, whereas heart rate decreases. Countless studies - on both vertebrates and invertebrates - have documented these allometries and explored their functional implications (Calder, 1984; Schmidt-Nielsen, 1984).

Much less is known about the causes of allometry. Studies on allometry in insects showed that some scaling relationships are readily modifiable by natural or artificial selection (see Emlen and Nijhout, 2000; Frankino et al., 2005). This finding suggests that even tight scaling laws are not immutable, which would explain why many traits scale differently (e.g., with different exponents) in different taxonomic groups (Pagel and Harvey, 1989). A very different, more theoretical line of research has shown that numerous allometries, specifically those with power law exponents that are multiples of 1/4, may have evolved because the optimal means of delivering metabolic energy to cells is through an hierarchically branching, fractal network of vessels whose termini (e.g., capillaries) are body size-invariant (West et al., 1997; Savage et al., 2004; West and Brown, 2005). This theory is mathematically complex and still controversial (Kozlowski and Konarzewski, 2004; Brown et al., 2005; Hoppeler and Weibel, 2005), but it is elegant. Furthermore, because the theory of West et al. is based in part on the assumption that natural selection optimizes phenotypes, it is consistent with the aforementioned finding that allometries are modifiable by selection. However, West et al.'s (1997) theory cannot explain (or does not yet explain) why some organs, such as the brain, scale with exponents that are not multiples of 1/4. Nor can it easily explain taxonomic differences in scaling exponents. Thus, the causal - physiological and/or developmental - bases of allometry are coming into focus but remain, for now, mysterious.

Brain scaling, in particular, remains quite poorly understood (see Principles of Brain Scaling, Scaling the Brain and Its Connections, How to Build a Bigger Brain; Cellular Scaling Rules for Rodent Brains). The discovery that brains become proportionately smaller with increasing body size dates back to the late eighteenth century (Haller, 1762; Cuvier, 1805-1845). Since then, numerous studies have documented brain allometry in all the major groups of vertebrates (Deacon, 1990a; van Dongen, 1998) and even some invertebrates (Julian and Gronenberg, 2002; Mares et al., 2005). Generally speaking, those studies confirmed that in double logarithmic plots of brain size versus body size, the data points for different species within a given lineage tend to form a reasonably straight line, indicating the existence of a simple power law. The slope of those best-fit lines are almost always less than 1, which reflects the aforementioned fact that brains generally become proportionately smaller with increasing body size. The large body of work on brain-body scaling further revealed that data points for different taxonomic groups often form lines with similar slopes but different y intercepts. These differences in y intercepts are known as differences in relative brain size or encephalization. They seriously complicate efforts to draw a single allometric line for any large taxonomic group (Pagel and Harvey, 1989), but they allow us to identify evolutionary changes in relative brain size among some smaller taxonomic groups. For example, they allow us to determine that relative brain size increased with the origin of mammals, with the origin of primates, several times within primates, with the origin of the genus Homo, and, last but not least, with the emergence of Homo sapiens (see Primate Brain Evolution in Phylogenetic Context, The Hominin Fossil Record and the Emergence of the Modern Human Central Nervous System, The Evolution of Human Brain and Body Growth Patterns). Overall, such phylogenetic analyses suggest that, among vertebrates, relative brain size increased more frequently than it decreased (Striedter, 2005).

Enormous effort has gone into determining the functional significance of evolutionary changes in brain-body scaling. Darwin, for example, had argued that relative brain size is related to "higher cognitive powers'' (Darwin, 1871), but defining those powers and comparing them across species has proven difficult (Macphail, 1982). Consequently, most subsequent investigators shied away from the notion of general intelligence, or 'biological intelligence' (Jerison, 1973), and focused instead on more specific forms of higher cognition. Parker and Gibson (1977), for example, proposed that a species' degree of ence-phalization is related to its capacity for extracting nutritious fruits and nuts from their protective shells. Several authors have stressed correlations between brain size and 'social intelligence' (Byrne and Whiten, 1988; Dunbar, 1998; Reader and Laland, 2002). Collectively, these studies reinforced the sense that relative brain size is, somehow, related to some forms of intelligence. However, relative brain size also correlates with several other attributes, such as longevity, home-range size, diet, and metabolic rate (for a review, see van Dongen, 1998). The latter correlations, with diet and metabolism, have received particularly lavish attention (Martin, 1981; McNab, 1989; Aiello and Wheeler, 1995). Paradoxically, the discovery of so many correlations has led some evolutionary neuroscientists to despair: there are too many correlates of relative brain size, and many of them come and go, depending on which taxonomic group is being examined and which statistical methods are used for the analyses (e.g., Bennet and Harvey, 1985; Iwaniuk et al., 1999; Deaner et al., 2000; Beauchamp and Fernandez-Juricic, 2004; Jones and MacLarnon, 2004; Martin et al., 2005). Too many contested hypotheses, too little certitude.

There is not much clarity on why brains scale so predictably with body size. Early workers argued that brains generally scale against body size with a power law exponent close to 2/3 because the brain's sensory and motor functions were related to the body's surface area, which presumably scales with that same exponent (Snell, 1891; Jerison, 1973). According to this view, brain sizes in excess of that predicted by the 2/3 power law are due to increases in the brain's nonsomatic, cognitive regions. This would explain the correlations between relative brain size and some forms of intelligence. Unfortunately, there are two major problems with this view. First, brain-body scaling exponents often differ substantially from 2/3 (van Dongen, 1998; Nealen and Ricklefs, 2001). The second problem is that the brain's more cognitive regions also scale predictably with body size (Fox and Wilczynski, 1986), undermining the assumption that brains are divisible into regions that scale with body size and regions that do not. Therefore, the excess neuron hypothesis (Striedter, 2005) is dead. In searching for an alternative, some have suggested that brain-body allometry is linked to the scaling of metabolic rates. This hypothesis is based on the observation that, in at least some taxonomic groups, brain size and basal metabolic rate scale against body size with similar exponents (Martin, 1981; Mink et al., 1981). However, other studies have shown that the correlation between brain size and metabolism is not tight, once the mutual correlation with body size is factored out (McNab, 1989). This correlational slack presumably arises because species differ in how much of the body's total energy supply they deliver to the brain (Aiello and Wheeler, 1995; Kaufman, 2003), but this just underscores that relative brain size is not so tightly linked to metabolic rate.

Overall, the lack of clarity on what causes brains to scale predictably with body size, and how to interpret deviations from the scaling trends, has caused interest in relative brain size to fade. Increasingly, evolutionary neuroscientists have turned away from relative brain size and asked, instead, how the size of individual brain regions correlates with various behavioral parameters (Harvey and Krebs, 1990; see Brain Size in Primates as a Function of Behavioral Innovation, Mosaic Evolution of Brain Structure in Mammals). This shift in research strategy makes sense, because, after all, the brain is functionally heterogeneous. However, even studies that focus on correlations between single brain areas and specific behaviors -some refer to them as neuroecological studies - are controversial because: (1) the behavioral parameters are difficult to quantify and/or define (Bolhuis and Macphail, 2001), (2) neuronal structure-function relationships are complex and often poorly understood, (3) it is difficult to decide a priori whether one should correlate behavioral parameters against a region's absolute size, its proportional size, or its size relative to expectations (Striedter, 2005), and (4) the methods for establishing statistically significant correlations in phylogenetic data remain debatable (Felsenstein, 1985; Garland et al., 1992; Smith, 1994; Martin et al., 2005). Brave neuroscientists are continuing to tackle those problems, but the larger problem of how to deal with relative brain size - how to find its causes and its functional significance - is fading from view. Perhaps we need a new approach to understanding relative brain size - perhaps one that is linked more directly to the physiological and geometric properties of brains (West and Brown, 2005) - but this novel direction is not yet apparent.

As interest in relative brain size waned, interest in absolute brain size waxed, mainly because many of the brain's internal structural and functional features turn out to scale predictably with absolute brain size. Best studied is the phenomenon of size-related shifts in brain region proportions (Sacher, 1970; Finlay and Darlington, 1995). In mammals, for example, the neo-cortex becomes disproportionately large as absolute brain size increases, whereas most other regions become disproportionately small. A second interesting scaling law is that a brain's degree of structural complexity tends to increase with absolute brain size. Within the neocortex, for example, the number of distinct areas increases predictably with neocortex size (Changizi and Shimojo, 2005). A third fascinating aspect of brain scaling is that the amount of white matter within mammalian brains scales allometrically with absolute brain size (Ringo, 1991; Zhang and Sejnowski, 2000). This connectional allometry, taken together with the fact that synapse size and density are relatively size-invariant, indicates that brains become less densely interconnected, on average, as they increase in size (Stevens, 1989; Deacon, 1990a, 1990b; Striedter, 2005; see Scaling the Brain and Its Connections). All of this signifies that brains change structurally in many ways as they vary in absolute size. Many of those changes have clear functional implications. For example, it has been suggested that, as hominid brains increased in size, the axons interconnecting the two cerebral hemispheres became so sparse and long that the hemispheres became less capable of interacting functionally, which led to an increase in functional asymmetry (Ringo et al., 1994; see Cortical Commissural Connections in Primates, The Evolution of Hemispheric Specializations of the Human Brain). Considerations such as these suggest that absolute brain size is a much better predictor of brain function than relative brain size, at least among close relatives (Striedter, 2005).

In retrospect, we can say that evolutionary neuros-cientists historically have overemphasized relative brain size. As Dunbar (2006) put it, comparative neu-robiologists have too long been "dragooned into worrying about relativizing brain size by a very peculiar view that body size must be the default determinant of brain volume.'' Can we explain this undue emphasis? Partly, evolutionary neuroscientists may have worried that focusing on absolute brain size and linking it to higher cognitive powers would force us to conclude that whales and elephants, with their enormous brains, are smarter than humans (see Cetacean Brain Evolution, Evolution of the Elephant Brain: A Paradox between Brain Size and Cognitive Behavior). This is a valid concern, for few would doubt that humans are - or at least can be - the most intelligent creatures on earth. However, whales and elephants are behaviorally complex, and humans may well be special because they are unique in possessing symbolic language (Macphail, 1982). Furthermore, it seems to me that large whales, with large brains, are more intelligent (both socially and in their hunting strategies) than dolphins or small whales. This hypothesis remains to be tested, but it points to a strategy for reconciling absolute and relative brain size: among close relatives, comparisons of absolute brain size are most informative, but in comparisons of distant relatives (e.g., whales and humans), relative brain size is a more potent variable (Striedter, 2005). This view is consistent with the finding that, among primates, social group size correlates more strongly with absolute brain size than with relative brain size (Kudo and Dunbar, 2001; Striedter, 2005). It also serves as a productive counterweight to the field's traditional, almost exclusive emphasis on relative brain size.

Was this article helpful?

0 0

Post a comment