Research

Research in my lab is focused on understanding why individual animals behave differently from each other. In other words, why do individuals of the same species have different personalities? Although we tend to think that 'personality' is confined to humans and perhaps nonhuman primates, there is growing evidence for something akin to personality in nonhuman animals as well. Even an individual fish, for example, behaves differently from other fish, through time and across situations.

Personality in animals raises at least two interesting evolutionary questions. First, individual consistency in behavior is mysterious because if individuals have a tendency to behave a certain way generally, then this tendency might carry over to influence behavior in other situations. In such a case, individuals might not be able to switch to the optimal behavior in a different set of circumstances (Sih et al. 2004). For instance, if an aggressive male bird that fights with his neighbors over territory boundaries can't 'turn off' the release of testosterone fast enough after a fight, then testosterone might carry over to influence the male's behavior after the fight, perhaps even during parental care, when nonaggressive behavior is favored.

Second, individual differences in behavior are difficult to explain from an evolutionary point of view because natural selection erodes heritable variation among individuals that is related to fitness. For example, if bold individuals are more likely to be consumed by predators, and if there is a heritable basis to boldness, then over many generations, the entire population of individuals will be timid. So how can we explain the persistence of bold individuals within a population that is under predation pressure?

In my lab, we study the proximate mechanisms underlying personality and the ultimate (evolutionary) consequences of personality using threespined stickleback fish as a model system. In addition to their long history as a favorite study animal of the classical ethologists, sticklebacks are an emerging model organism with growing genomic resources, including a full genome sequence. Sticklebacks are particularly great subjects for studying evolution because they have undergone an adaptive radiation and exhibit tremendous phenotypic variation among populations. Independent populations rapidly evolved convergent phenotypes in response to similar selective pressures, such that the same traits arose repeatedly and independently from the same ancestor. Research in the lab is currently focused on answering the following questions using sticklebacks as a model organism:

1) What is the genetic basis underlying personality traits?

One of the primary behavioral axes along which individual animals, including sticklebacks, vary is their propensity to take risks. Some individuals are relatively risk-averse, in that they hide or freeze in the presence of a predator and are relatively unaggressive toward other sticklebacks. Other individuals are generally more risk-prone in dangerous situations, in that they are willing to forage when predators are present, or to engage in dangerous fights with conspecifics.

One hypothesis to explain behavioral consistency is that individuals behave in a consistent way because of an inherited tendency to do so. Previous studies have shown that there is a heritable basis to risk-taking behaviors in sticklebacks, but the identity of the genes underlying such behaviors is unknown. One tactic for identifying the genes responsible for a particular phenotype is to perform controlled laboratory crosses between divergent phenotypes, and to relate phenotypic variation to genetic variation using quantitative trait locus (QTL) mapping. Together with Katie Peichel, we initiated a QTL project to map risk-taking behaviors by crossing sticklebacks from a population that was generally risk-prone toward predators and conspecifics with sticklebacks from a population that was generally risk-averse, and the work is ongoing. Based on the results of this experiment, we will perform another cross between a different pair of populations to determine whether the same QTL (and, ultimately genes) are related to risk-taking behaviors in independent populations. Lindsay Marquardt is a graduate student working on this project.
            As mentioned above, some behavioral variation in sticklebacks is inherited and genetically-based, but these behaviors are also influenced by experience. Indeed, a resounding message from decades of research on personality in humans and nonhuman animals is that although there is often a heritable component to personality, there is also important room for environmental effects -- personality changes with age, and is influenced by life events. Some of the most intriguing work on the genetic basis for personality variation in humans has found that whether or not a particular gene is associated with personality depends on the environment that their subjects experienced (GxE interactions). Altogether, the work on humans suggests that our attempts to find genes associated with personality are most likely to be successful if we simultaneously consider both genetic and environmental effects.

Therefore we are using new genomic resources for this species to identify both inherited and environmentally-responsive genes associated with risk-taking behaviors. In a project funded by an R01 grant from the National Institutes of Health, we are using expression microarrays to identify genes associated with risk-taking behaviors. Osee Sanogo is a postdoc working on this project.

2) Does personality result from selection or plasticity?
            In previous work, I found that behavioral consistency is favored when predation pressure is high. Specifically, whereas sticklebacks from a relatively 'safe' population, where predators are rare, do not behave consistently through time or across contexts, sticklebacks from a population under strong predation pressure do behave consistently (Bell 2005, see also Dingemanse et al 2007). We identified predation as the specific selective factor favoring the evolution of personality in sticklebacks by applying real predation pressure by trout to sticklebacks from a 'safe' population (which did not behave consistently) and found that behavioral consistency emerged among the survivors (Bell and Sih 2007).

An unanswered question that was prompted by this experiment is whether predation caused behavioral consistency via nonrandom mortality of certain behavioral types, or via phenotypic plasticity, or both. That is, perhaps trout selectively ate individuals that did not behave consistently. Alternatively, individuals might have become more consistent after exposure to predation risk. Together with Niels Dingemanse, we are testing the relative importance of each of these two mechanisms by assessing whether exposure to predator cues (but not actual predation) is sufficient to generate behavioral consistency.

3) Do parents influence the personality of their offspring via non-genetic parental effects?
            In sticklebacks, as in many fishes, the father provides all the parental care toward the developing offspring. We have noticed that there is tremendous variation among males in how they behave as parents -- some males are highly responsive toward intruders and predators during the parental period, whereas other males are consistently less aggressive toward intruders, but possibly more 'attentive' toward their offspring. There is also some published data (Tulley and Huntingford 1987, Animal Behaviour, 35, 1570-1572) indicating that how males behave toward the developing fry influences the risk-taking behaviors of the offspring toward predators later in life.

Even though males exclusively provide parental care after females lay eggs, there is the intriguing possibility that in utero conditions in the mother might also influence the risk-taking behaviors of her offspring. Preliminary data from Eric Giesing, a graduate student in the lab, show that eggs from mechanically stressed females contain higher concentrations of cortisol, a stress hormone, than eggs from nonstressed females, a result that is consistent with an emerging literature in birds showing that females deposit hormones into the yolks of developing eggs, and that early exposure to hormones shapes the behavioral development of the chicks. Therefore we are investigating the interactions between maternal and paternal effects on risk-taking behaviors, and compare the interaction among different populations that vary in the severity and frequency of stressors, e.g. predators, in the environment (Bell et al 2007). This work could set the stage for studying some very interesting interactions between males and females during mate choice. For example, the type of male that would be the optimal parent for a female's offspring might depend on the female's recent exposure to stressors.

4) Measuring stickleback personality in the field.

Although we know a lot about individual differences in behavior in the lab, we do not know whether the behavioral differences we observe between individual stickleback in the lab reflect natural variation in behavior in the field. Therefore, Kate Laskowski, a PhD student in the lab, is validating the behavioral assays used in our lab experiments by carrying out similar tests in the field, and then re-measuring the same individuals in the lab.

Finally, an understanding of the brain of this fascinating organism has lagged far behind our understanding of its behavior and genome. Therefore, Molly Kent, a Neuroscience graduate student in the lab, has initiated a project to develop a brain atlas for sticklebacks. The atlas will lay the groundwork for subsequent neuroanatomical and neurofunctional investigations of the neural mechanisms that are the link between genomic variation and behavioral variation. We will make this resource widely available to the growing community of biologists studying this emerging model organism.

In conclusion, consistent individual differences in behavior have been observed by students on animal behavior for many years, but have generally been considered to be 'noise' around an optimal mean. However, there is accumulating evidence that individual differences in behavior are often heritable and related to fitness. My general goal is to understand both from a proximate and ultimate perspective why personality exists in the first place.

Department of Animal Biology
Institute for Genomic Biology
University of Illinois

Updated 02/20/2009