The first approach to evolutionary neuroscience, sometimes, is made by individuals who have a good grasp of neuroscience themes, but little comprehension of some questions of evolutionary biology – at least, that was my case! What follows is a short explanation of the concept of “sameness” in evolutionary biology. I will attempt to relate the manifold concept of sameness to different theoretical frameworks, and will use evolutionary neuroscience as long as it is possible.
Homology and homoplasy
The concept of homology is central to biology. It originated in comparative biology, systematics and evolutionary biology, but is also used in developmental and molecular biology. It is a “natural kind term”, in the sense that homologues are characters of organisms that are grouped together because they have a unity of form, which is assumed to be due to some non-trivial underlying mechanism – as opposed to being a “normative kind term”, in the sense that homologues would be characters of organisms that are grouped together arbitrarily (Wagner, 1996) “Homology” refers to similarity by conservation – that is, a given character or trait in species A is homologous to a trait in species B if it can be demonstrated that the character was conserved in evolution, which means that it is primitive. In general, inferences of homology involve techniques to determine the “ancestor state” of the character – that is, if it was present or absent in the closest common ancestor of both species A and B. For example, cladistic analysis established that the ventral subpallial nuclei of teleosts is homologous to the septum of mammals (Wullimann and Mueller, 2004). However, it does not suffice to state homology as a relation between structures; it is also necessary to state homology in relation to derivation. Consider the following example, found in Butler and Hodos (2005): the following statements are both true:
The wing of a bird is homologous to the wing of a bat
The wing of a bird is not homologous to the wing of a bat
The “specification” requirement is sufficient and necessary to distinguish the situations in which each statement is true. The wing of a bird is homologous to the wing of a bat as a deravative of the forelimb, which means that the common ancestors of birds and bats probably possessed forelimbs of a similar basic structure, and the wings of bats and birds are derivatives of those forelimbs. On the other hand, the wing of a bid is not homologous to the wing of a bat as a wing, because the forelimbs of the common ancestors of birds and bats were not wings. Citing Butler and Hodos (2005, pp. 8-9): “In statements of homology, unless the specification is obvious and unmistakable, the specific characteristic being compared must be included in the statement for the statement to be meaningful”. Thus, in our example of Vl-Vv homologies with septum, it is necessary to define whether Vl-Vv and septum are homologous as septal areas or as derivatives of subpallial areas. In the first case, it is not yet possible to fully appreciate the ancestral state of the character as a septal area, because the Vl-Vv of the common ancestor of teleosts and mammals is not known. Statistical techniques, such as phylogenetic generalized least squares estimation of ancestor states (Martins and Hansen, 1999), are available to generate estimates for this trait, but there is still the problem of defining which properties of the trait should be included in the analysis.
The concept of homology as the relationship between two characters in two different species as inherited from a common ancestor is called historical homology. It is based mainly on a “transformational” approach, which states that a character A in species X is considered homologous to a character A' in species Y if it takes “fewer steps to transform a into A' than it takes to transform A into A' than it takes to transform A into B'” (McKitrick, 1994). This “transformational” approach is superimposed by taxonomic considerations – that is, the distribution of the character on a phylogenetical tree. The determination of historical homology is commonly based on the analysis of fossils; however, in evolutionary neuroscience, one must face the problem of the impossibility of fossilization of the brain. Simpson (1961) proposed that other criteria should be used to establish historical homology hypotheses for neuroanatomical data. These include:
Similarity of axonal connections (hodology)
Similartiy in the relationships between the group of neurons in analysis to some consistent feature of the species analysed
Similarity of embriological derivation
Similarity in the morphological features of individual neurons that form the group
Similarity in the neurochemical attributes of the neurons that form the group
Similarity in the physiological properties of the neurons that form the group
Similartiy in the behavioral outcomes of neuronal activity
This is known as the “requirement of total evidence” epistemological principle (Fitzhugh, 2006), and gave rise to an endless controversy on whether the last two criteria – being functional criteria – should be included in homology analyses. Despite these controversies, “the more of these criteria that can be satisfied, the stronger the support for an hypothesis of historical homology” (Butler and Hodos, 2005, p. 9).
The opposite of historical homology is homoplasy, defined as the structural similarity between two traits in two species without phyletic continuity – which is equivalent to saying that, even though the traits are similar, the common ancestor of species A and B did not present the trait. There are three different types of homoplasy: convergence, parallelism, and reversal. The first type refers to the evolution of similar traits in response to similar adaptive pressures, but not to similar genes and developmental processes; an example of convergence can be found in the eletroreception of mormyrids and gymnotoids: while the organs responsible for this perceptual capacity are similar, they are not derived from a common ancestor. Parallelism occurs in closely related taxa, and is defined as the independent development of a descendant character that is not present on a common ancestor. Parallelism occurs when two taxa develop the same character after evolutionary divergence; since the trait is absent in a common ancestor, but present in both descendant species, it is probable that the developmental genetics that produces the structures in the different taxa is the same, which means it was inherited from the common ancestor. Thus, there is homology between the developmental and genetic materials, but not on the final structure. Parallelism is the great challenge for statistical cladistic approaches, since they cannot detect whether the character was absent in the common ancestor without complementary approaches. Hennig (apud Butler and Hodos, 2005) stated an auxiliary principle that goes like this: “Never assume convergent or parallel evolution; always assume homology in the absence of contrary evidence”. This is based on the “transformational” approach to parsimony. Reversals are instances of homoplasy in which a character appears, subsequently disappears, and later reappears along the descendants in one lineage. Statements of reversals, instead of parallelism and convergence, must be analysed through the “transformational” approach to parsimony.
There has been considerable debate on whether the historical homology concept is satisfactory to phylogenetical analyses. It has been argued that this concept is not capable of recognizing the importance of genetic and developmental continuity as bases for homology. Two alternatives to this approach appeared. Biological homology is defined as the morphological identity of characters. It focuses, thus, “on the developmental pathways and the behavior of morphogenetic fields [ie, discrete units of embryonary development] to account for the variability of character expression but does not define sameness by them”. Thus, a trait A in species A is considered homologous to a trait A' in species B if A and A' share a set of developmental constraints (Wagner, 1989). The requirement of similar phenotypic expression from historical homology is not taken as valid, and the requirement of phyletic continuity applies only to developmental units.
The concept of generative homology is a second alternative to historical homology, and proposed by Butler and Saidel (2000). It is defined as “the relationship of a given character in different taxa that is produced by shared generative pathways” (Butler and Saidel, 2000, p. 849), ie, shared genetic and/or morphogenetic basis. Generative homology encompases parallelism, reversals, and most cases of historical homology. It's opposite is convergence, and not all instances of homoplasy.
It must be noted that these different concepts of homology normally overlap, and are segregated only by different uses in different areas. While systematists are concerned with the reconstruction of phylogenetic relationships and the recognition of monophyletic groups – thus relying on historical homology –, evolutionary biologists are interested in the evolution of a trait, and rely on biological homology. Comparative developmental biologists, on the other hand, are concerned with evolution of generative pathways for traits, and, accordingly, use the concept of generative homology. Evolutionary neuroscience is normally interested in the evolution of a trait, and can sometimes be interested in the comparative developmental character of a trait; thus, evolutionary neuroscientists normally rely on biological and generative homology concepts. This does not mean that the criteria for homology developed by Simpson (1961) should be discarded; rather, they are complementary criteria for deciding on hypotheses of homology. A complete account of the evolution, generative status and cladistic status of a trait should refer to those criteria.
A considerable ammount of neuroscientists are not that interested in analysing form; most of all, they are interested in the function of a given structure. This is specially relevant in the case of behavioral neuroscience – the field from where I came. “Yes, very neat, this structure is homologous in species A and B, and it presents such and such properties, but what does it do?”: this is the behavioral-neuroscientific equivalent of “but does it make coffee?”.
In contrast, evolutionary biologists are not that much worried about function. The concept of analogy was created to address functional questions in that field. Analogy refers to similarity of function, regardless of phyletic relationships. The rationale for identifying analogy as a different concept than homology is that structures that present different morphology and origin (be it phyletic or embryological) can have similar functions. As an illustration, let's return to the bat and bird example again. The wing of a bird and the wing of a bat are homoplastic as wings, but share the same function – flying. This distinction is important, because it contains in itself the idea that relations between form and function are not always correlated to relations between form and phyletic relationships. The establishment of such relations must be made independently. A good strategy is to relate, via statistical analyses, the phylogenetic correlations between a given structure and a given function, which can be physiological or behavioral. It is teoretically sound that physiological functions will present greater correlations with morphological traits – since it is a basic tenet of neuroscience that the physiological properties of a given structure are reducible (either literally or in form of supervenience) to the underlying organization. The relations between behavior and morphological traits, on the other hand, are less well established.
The field of behavioral phylogenetics is considerably new, and, so far, was not totally incorporated in evolutionary neuroscience. Its aim is to analyse the phylogenetical distributon of a given behavioral trait, using mainly statistical techniques. Phylogenetic questions about behavior is hindered by the fact that behavior patterns reflect individual variability, plasticity, and responses to environmental change; thus, homology hypotheses on behavior should be rather weaek (Gittleman and Decker, 1994). However, diverse malleable patterns of behavior present some phylogenetic inertia (Wilson, 1975), because closely related species tend to evolve in similar niches, have similar genetic variance for selection to act upon, and have similar phenotypes, tending to respond to environmental change in a similar fashion (Harvey and Pagel, 1991). Ideally, statistical techniques could be used to analyse the rate of change and conservation in a given set of traits – say, allometric data on amygdala and homologues and behavioral measures of fear. However, this is not enough to establish that a given analogous trait presents homology, be it historical, biological or generative. To my knowledge, there have been so far no studies attempting to correlate neural and behavioral traits in evolution. I may attempt to do so in the near future, as long as I can access a reliable data set.
That is it for today. Now, go away!
Butler AB, Saidel WM (2000). Defining sameness: Historical, biological and generative homology. BioEssays 22: 846-853. http://doi.wiley.com/10.1002/1521-1878(200009)22:9%3C846::AID-BIES10%3E3.0.CO;2-R
Fitzhugh K (2006). The 'requirement of total evidence' and its role in phylogenetic systematics. Biology and Philosophy 21: 309-351. http://dx.doi.org/10.1007/s10539-005-7325-2
Gittleman JL, Decker DM (1994). The phylogeny of behaviour. In: PJB Slater, TR Halliday (eds.), Behaviour and Evolution. Cambridge: Cambridge University Press.
Harvey PH, Pagel MD (1991). The Comparative Method in Evolutionary Biology. Oxford: Oxford University Press.
Martins EP, Hansen TF (1999). Phylogenies and the comparative method: A general approach to incorporating phylogenetic information into the analysis of interespecific data. American Naturalist 149: 646-667.
McKitrick M (1994). On homology and the ontological relationships of parts. Systematic Biology 43: 1-10.
Simpson GG (1961). Principles of Animal Taxonomy. New York: Columbia University Press.
Wagner GP (1996). Homologues, natural kinds and the evolution of modularity. American Zoologist 36: 36-43.
Wilson DS (1975). Sociobiology: The New Synthesis. Cambridge: Harvard University Press.
Wullimann MF, Mueller T (2004). Teleostean and mammalian forebrains contrasted: Evidence from genes to behavior. Journal of Comparative Neurology 475: 143-162. http://dx.doi.org/10.1002/cne.20183