Personality Neuroscience and the Biology of Traits

Personality neuroscience involves the use of neuroscience methods to study individual differences in behavior, motivation, emotion, and cognition. Personality psychology has contributed much to identifying the important dimensions of personality, but relatively little to understanding the biological sources of those dimensions. However, the rapidly expanding field of personality neuroscience is increasingly shedding light on this topic. This article provides a survey of progress in the use of neuroscience to study personality traits, organized using a hierarchical model of traits based on the Big Five dimensions: Extraversion, Neuroticism, Agreeableness, Conscientiousness, and Openness ⁄ Intellect. Evidence is reviewed for hypotheses about the biological systems involved in each trait. The mission of personality psychology is ‘‘to provide an integrative framework for understanding the whole person’’ (McAdams & Pals, 2006; p. 204), and many different methods must be brought to bear to accomplish such an ambitious goal. The field of personality neuroscience rests on the premise that the whole person cannot be understood without understanding the brain. In this article, I discuss the role that neuroscience can play in personality research and review the progress of this rapidly expanding field. (For more in depth review of the field and its influential theories see DeYoung & Gray, 2009; and Zuckerman, 2005.) Personality psychology’s attempt to understand the whole person calls for a broad conception of personality itself, such as the one provided by McAdams and Pals (2006, p. 212): Personality is an individual’s unique variation on the general evolutionary design for human nature, expressed as a developing pattern of dispositional traits, characteristic adaptations, and integrative life stories, complexly and differentially situated in culture. This definition describes three levels at which personality can be analyzed: traits, characteristic adaptations, and life stories. Personality neuroscience has focused primarily on traits, which are relatively stable patterns of behavior, motivation, emotion, and cognition (Pytlik Zillig, Hemenover, & Dienstbier, 2002; Wilt & Revelle, 2009) that are not specific to a particular social milieu or culture. This is not to say that traits are evident to the same extent or with identical manifestations in all cultures, but rather that any trait can be observed in a subset of situations in any culture. In contrast to traits, characteristic adaptations and life stories describe the individual’s specific responses to his or her particular life circumstances. Obviously, the latter two levels of analysis are crucial for understanding any individual, but their complexity renders them less amenable to study by neuroscience, and this article focuses on the biology of traits. Traits can be considered probabilistic descriptions of the frequency and intensity with which individuals exhibit various behavioral, motivational, emotional, and cognitive states Social and Personality Psychology Compass 4/12 (2010): 1165–1180, 10.1111/j.1751-9004.2010.00327.x a 2010 The Author Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd (Fleeson, 2001; Fleeson & Gallagher, 2009). Individuals who are high in some trait will experience the states associated with that trait more often and more intensely than individuals low in that trait. For example, someone high in Extraversion will be talkative, outgoing, and excited more often than someone low in Extraversion, but even the person low in Extraversion may experience those states occasionally. The aim of personality neuroscience is to understand both the biological systems that are responsible for the states associated with traits and the parameters of those systems that cause them to function differently in different individuals. The systems themselves are presumed to be present in every intact human brain – hence McAdams and Pals’ (2006) reference to ‘‘the general evolutionary design for human nature’’ – but their parameters will vary from person to person. (For example, all people have brain systems that respond to rewards, but in different individuals these systems will respond with different degrees of vigor to a particular reward, and the systems’ average level of response may be associated with some personality trait.) When considering the biological sources of personality, one must distinguish between proximal and distal sources. The proximal sources, just described, consist of stable differences in the functioning of the neural systems that produce the states associated with traits. The distal sources are both genetic and environmental, as indicated by the fact that heritability estimates for personality traits are in the range of 40% to 80%, depending on trait and method (Bouchard & McGue, 2003; Riemann, Angleitner, & Strelau, 1997). (The heritability of a trait indicates the amount of its variation in a population due to genetic, rather than environmental, variation.) It is important to remember that when either genes or situations have lasting effects on traits, they must do so by changing the brain; thus, personality differences are ‘biological’ regardless of their heritability, in the sense that they must be proximally generated by the brain no matter whether they originated in genes or environment. Methods in Personality Neuroscience Scientific investigation of the biological basis of personality has been limited, until relatively recently, by a lack of technology to examine brain structure and function in living human beings. Prior to the development of neuroimaging, the sole means for assessing brain activity was the electroencephalogram (EEG), which measures the brain’s electrical activity at the scalp. Today a number of methods are available for personality neuroscience. Five important categories of neuroscientific methods are (1) neuroimaging (e.g., magnetic resonance imaging [MRI] or positron emission tomography [PET]), which allows assessment of brain structure and function with a relatively high spatial resolution; (2) molecular genetics, which allows assessment of variation in specific genes that are expressed in the brain; (3) EEG, which provides the highest temporal resolution of neural activity of any available method; (4) assays of endogenous psychoactive substances or their byproducts (e.g., hormone levels in saliva or neurotransmitter metabolites in cerebrospinal fluid); and (5) psychopharmocological manipulation (e.g., tryptophan depletion or augmentation to alter levels of the neurotransmitter serotonin). Personality neuroscience employs these methods in conjunction with the methods of personality psychology, attempting to link biological variables to traits. Measurement of traits is typically accomplished by questionnaire, through self-report and ⁄or ratings by peers or other knowledgeable informants. Questionnaires provide a convenient and reliable method for assessing a broad range of stable individual differences, drawing on raters’ experiences over a far greater span of time than is available in the laboratory. However, 1166 Personality Neuroscience a 2010 The Author Social and Personality Psychology Compass 4/12 (2010): 1165–1180, 10.1111/j.1751-9004.2010.00327.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd traits can also be assessed in other ways, such as through laboratory tasks, behavioral observation, or experience sampling, and personality should not be identified exclusively with the variables provided by personality questionnaires. What we want to explain in personality neuroscience is not how people answer questionnaires, but rather why they exhibit stable patterns of behavior, motivation, emotion, and cognition. Most studies discussed in this review begin with some psychological trait and attempt to discover its biological basis. Another valid and useful approach is to begin with stable individual differences in some biological parameter (such as asymmetry in the level of left versus right cortical hemisphere activity, measured at rest by EEG) and then to attempt to discover what psychological trait or traits are associated with that biological trait. An important caveat for anyone exploring the literature on personality neuroscience or contemplating entering the field as an investigator is that much inconsistency exists in the findings to date. In neuroimaging research, inconsistency is probably due in large part to the use very small samples, due to cost. Unfortunately, small samples often lack statistical power to detect true effects (Yarkoni, 2009). Worse still, low power has the unfortunate additional consequence of increasing the proportion of false positives among significant findings (Green et al., 2008). In genetic studies, though larger samples are typically used, power may still be an issue, and inconsistency may arise, because each trait is influenced by many genes, each accounting for only a very small amount of the variance of the trait (Green et al., 2008). These difficulties highlight the importance of testing reasonably focused hypotheses, rather than simply exploring associations with biological variables in the absence of any theory of the causal mechanisms that might underlie a given trait. The Structure of Personality Traits Personality neuroscience can usefully be guided by existing knowledge about the structure of personality – that is, knowledge about how various traits relate to one another and to the major dimensions of personality. In order to produce a coherent overview of the progress of personality neuroscience, one needs to relate findings to a reasonably comprehensive taxonomy of traits, such as that provided by the Five Factor Model or Big Five, which categorizes the majority of traits within five broad domains, typically labeled Extraversion, Neuroticism, Agreeableness, Conscientiousness, and Openness ⁄ Intellect (Digman, 1990; John, Naumann, & Soto, 2008). The Big Five model is the result of decades of work using factor analysis to map the patterns of covariation among traits. In order to be accurate, such factor analyses require a reasonably comprehensive and unbiased pool of traits to analyze. The lexical hypothesis states that natural language (as represented in dictionaries) provides just such a pool of trait descriptors (John et al., 2008; Saucier, 2009). Existing personality questionnaires constitute another source of a large and broad pool of traits in which to locate general factors. Lexical and questionnaire research have both provided evidence for the Big Five (Digman, 1990; John et al., 2008; Markon, Krueger, & Watson, 2005). The suitability of the Big Five model for personality neuroscience may not be immediately obvious because it was developed as a purely descriptive model, remaining silent on why those particular five dimensions describe the major axes of covariation among personality traits. However, explanatory models for the Big Five should be possible because a key premise of the factor-analytic approach is that multiple traits covary, indicating a factor, due to some shared underlying cause (Haig, 2005). This cause need not be exclusively biological (an alternative cause, for example, might be a frequently occurring class of situations that evokes behaviors related to multiple traits with distinct biological Personality Neuroscience 1167 a 2010 The Author Social and Personality Psychology Compass 4/12 (2010): 1165–1180, 10.1111/j.1751-9004.2010.00327.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd causes), but shared biological causation is at least a reasonable hypothesis. Individual differences in the Big Five are strongly genetically influenced (Bouchard & McGue, 2003; Riemann et al., 1997), and the genetic factor structure of the Big Five appears to be invariant across European, North American, and East Asian samples (Yamagata et al., 2006). We have good reason, therefore, to look for biological systems underlying the Big Five. However, the Big Five do not constitute the only level of personality traits of interest. Traits can be arranged hierarchically, with correlated groups of more specific traits categorized together in broader traits (Markon, 2009). The existence of traits both narrower and broader than the Big Five may indicate additional biological causes both more and less specific than those that influence the Big Five. At least one level of structure exists above the Big Five. Although the Big Five were originally conceived as independent traits and the most general level of personality description, research has shown that they are regularly intercorrelated and possess a higher-order factor structure (DeYoung, 2006; Digman, 1997; McCrae et al., 2008). The shared variance of Neuroticism (reversed), Agreeableness, and Conscientiousness constitutes one higher-order factor or metatrait, labeled a or Stability, and the shared variance of Extraversion and Openness ⁄ Intellect constitutes another, labeled b or Plasticity. Additionally, multiple levels of the trait hierarchy exist below the Big Five. Each Big Five domain comprises the shared variance of a large number of lower-level traits, often called facets, with no consensus as to how many facets exist for each domain. Each facet has been shown to have a unique genetic contribution, suggesting the existence of specific biological parameters that differentiate facets within a single domain (Jang, McCrae, Angleitner, Riemann, & Livesley, 1998). Finally, another level of personality structure appears to exist between the Big Five and their facets. Jang et al. (2002) found that two genetic factors were necessary to account for the shared genetic variance among facets within each of the Big Five. If the Big Five were the next level above the facets, only one genetic factor should have been necessary for each domain. These mid-level factors were characterized in more detail through factor analysis of many facets for each domain, followed by examination of how the resulting factors were correlated with over two thousand personality items (DeYoung, Quilty, & Peterson, 2007). This analysis indicated that each of the Big Five comprises two separable but correlated aspects, which constitute a level of personality structure between the facets and the Big Five. The four levels of the personality hierarchy just described are depicted in Figure 1. (Remember that labels for personality traits are always imperfect attempts to capture what is shared among the various characteristics encompassed by each trait.) Each level can be considered to reflect causal influences of differing breadth. The existence of the bottom level indicates that some causes are unique to each facet. At the second level, causes are shared between facets within each aspect but are not shared with facets in the complementary aspect in each pair. At the third level, all traits within each of the Big Five share causes that are not shared with traits in other Big Five domains. Finally, the existence of the metatraits implies that traits share causal influences even across some of the Big Five domains. At least one caveat must be made regarding the model in Figure 1. It is necessarily an oversimplification because personality does not have simple structure, which means that some facets and aspects show correlations, not modeled by the hierarchy in Figure 1, with facets and aspects in other Big Five domains. As one biologically pertinent example, Assertiveness is negatively correlated with Politeness (which has belligerence at its low pole), despite the fact that these traits are subsumed by two Big Five traits that are unre1168 Personality Neuroscience a 2010 The Author Social and Personality Psychology Compass 4/12 (2010): 1165–1180, 10.1111/j.1751-9004.2010.00327.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd lated (DeYoung et al., 2007). This association is likely to be of interest for personality neuroscience because these two traits appear to be influenced in opposite directions by testosterone (Luxen & Buunk, 2005; McIntyre et al., 2007; Netter, 2004), which may account for their covariation. To include a trait reflecting this putative causal influence of testosterone in Figure 1, we would need to add an additional variable outside of the existing hierarchy (perhaps labeled ‘Competitiveness’ or ‘Dominance’), with arrows directed at Assertiveness and Politeness. To account for the full complexity of personality structure, many such additional variables are likely to be necessary. Nonetheless, the model in Figure 1 can effectively be used to organize most of what is currently known about the neurobiology of traits. Neurobiology of the Trait Hierarchy A crucial step in the investigation of the biology of personality traits is identification of the psychological functions shared by all of the lower-level traits within each trait of interest. Because much is known about how the brain enacts a wide variety of psychological functions, development of a hypothesis about the psychological functions responsible for a personality trait provides the necessary stepping stone to a hypothesis about the brain systems involved in that trait. Personality psychologists have produced many theories about the psychological functions underlying various traits, but only a few have attempted to identify the psychological functions underlying all of the Big Five (Denissen & Penke, 2008; DeYoung, 2010; DeYoung & Gray, 2009; Nettle, 2006, 2007; Van Egeren, 2009). Such theories offer psychological explanations for the Big Five that go beyond mere description. Considerable agreement is evident among these theories about the functions associated with the Big Five. Extraversion and Neuroticism are thought to be the primary manifestations in personality of sensitivity to reward and positive affect (Extraversion) and sensitivity to punishment and negative affect (Neuroticism). Agreeableness reflects the tendency toward altruism as opposed to exploitation of others. Conscientiousness reflects top-down control of behavior and impulses in order to follow rules and pursue non-immediate goals. And Openness ⁄ Intellect reflects the tendency to detect, explore, appreciate, and utilize patterns of abstract and sensory information. Armed with these hypotheses, one can begin to develop and evaluate a model specifying which brain systems are likely to be involved in which traits. Stability Plasticity Metatraits: