Evolution of Plasticity in Brain Morphology

Phenotypic plasticity, the ability of a single genotype to result in different phenotypes in response to environmental cues, is ubiquitous across life. Recently, the degree to which a trait can be plastic has itself been recognized as an important characteristic of organisms that can be selected upon and evolved [1]. Since plasticity can produce adaptive responses to rapidly changing environments [2‒4], it is critical to understand how plasticity of fitness-related traits evolves in response to natural selection.

Brain morphology has long been recognized as a trait that is critical to organism performance across environmental variation. Brains vary in size across animal species by many orders of magnitude [5]. This variation has been linked to cognition and behavioral variation [5‒9]. Brains are subdivided into regions that are associated with specific aspects of cognition, behavior, and sensory perception [5, 10‒15]. Differences in brain morphology (brain size and the proportional size of brain regions) have been linked to differences in ecological characteristics both between [16‒21] and within species [22‒27]. Further, brain morphology can be influenced by both genetic variation and phenotypic plasticity, as demonstrated by common garden studies [26, 27], artificial selection studies [8], and experimental plasticity studies (outlined in detail below). There has been, however, only limited research examining the degree to which plasticity in brain morphology can itself evolve, as well as the environmental patterns that may select such plasticity. Both brain morphology and phenotypic plasticity are hypothesized to be critical traits influencing responses to rapidly changing environments, so gaining a better understanding of their interaction may provide important insight into how organisms respond to a variety of environmental challenges.

The goal of this review was to provide a framework to improve our understanding of the evolution of plasticity in brain morphology. We do so by first summarizing what we know about plasticity in brain morphology. We focus on the range of external cues that can lead to plastic responses rather than the adaptive function of the observed plasticity. Second, we outline proposed ecological patterns that we hypothesize may cause the evolution of brain plasticity, with the goal of establishing a framework for future work examining the evolution of brain plasticity.

Numerous environmental cues lead to plastic responses in brain morphology. First, the most typical environmental cue inducing plasticity in brain morphology, typically brain size, is environmental enrichment. Deprivation of environmental stimuli can lead to the development of smaller brains across taxa, including in fish [28‒31], birds [32], and mammals [33]. This also results in a common observation that animals held in captivity, particularly in laboratory-based environments, develop smaller brains than wild individuals [34‒37].

Second, a related pattern, known as the social brain hypothesis, predicts that larger brain size should be associated with more complex social situations. Social rearing has been shown to influence brain morphology development. For example, common frog tadpoles (Rana temporaria) reared at high density developed larger brains [38] and larger optic tecta (the brain region associated with visual information processing) than those reared at low density [39]. Similar results were found by Axelrod et al. [24], with Trinidadian guppies (Poecilia reticulata) developing larger brains when reared in groups than when reared alone. Further, Ott and Rogers [40] found similar results in the desert locust (Schistocerca gregaria), with individuals reared at high density developing 30% larger brains than those reared at low density. However, the opposite effect was found in ninespine stickleback (Pungitius pungitius), with larger brains developing at lower density than higher density [26]. A more complex social effect was found in response to the sex ratio in Trinidadian guppies, with male guppies developing larger brain size when reared with females than when reared with males [41].

Third, specific ecological factors have also been found to cause plastic variation in brain morphology. The presence of predators, usually tested by introducing predator olfactory or visual cues, has been found to influence the development of brain morphology, though in varying ways across species. For example, Gonda et al. [38] found that common frog tadpoles reared with predator cues developed smaller brains than those reared without. Further, Trokovic et al. [39] found that frogs developed a smaller diencephalon, the region of the brain thought to integrate information across other regions, when reared with predator cues. Both these studies suggest a reduced importance of cognitive ability in the presence of predators. However, the opposite trend has been shown in fish, with Trinidadian guppies developing larger brains when reared with predator cues (Axelrod et al. [24]). Further, both ninespine stickleback [42] and Trinidadian guppies (Axelrod et al. [24]) developed larger olfactory bulbs when reared with olfactory predator cues. It is possible that inconsistencies in responses to predator and social cues are the result of differences in the conditions of the experiments, though these contradictory findings suggest that the impact of predation and social environment on brain morphology may vary across systems.

Fourth, abiotic characteristics of the environment, such as temperature, light, and oxygen, can also influence the development of brain morphology. Temperature during development has been shown to affect brain morphology across taxa, though with inconsistent results. Eastern three-lined skinks (Bassiana duperreyi) hatched under colder temperatures developed larger telencephalons, the region of the brain associated with higher order reasoning, than those hatched under warmer temperatures [43]. Conversely, Gu et al. [44] found no relationship between brain morphology and temperature in Gunther’s frogs (Hylarana guentheri). The light environment during ontogeny can also influence brain development. Oriental fire-bellied toad tadpoles (Bombina orientalis) reared in the dark developed smaller brains and visual centers of the brain than those reared in light environments [45], and lesser earless lizards (Holbrookia maculate) reared against dark sand developed a larger medial cortex than those reared against white sand [46]. Oxygen levels can also influence brain morphology development, with African cichlids (Pseudocrenilabrus multicolor) reared in high oxygen conditions developing larger relative brain sizes than those reared in low oxygen conditions [47].

Fifth, broader environmental conditions can also influence brain morphology through seasonal changes. For example, round gobies (Neogobius melanostomus) were found to have larger telencephalons in spring than in autumn. Further, female Western Fence Lizards (Sceloporus occidentalis) were found to have a larger dorsal cortex during the breeding season than during the post-breeding season [48]. Further, seasonal effects on brain morphology have been found in birds. Nottebohm [49] found that the volume of the telencephalic nuclei related to song control of Canaries (Serinus domestica) was larger in the spring than in the fall, likely related to song learning. Similar results showed plasticity in brain morphology across seasons in the Black-capped chickadee (Parus atricapillus) [50], European Starling (Sturnus vulgaris) [51, 52], and Ruffed Grouse (Bonasa umbellus) [53]. Though these differences appear to be caused by environmental cues, it is important to note that, because these changes to brain morphology occur on predictable timelines across seasons, it is possible that they represent pre-programed shifts during ontogeny rather than changes induced by exposure to particular cues.

The final category of cue that has been found to influence brain morphology is human-created chemicals that animals may encounter in nature. For example, Leopard frog tadpoles (Lithobates pipiens) exposed to the pesticide chlorpyrifos developed smaller brains than controls [54]. Similarly, Campbell et al. [55] found that leopard frogs develop a smaller cerebellum when reared with a noenicitinoid-based insecticide.

One observation that has been suggested previously but rarely demonstrated empirically is that brain morphology plasticity differs across broad taxonomic groups. While invertebrates, fish, amphibians, and reptiles show plasticity in brain morphology in response to various environmental cues, birds and mammals only show plasticity in response to structural enrichment or seasonality, as outlined above for birds. Even fewer examples of plasticity in brain morphology have been shown in mammals, with only environmental enrichment having been shown to increase glial cell depth in rat brains [33]. One proposed example of brain morphology plasticity in humans is an increase in hippocampus size (the brain region related to navigation) in London taxi drivers [56]. It remains unclear why taxa show these differences in capability for brain morphology plasticity.

The examples outlined above illustrate the range of environmental cues that brain morphology plastically responds to during development. However, there remain broad gaps in the scope of our knowledge that future work can address, in addition to the gaps related to the evolution of brain morphology plasticity that we discuss in the next section. First, as we outlined above, broad taxonomic groups are expected to differ in levels of plasticity in brain morphology. Indeed, we found very limited examples of experimental plasticity in birds and mammals (8 studies, though 6 of these studies were correlational across seasons), with these only showing brain morphology plasticity to seasonality and environmental enrichment. However, we found many examples of plasticity in fish and amphibians (23 studies), with these responding to a wide variety of cues. This may represent bias in the choice of animals that tend to be used in these kinds of experiments, perhaps due to ethical concerns or ease of laboratory rearing, or might indicate genuine physiological differences across groups. For example, while mammals stop widespread neurogenesis throughout the brain once adulthood is reached, fish can maintain the generation of new neurons across the brain into adulthood [57], which may allow them to exhibit more plasticity throughout life. One potential explanation is that mammals and birds may, instead of broad plasticity in brain size or region sizes, make use of more subtle morphological plasticity or cellular plasticity, such as neurotransmitter levels or neuronal structure, to respond to environmental variation. Further research across taxa is needed to demonstrate the mechanism underlying differences in brain morphology plasticity. A second question is whether there are consistent patterns to the function of changes in brain morphology in response to particular environmental cues. Due to differences in the specific nature of experiments and characteristics of different species, it is not yet clear why some cues elicit particular plastic responses in the brain and others do not, or why cues can elicit different responses in different species. More standardized tests of the effect of plasticity cues on brain morphology are needed to create generalized expectations of how environmental cues may influence brain morphology. A final question is how brain morphology plasticity may impact species responses to changing environmental conditions.

Much of the research into brain morphology plasticity asserts that observed plastic changes are adaptive, with plastic shifts resulting in a better match between brain morphology and environmental conditions. If true, this implies that plasticity is shaped by evolutionary change over generations. There has been, to this point, limited research testing evolutionary change in brain morphology plasticity. However, the small amount of research that has been done suggests that this plasticity does have the potential to evolve and respond to natural selection [24, 26, 42, 47, 58]. As shown above, we have a general sense of what environmental cues can impact brain morphology through plasticity. What remains unclear is what environmental conditions may select for plasticity in brain morphology. Here, we outline four hypothesized patterns of environmental variation that may lead to the evolution of increased brain morphology plasticity (summarized in Table 1).

The evolution of increased plasticity in morphology is generally expected to be linked to temporal variation in ecological conditions relevant to the function of a particular trait [1, 59]. For example, Huber et al. [60] found that Common Jewelweed (Impatiens capensis) that exhibited plasticity in shade avoidance behavior was favored by selection in more heterogenous microhabitats. A similar process may select for increased brain morphology plasticity. Brain morphology variation is thought to influence animal performance and fitness through functional links to cognition, behavior, and sensory perception [5]. If the specific optima of these characteristics are different between environmental conditions, then different sizes of brains and brain regions will similarly have different optimal levels across environments. When these environments vary within the lifetime of an individual, then plasticity in cognitive function will be favored, which in turn may favor plasticity in brain morphology. Though few experimental tests of this hypothesis exist, some evidence for this hypothesis has been found. For example, African cichlids (P. multicolor victoriae) with more potential for dispersal, and therefore a greater likelihood of encountering variable environmental conditions, were found to have more plastic brain morphology than those in more static environments [47]. Similarly, sticklebacks from pond environments (which are thought to be more temporally variable) show more plasticity in brain morphology than those from the ocean (Gonda et al. [26, 38]).

The eco-cognitive variation hypothesis predicts that brain morphology should be more plastic under more environmentally variable conditions, particularly when these conditions differ in their specific cognitive requirements. Testing this hypothesis requires comparing levels of plasticity between groups that differ in their levels of environmental variability but ideally do not differ in other characteristics that are predicted by other hypotheses outlined below. The ideal experiment would use populations of a species that differ in their levels of brain morphology plasticity. These could use pre-existing genetic variation in natural populations or be created using artificial selection. These populations would then be introduced to artificially controlled experimental conditions that have manipulated differences in environmental variation. The eco-cognitive variation hypothesis would predict that the population with greater plasticity would show higher fitness under the more variable conditions.

Though environmental variation in cognitive requirements is generally expected to be the primary force selecting morphological plasticity in brains, it is possible that brain morphology, and in particular brain size, may operate differently because of the relationship between brain size and cognitive function. Flexibility of trait function is usually expected to be linked to trait form; however, this is not necessarily true with respect to brain size. Greater levels of cognitive flexibility, rather than being associated with flexibility in brain size, have been found to be more related to larger brain size [7, 61‒63]. Big brains, not flexible brains, lead to cognitive flexibility. As such, variability in eco-cognitive conditions that require cognitive flexibility may select for large brains rather than plastic brains. So, if not flexibility in the functional requirements of traits, then what environmental conditions select for brain morphology plasticity?

The energy variation hypothesis posits that variability in the nutritional environment of animals is the primary pattern selecting for plasticity in brain morphology. Brain tissue is highly metabolically costly [8, 64‒66]. As such, maintaining a large brain size or large brain region sizes requires an adequate influx of nutrition. When food is limited, a large brain may reduce fitness by diverting metabolic resources away from other necessary biological functions such as body growth, digestions, and reproduction. Ledón-Rettig et al. [67] showed that a higher quality diet can induce a plastic increased telencephalon size in spadefoot toads (Spea bombifrons). Since differences in the amount of available food or nutrition can influence the fitness consequences of different brain sizes, then temporal variability in food availability could select for plasticity in brain morphology to account for such variation. If this is the primary driver of the evolution of brain morphology plasticity, then we would expect to see higher plasticity in environments that vary in their nutrient availability, for example, habitats with highly seasonal variation in food availability or areas with highly variable community dynamics.

Traits do not occur in isolation but develop in the context of other characteristics of organisms. These other characteristics can influence the development and evolution of traits through functional or developmental correlation. Brain morphology, and brain size in particular, has been shown to be correlated with a variety of traits across taxa. In mammals and birds, brain morphology is, by functional necessity, highly correlated with skull morphology [68, 69]; This correlation is not present to the same degree in other taxa, at least within species as brain tissue is usually not pressed up against the inside of the skull [12]. Other traits that have been found to be negatively correlated with brain size are gut size [64, 70, 71], fat storage [71, 72], and testis size [73, 74]. This correlation is thought to be the result of a tradeoff in energy allocation between energetically expensive tissues (Aiello and Wheeler [64]).

These trait correlations may be critical to the evolution of brain morphology plasticity if they cause correlated plasticity of traits. If two traits are correlated, then selection on one trait will be expected to cause selection on the other. Therefore, if the plasticity of two traits is correlated, we can hypothesize that selection on the plasticity of one trait may cause correlated selection on the plasticity of the other trait. If true, then environmental variability that leads to selection on the plasticity of any trait correlated with brain morphology, for example, skull shape, gut size, or reproductive physiology, will also cause selection for increased or decreased plasticity in brain morphology. This hypothesis predicts that shifts in the level of plasticity in brain morphology should coincide with similar shifts in plasticity in other correlated traits. The ideal test of this hypothesis would employ artificial selection techniques to select the plasticity of a trait expected to be correlated with brain morphology, with the expectation that brain morphology plasticity would also evolve along with the focal trait.

One final environmental pattern that has been hypothesized to lead to the evolution of plasticity generally is rapid environmental change, typically through colonization of novel habitats. This idea is commonly known as the “plasticity first” evolution hypothesis [75]. This process occurs when two environments differ in their optimum phenotypic value of a particular trait, and a population moves from one of these environments to the other. After the initial move, individuals in the population that are able to shift their phenotype to more closely match the conditions of the new environment, generally those with higher levels of phenotypic plasticity, will best survive and reproduce. This pattern requires no temporal variability within individual environments and selects for plasticity as a byproduct of the difference in optimum phenotype between environments. It will apply to brain morphology if organisms colonize new habitats that differ in their optimal brain morphology. Evidence for this pattern has been found in fish brain morphology, with higher levels of brain morphology plasticity in colonized compared to ancestral populations of Pumpkinseed Sunfish (Lepomis gibbosus) [58] and Trinidadian Guppies (P. reticulata) [24]. Further, Gonda et al. [26, 38] are consistent with this pattern as they found greater plasticity in colonized pond sticklebacks than ancestral marine sticklebacks.

Generally, the colonization plasticity selection hypothesis predicts that populations in more recently colonized habitats, specifically those that differ from ancestral habitats in required cognitive or energetic characteristics, should show greater plasticity. Further, it predicts that changes in brain morphology plasticity should decline after the initial colonization event as selection for plasticity would no longer be acting, an idea known as genetic assimilation [75‒77].

An important final question in examining the patterns of selection that cause the evolution of brain morphology plasticity is what environmental conditions lead to selection against such plasticity rather than for it. Generally, as highly variable environments are expected to select for plasticity, highly stable environments are hypothesized to select against plasticity (Schlichting and Pigliucci [1]). With respect to brain morphology plasticity, we can hypothesize a similar pattern, though each of the presented hypotheses points to environmental stability of a different sort selecting against plasticity in brain morphology. The eco-cognitive variation hypothesis points to environments with stable eco-cognitive conditions, the energy variation hypothesis points to stability in the nutritional environment, the correlation plasticity evolution hypothesis points to stability in environmental conditions specifically related to traits correlated with brain morphology, and the colonization plasticity selection hypothesis points to stable habitats over time. Finally, the colonization plasticity selection hypothesis predicts that populations living in the same habitat for long periods of time should show selection against plasticity in brain morphology. However, all of these patterns make the assumption that plasticity in brain morphology presents a cost to fitness in stable environments, as this is needed for selection to act against a trait [78]. To date, no empirical evidence has demonstrated that there are metabolic or functional costs innate to plasticity in brain size or morphology, so future work is needed to demonstrate why plasticity is not favored, or at least neutral, in all circumstances. This is of particular interest when revisiting the question of why levels of brain morphology plasticity differ across taxa as these differences suggest that this plasticity may have been broadly disfavored in mammals and birds.

Both phenotypic plasticity and brain morphology have been proposed as characteristics of organisms that will be critical to population adaptation and persistence in the face of rapidly changing environmental conditions. Therefore, understanding the way these traits interact through plasticity in brain morphology and particularly how this plasticity may evolve, may become critical as we push our understanding of evolution. Though very little empirical evidence exists showing population-level differences in plasticity of brain morphology, the presence of a handful of studies in fish [24, 26, 42, 47, 58] indicates that it may be an important mechanism of population diversification. Moving forward, more studies explicitly testing the outlined hypotheses by comparing populations that differ in their levels of environmental variability with respect to cognitive and nutritional traits, as well as the colonization history of populations are needed. Finally, artificial selection studies, imposing selection on plasticity, would be ideal for deciphering the causes and consequences of selection on plasticity in brain morphology. These proposed experiments would be logistically complex, would require detailed knowledge of particular study systems, and would likely involve multiple iterations to refine techniques and test alternative hypotheses. However, we believe they represent the best way forward toward a better understanding of the evolution of plasticity in brain morphology.

We thank A. López-Sepulcre for comments on the article.

The authors declare no conflicts of interest.

This project was provided by funds from Cornell University (awarded to S.P.G.).

C.J.A. and S.P.G. conceived of the manuscript. C.J.A., H.S., S.M.T., D.C.D., N.M.F., N.G., M.R., and N.V. contributed to the reviewing the literature. C.J.A. drafted the manuscript. All authors provided comments and edits to the manuscript.