Biological Analogs of Personality
Biological Analogs of Personality
Abstract and Keywords
Individual differences in behavior that are stable over time and correlated across different contexts can be found in a wide range of species across the animal kingdom. Such structured behavioral differences have been termed “animal personalities” (behavioral syndromes). This chapter provides a brief introduction into this research area and discusses the two most common examples of behavioral variation associated with animal personalities (boldness-aggression syndrome and responsiveness to environmental stimuli); key genetic and physiological correlates of animal personalities; and fitness consequences of animal personalities in natural populations. The discussion makes clear that animal personality is a ubiquitous characteristic of animal populations, that personality variation is heritable and underpinned by variation in neuroendocrine and metabolic profiles, and that fluctuating selection pressures act on this variation in a wide variety of taxa. The widespread existence of personality variation implies that the study of decision making should explicitly incorporate between-individual variation.
Individual differences in behavior are a ubiquitous phenomenon in animal populations. Great tits (Parus major), for example, differ in the speed with which they explore a novel object, three-spined sticklebacks (Gasterosteus aculeatus) differ in their aggressiveness toward territorial intruders, and mice (Mus musculus) and rats (Rattus norvegicus) differ in how quickly they solve a maze problem. The existence of quantitative behavioral differences such as these, however, is hardly surprising. After all, for most quantitative traits, phenotypic differences are to be expected due to, for example, noise in the development of the phenotype or stochastic state differences among individuals (e.g., in energy reserves). At first sight, these differences do not seem to call for a deeper explanation.
Two basic observations suggest, however, that there may be more to behavioral variation than meets the eye (Gosling 2001; Sih et al. 2004; Bell et al. 2009). First, behavioral differences are often stable for some period of time. (p.218) A great tit that explores a novel object faster than a conspecific will tend to be faster also several weeks later (Verbeek et al. 1994). Similarly, the rank order in the level of aggression that individuals show toward territorial intruders tends to remain stable throughout the breeding cycle in sticklebacks (Huntingford 1976).
Second, behavioral differences often extend to a range of different situations and contexts in a systematic way. The faster a great tit approaches a novel object, for example, the faster it explores a novel environment, the more aggressively it behaves toward conspecifics, the quicker it forms rigid foraging habits, and the more willing it is to take risks (Groothuis and Carere 2005). Similarly, the more aggressively a stickleback behaves toward a conspecific territorial intruder during the breeding cycle, the more aggressive it is also toward a hetero-specific intruder and the bolder it is when approaching a predator outside the breeding season (Huntingford 1976).
Such findings are surprising. If behavioral differences were solely due to random factors, such as noise in the development of the phenotype, these differences should be uncorrelated, both through ontogeny and across different situations and contexts. The above findings, however, indicate that behavioral differences are often much more structured. Behavioral differences that are maintained through time and across contexts are termed personalities in humans (Pervin and John 1999) and, analogously, the term animal personalities (or coping styles, temperament, behavioral syndromes) has been adopted in the literature. Throughout, we will follow this usage, using the term animal personalities to refer to behavioral differences that are (a) stable through part of the ontogeny of individuals and (b) correlated across a range of situations and contexts.
Compared to the study of animal behavior, human psychology has a long history of personality research. Others have discussed this huge literature extensively (Buss and Hawley 2010); thus we will limit ourselves here to a summary of a few basic facts on human personalities.
Personality research is, almost by definition, concerned with those traits that show some stability through ontogeny, since the term personality refers to “those characteristics of individuals that describe and account for consistent patterns of feeling, thinking, and behaving” (Gosling 2001). This is not to say, however, that personality characteristics are thought to be absolutely stable (Fleeson 2004). In fact, human personality traits are known to differ in their stability and to undergo systematic changes with age, maturational events, and social-contextual transitions (Caspi et al. 2005).
What are the basic dimensions of human personality differences? The “currently most widely accepted and complete map of personality structure” (p.219) (Gosling and John 1999) is the Five Factor Model of human personalities. This descriptive model is based on the following approach: Subjects are asked to rate themselves, or people they know, according to a long questionnaire describing behavioral dispositions. These ratings are then distilled into a smaller set of variables using factor analysis. The consensus from a large number of studies that have followed this approach appears to be that behavioral variation can be described along five independent axes: extraversion, agreeableness, conscientiousness, neuroticism, and openness (Pervin and John 1999). These axes reflect distinct dispositions, each subsuming a large number of phenotypic attributes that tend to be intercorrelated. Individuals that score high on the extraversion axis, for example, tend to be talkative, assertive, active, energetic, outgoing, outspoken, dominant, forceful, enthusiastic, sociable, adventurous, noisy, and bossy. A number of studies have shown that an individual’s personality score predicts outcomes in his or her life, such as juvenile delinquency, performance in school and at work, psychopathology, and longevity (Ozer and Benet-Martinez 2006; Roberts et al. 2007). Twin studies have fairly consistently reported, for each of the five axes, broad-sense heritabilities in the range of 0.4 to 0.6 (Bouchard and Loehlin 2001), implying that roughly 50% of variation in personality is of genetic origin.
The study of personalities in nonhuman animals has a much shorter history. An early landmark was a set of studies by Nobel Laureate Ivan Pavlov, in which he identified three temperamental types in dogs (Pavlov 1928:363–364):
Finally, we have been able to distinguish several definite types of nervous systems. To one of these types, then, I take the liberty to direct your attention. This type of dog is one which judging by his behavior (especially under new circumstances) everyone would call a timid and cowardly animal. He moves cautiously, with tail tucked in, and legs half bent. When we make a sudden movement or slightly raise the voice, the animal draws back his whole body and crouches on the floor. …As I gradually analyzed the types of nervous systems of various dogs, it seemed to me that they all fitted in well with the classical description of temperaments, in particular with their extreme groups, the sanguine and the melancholic. …Between these extremes stand the variations of the balanced or equilibrated type, where both the process of excitation and the process of inhibition are of equal and adequate strength, and they interchange promptly and exactly.
Despite several other seminal contributions (Hebb 1946; Huntingford 1976; Yerkes 1939), research on animal personalities was almost nonexistent during most of the 20th century. Several explanations are offered for this in the literature. Clark and Ehlinger (1987) attribute the neglect of individual differences to the focus of early ethologists on stereotyped, species-typical behaviors and the (p.220) assumption of comparative psychologists (behaviorists) in search of general laws of learning and cognition that variation in their subjects’ responses results from uncontrolled factors in the environment. Wilson (1998a) emphasizes the historical trend in the use of natural selection to explain differences between organisms at increasingly finer scales, starting with adaptive explanations for differences between genera and higher-level taxa (1940s) to differences between closely related species (1960s) to adaptive explanations for differences among subpopulations (1980s). He argued that differences between individuals are the next in line to be considered from an adaptive perspective. Several other authors point to a rigid refusal to “anthropomorphize” (e.g., Réale et al. 2007).
Times have changed, and during the last decade animal personalities have been the subject of considerable scientific interest. A survey by Gosling (2001, 2008) counted personality studies from a wide range of vertebrate and invertebrate taxa, including mammals, fish, birds, reptiles, amphibians, arthropods, and mollusks. The research questions in these studies are diverse, addressing methodological issues (e.g., assessment of the reliability and validity of particular behavioral tests), the structure of personalities (e.g., the correlation of traits with each other and the stability of these correlations), and the causes and consequences of personalities (e.g., the physiological correlates of personalities, the effect of experience on personalities, the issues of whether personalities are favored by selection). Reviews of this huge and growing literature are provided elsewhere (Dingemanse and Réale 2012; Dingemanse and Wolf 2010; Réale et al. 2007; Wolf and Weissing 2012; Sih et al. 2012). Here, we limit ourselves to a brief introduction of two common behavioral characteristics of animal personalities which figure prominently in the literature. Thereafter we discuss the genetic and physiological correlates of animal personalities and provide an overview of selection pressures which act on animal personality traits. We hope that this discussion increases awareness that humans and other animals vary consistently in suites of correlated behavioral traits, and that such variation warrants inclusion in a Darwinian framework of decision making.
Two Common Behavioral Characteristics of Animal Personalities
The Boldness-Aggression Syndrome
The term “boldness-aggression syndrome” refers to a positive correlation between individual differences in boldness (e.g., in the response to a predator or in exploring a novel environment) and level of aggression toward conspecifics. To our knowledge, such differences were first described in three-spined sticklebacks by Huntingford (1976), who found that (a) within the breeding season, male sticklebacks differed significantly in their aggressiveness toward territorial intruders, (b) the rank order of aggressiveness remained stable both when confronted with different types of intruders throughout the breeding cycle, and (p.221) (c) individuals that were more aggressive toward territorial intruders during the breeding season tended to be bolder in response to a predator outside the breeding season.
Since Huntingford’s (1976) seminal study, the boldness-aggression syndrome has been described for a variety of other taxa (Sih et al. 2004), including birds and rodents. At present, the boldness-aggression syndrome is one of the most reported findings in the animal personality literature. Having said this, it is far from universal. For example, even within a single species, population comparisons have shown that the boldness-aggression syndrome is present in certain types of populations (those with high predation pressure) but not in others (Dingemanse et al. 2007).
Responsiveness to Environmental Stimuli
Individual differences in responsiveness to environmental stimuli are a second behavioral characteristic of animal personalities that appears to have some universality. While some individuals appear to be very responsive to all kinds of stimuli and readily adjust their behavior to the prevailing conditions, others show more rigid, routine-like behavior. Consistent differences in responsiveness (also referred to as coping style, reactivity, flexibility, and plasticity) have been documented in several taxa including birds, rodents, pigs, and humans (Dingemanse et al. 2010; Koolhaas et al. 1999). In both mice and rats, for example, individuals differ substantially in their responsiveness to environmental changes in a maze task. Some individuals quickly form a routine, are not influenced by minor environmental changes, and perform relatively badly when confronted with a changing maze configuration. Others do not form a routine, are strongly influenced by minor changes, and perform relatively well when confronted with changing maze configurations (Koolhaas et al. 1999). Similarly, some great tits readily adjust their foraging behavior to a change in the feeding situation while others stick to formerly successful habits (Verbeek et al. 1994). Individual variation in responsiveness is also increasingly studied in the wild (Dingemanse et al. 2010). In great tits, for example, individuals in the wild differ in how quickly they habituate behaviorally to repeated exposure of novel environment rooms in each of four West European populations (Dingemanse et al. 2012).
Genetic and Physiological Correlates of Animal Personalities
At the proximate level, the behavioral phenotype of an individual is shaped by its genetic, physiological, neurobiological, and cognitive systems. It is therefore not surprising that differences in personalities are often associated with differences in such systems.
The question of whether and to what extent animal personality variation has a heritable basis has received considerable attention in the recent literature (Dochtermann 2011; Réale et al. 2007; van Oers et al. 2005). The genetic basis of personality has been addressed in quite some detail using both quantitative and molecular genetics tools in a wide diversity of taxa, including birds (Drent et al. 2003), fish (Dingemanse et al. 2009), and primates (Weiss et al. 2000). Both approaches suggest that animal personalities are heritable but that their genetic underpinning might be complex.
Animal personality variation, addressed from a quantitative genetics perspective, has primarily focused on estimating how much of the variation in behavior can be explained by additive effects of genes. Such studies have largely confirmed that (a) behavioral traits are moderately repeatable (i.e., they differ consistently between individuals, and repeatability values of behavior are roughly 20–40% according to a recent meta-analysis; Bell et al. 2009) and (b) behavioral traits are also moderately heritable (i.e., the consistent differences between individuals have a heritable basis; narrow-sense heritability estimates range roughly between 10–40% for behavioral traits; Réale et al. 2007; van Oers et al. 2005). These estimates of behavioral heritability are based either on classic artificial selection experiments or variation in relatedness within pedigreed natural populations.
Quantitative genetics studies have given rise to a number of further key insights concerning the genetic underpinning of personality. First, genetic correlations between the same behavior expressed in different contexts (e.g., activity in the presence vs. absence of predators, aggressiveness in juvenile vs. adult life-history phases, or exploratory behavior of novel environments vs. objects) are often extremely tightly correlated; genetic correlation values are close to 1 (Dingemanse et al. 2009). In other words, it appears that the same genes are often expressed when the same behavior is expressed in different contexts, giving rise to the cross-context correlations that characterize animal personalities. Second, genetic correlations between functionally distinct behaviors are, in contrast, relatively labile, as exemplified by work on stickleback which shows that both phenotypic and genetic correlations between components of the boldness-aggression syndrome vary between populations on relatively small spatial scales (Bell 2005). In other words, genetic correlations between different types of behavior might typically occur not because of genes with pleiotropic effects1 but rather because of linkage disequilibrium2 induced by natural selection (Dochtermann 2011).
Molecular geneticists studying animal personalities have focused primarily on candidate genes—genes known to affect personality in humans or husbandry animals (Munafò et al. 2008). In recent years, much research has centered on a single, though prominent, human personality gene: the dopamine receptor gene DRD4. This gene is part of the dopaminergic system that affects the motivation of behavior and influences human novelty seeking. Recent work shows that genetic variants (i.e., single nucleotide polymorphisms or SNPs)3 exist for this gene and predict exploration behavior in animals of a wide range of taxa (Munafò et al. 2008). For example, SNP830 in the DRD4 gene predicts exploration of novel environments in a Dutch nest box population of great tits (Fidler et al. 2007). However, the same polymorphism in the DRD4 gene does not predict exploration behavior in three other nest box populations of this species (Korsten et al. 2010). This implies that links between genes and personality may be relatively complex and influenced, for example, either by gene–gene or gene–environment interactions. Such a complex genetic underpinning has recently been shown experimentally in stickleback, where the expression of heritable variation in personality traits is influenced by perceived predation risk during ontogeny (Dingemanse et al. 2009).
Evidence is accumulating to indicate that personality differences are often systematically associated with the physiological setup of individuals.
Much research in this area has focused on the notion that individual animals might vary consistently in their physiological responsiveness to mild stressors, indicated by the term “coping style” (Koolhaas et al. 1999). “Proactive copers” are characterized by a relatively (a) low hypothalamic pituitary adrenal (HPA) axis reactivity, (b) high sympathetic reactivity, and (c) low parasympathetic reactivity. In contrast, “reactive copers” are characterized by a relatively (a) high HPA axis reactivity, (b) low sympathetic reactivity, and (c) high parasympathetic reactivity (Koolhaas et al. 1999). Over the last few years, such variation in physiological setup has been reported for a wide range of taxa, including birds, domestic pigs, rodents, fish, and primates (Carere et al. 2010).
Variation in stress physiology is systematically associated with the boldness-aggression syndrome and individual differences in responsiveness (Koolhaas et al. 1999; Koolhaas et al. 2010). Proactive copers—whose behavior is thought to be driven by internal mechanisms—are typically active, aggressive, and bold, and relatively unresponsive to environment stimuli. (p.224) Reactive copers, on the other hand, behave in a nonaggressive and cautious manner and are much more responsive to environmental stimuli (Koolhaas et al. 1999). Detailed experimental work in behavioral stress physiology suggests that variation in physiological and behavioral stress responsiveness is underpinned by two independent personality axes: one represents the quality of the response to challenging situations (coping style), the other the quantity of that response (stress reactivity) (Koolhaas et al. 2010). Finally, there is growing awareness that behavior and neuroendocrinology interact dynamically, and that one (endocrinology) should not necessarily be regarded as the proximate underpinning of the other (behavior) (Koolhaas et al. 2010).
Other recent research on the proximate underpinning of animal personality has focused on whether individuals differ consistently in their energy metabolism (e.g., resting metabolic rate) and whether such differences are systematically associated with personality differences. A recent meta-analysis shows that this is indeed the case (Biro and Stamps 2010). Individuals within several animal species differ consistently in resting metabolic rate, and these differences are associated with differences in activity levels, levels of aggression, and dominance status. Typically, active, aggressive, or dominant individuals have higher values of resting metabolic rate. These links appear plausible at first sight: highly aggressive and active individuals need the physiological machinery to fuel such costly activities. Phrased differently, individuals that have high metabolic rates might need to behave in such a way that they acquire resources needed to meet their daily energy requirements.
Natural Selection and Animal Personality
Why has natural selection not given rise to a single “optimal” behavioral type but rather to a mixture of behavioral types? From a theoretical perspective, several different lines of explanations have been proposed (Dingemanse and Wolf 2010; Wolf and Weissing 2010). Empiricists, however, have focused primarily on spatiotemporal variation in selection pressures as an explanation for coexistence.
Spatiotemporal Variation in Selection Pressures
Different behavioral types may be favored in different local environments (spatial variation) or at different times (temporal variation) and such variation in selection pressure may give rise to the coexistence of behavioral types. A recent review reported ten (out of eleven) studies reporting some form of spatiotemporal variation in selection pressures acting on behavior within the same population (Dingemanse et al. 2012). These studies comprised a large diversity of taxa (birds, insects, mammals, reptiles), implying that this form (p.225) of selection may be widespread in nature. Temporal variation, for example, has been documented in birds (great tits), mammals (bighorn sheep, Ovis canadensis, and North American red squirrels, Tamiasciurus hudsonicus). In red squirrels, aggressive females were favored in certain (but not in other) years, based on growth and survival of their offspring. Spatial variation in selection pressures has been documented, for example, in birds (great tits), where fast explorers had relatively high breeding success in high density areas but slow explorers in low density ones.
Missing Pieces of Information
Despite great interest in the fitness consequences of animal personality, two key questions have been largely neglected in the empirical literature (Dingemanse et al. 2012). First, how does natural selection act on behavioral consistency? Do consistent individuals enjoy higher fitness than individuals that are less inconsistent in their behavior? Second, how does natural selection act on behavioral correlations? Do bold animals, for example, have higher fitness when they are also relatively aggressive? Addressing such questions would enable us to better understand why individuals differ consistently in suites of correlated traits, and whether such differences are perhaps favored by natural selection.
Over the past few years, researchers have started to expand their horizons and restrict their research not solely to understanding the proximate and evolutionary mechanisms maintaining personality variation. Specifically, there is growing interest in the question of whether personality variation, in itself, has ecological and evolutionary consequences. A recent modeling study (Wolf and Weissing 2012; Sih et al. 2012), for example, shows that populations consisting of a mix of social and asocial individuals spread faster than populations consisting of either one (Fogarty et al. 2011).
Animal personality refers to individual differences in behavior that are stable over time and correlated across different contexts. Such differences can be found in a wide range of taxa across the animal kingdom. Here we have introduced the empirical research on animal personalities. At present, there seem to be two common (i.e., to some extent universal) behavioral characteristics of animal personalities: the boldness-aggression syndrome and individual differences in responsiveness to environmental stimuli. Work on the proximate underpinning of animal personality shows that animal personality is heritable, that the expression of the same behavior across different contexts is often influenced by the same genes, and that genetic correlations between different behaviors (e.g., aggression and boldness) may result from linkage disequilibrium rather than gene pleiotropy. Animal personality is often associated with (p.226) differences in stress physiology and metabolism, though behavior and stress physiology might interact dynamically and cause-effect relationships might therefore be complex. Selection on the behavioral phenotype often fluctuates in space and time, while selection pressures favoring personality per se (i.e., consistency, correlations between behaviors) await quantification. Finally, the ecological and evolutionary consequences of such variation await further study.
Both authors contributed in equal part to this manuscript. The introductory sections of this paper are modified versions from the introductory sections of M.W.’s Ph.D. thesis. N.J.D. was supported by the Max Planck Foundation.
(1) Pleiotropic effects occur when one gene influences multiple phenotyphic traits.
(2) In population genetics, linkage disequilibrium is the nonrandom association of alleles at two or more loci.
(3) Single nucleotide polymorphisms reflect DNA sequence variation and occur when a single nucleotide differs either between chromosomes of the same individual or different individuals.