Personality Neuroscience: Explaining Individual Differences in Affect, Behavior, and Cognition

Text-only Preview

Personality Neuroscience 1

Personality Neuroscience: Explaining Individual Differences in Affect, Behavior, and Cognition

Colin G. DeYoung Jeremy R. Gray
Department of Psychology, Yale University

To be published in the Cambridge University Press Handbook of Personality
Please to not cite or circulate without permission

Email: [email protected]

Personality Neuroscience 2
Personality Neuroscience: Explaining Individual Differences in Affect, Behavior, and Cognition

Human behaviors and experiences are generated by biological processes, primarily within
the brain. On this basis, we may assume that the regularities in these behaviors and experiences
that constitute personality are associated with regularities in the biological functions of the brain,
making personality neuroscience possible. It is increasingly easy to study psychologically
relevant individual differences using neuroscientific methods. Personality neuroscience
endeavors to understand the proximal sources of personality in the brain and to trace those brain
processes back to their distal sources in complex interactions between genes and environment.
Heritability estimates for personality traits are typically around 50% or higher, indicating that the
distal sources of personality lie in both the genome and the environment (Bouchard, 1994;
Loehlin, 1992; Reimann, Angleitner, & Strelau, 1997). Both genes and environment must make
their mark on the brain, however, if they are to have a lasting influence on personality.

Personality psychology attempts to answer some of the most fundamental questions about
people: Why are individuals the way they are? How and why do people differ from each other?
For much of the past century, personality psychology has been concerned more with describing
personality than with explaining it – that is, with how people differ from each other rather than
with why they differ from each other. One reason for this emphasis on description rather than
explanation was the immaturity of human neuroscience. Tools for investigating the
neurobiological underpinnings of individual differences have greatly increased in power with the
rise of neuroimaging and molecular genetics.

Another reason that personality psychology tended to focus on description rather than
explanation was the necessity of developing an adequate categorization system for personality

Personality Neuroscience 3
traits. Traits are relatively stable patterns of affect, behavior, and cognition (Fleeson, 2001;
Pytlik Zillig, Hemenover, & Dienstbier, 2002). The existence of a multitude of such patterns
raises the question of whether some relatively small number of broad trait categories can be used
to classify and organize the majority of traits. Over the past 25 years, this question has been
largely resolved by the emergence of the Five Factor Model or Big Five, which postulates that
almost all trait descriptors can be categorized within five broad domains (or as blends of two or
more of those domains): Extraversion, Neuroticism, Agreeableness, Conscientiousness, and
Openness/Intellect (Costa & McRae, 1992; Digman, 1990; Hofstee, de Raad, Goldberg, 1992;
John & Srivastava, 1999; chapter 10, this volume). Our emphasis on the Big Five should not be
misconstrued as a suggestion that the trait level of analysis alone can capture everything about
personality—it cannot. Characteristic adaptations to particular situations and self-defining life
narratives constitute two additional major levels of analysis that are more specific and better able
to capture the rich detail and uniqueness of individual personalities (McAdams & Pals, 2006;
Mischel & Shoda, 1998; Wood & Roberts, 2006). Incorporating these levels can complement
and enrich the trait approach. At this early stage in personality neuroscience, however, linking
traits to neurobiological substrates is a promising start.

The Big Five model offers a useful categorization scheme for personality neuroscience
and can usefully organize a review of this young field. Despite some debate in lexical research
on personality about whether a six- or seven-factor model might be more robust across languages
(Ashton et al., 2004; Saucier & Goldberg, 2001), the Big Five is the most widely used taxonomy
of personality and provides a useful common language for personality research, helping to ensure
that results are comparable across different studies. Behavior genetic research shows that the Big
Five are substantially heritable, with estimates ranging from 40% or 50% up to 80%, depending

Personality Neuroscience 4
on trait and method (Reimann et al., 1997). Recently, the genetic factor structure of the Big Five,
as measured by the Revised NEO Personality Inventory, was shown to be invariant across
European, North American, and East Asian samples, suggesting the biological universality of
these traits (Yamagata et al., 2006). When measures of abnormal and normal personality traits
are factor analyzed together, the standard Big Five solution appears (Markon, Krueger, and
Watson, 2005), suggesting the utility of the Big Five for studying psychopathology. Finally, the
Big Five appears to be an effective taxonomy of descriptors of individual differences in other
species (Gosling & John, 1999), and cross-species comparisons are often important in
neuroscience. All of these considerations suggest the potential utility of the Big Five for
personality neuroscience, as long as models can be developed that identify possible biological
substrates of the Big Five and lead to testable hypotheses. Fortunately, the field has now
developed to the point where a synthesis of the literature may contribute to this sort of model.

Our aims in this chapter are first to review the methods and history of personality
neuroscience and then to attempt a synthesis of findings across the range of personality traits, as
represented primarily by the Big Five. The goal here is not to summarize every result for any
particular method, but to determine how the array of findings to date contributes to a larger
picture of the relation between variation in the brain and variation in personality. This strategy
will highlight a point that we consider to be of major conceptual importance, namely that
theories of personality must not be limited to a particular domain of information processing, but
must consider individual differences in affect, behavior, and cognition, as well as how these
different domains are integrated and interact (Gray, 2004; Gray, Braver, & Raichle, 2002). Only
by taking a broad view will the field be able to fulfill the promise of personality psychology to
understand the individual as he or she actually functions as a whole.

Personality Neuroscience 5
Methods in Personality Neuroscience

Personality neuroscience reflects the conjunction of personality psychology with methods
for discovering the biological sources of individual differences. The measurement of personality
is typically accomplished by questionnaire, through self-report and/or through ratings by peers or
other knowledgable informants. “Personality” is sometimes taken to mean the set of variables
that result from questionnaire measures, but this confuses the instruments with the constructs.
Questionnaires are simply a convenient and reliable method for assessing a broad range of stable
individual differences, drawing on subjects’ experiences over a far greater span of time than is
available in the laboratory. Other methods can be used to measure personality, if they can be
validated psychometrically (i.e., as measuring stable patterns of affect, behavior, and cognition).
Measures of intelligence and of the ability to delay gratification (Ohmura, Takahashi, Kitamura,
& Wehr, 2006) are good examples. The challenge with such behavioral measures, because
models of personality structure like the Big Five are typically operationalized with
questionnaires, is to relate them back to those models so that they may be incorporated into the
larger nomological network of personality psychology.

The general categories of neuroscientific methods that we see as currently most relevant
to personality are (1) neuroimaging (e.g., magnetic resonance imaging (MRI) or proton emission
tomography (PET)), (2) molecular genetics (a.k.a. genomics), (3) electrophysiological techniques
(e.g., electroencepholography (EEG) or measurement of electrodermal activity (EDA)), (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 serotonin).

Personality Neuroscience 6

Studies of personality using structural and functional MRI and PET are appearing at a
rapid rate; over 50 have been published since 2003, almost tripling the existing number (see
chapter 16, this volume). Molecular genetics has seen a similar explosion of personality research
since the first studies of the effects of genetic variation on normal personality traits appeared in
1996 (see chapter 13, this volume). Electrophysiological research was the primary tool for
investigating the biology of personality, prior to the advent of neuroimaging and molecular
genetics; with a few exceptions, however, associations of electrophysiological variables with
personality have been inconsistent (Zuckerman, 2005). Excellent reviews of research on the
influence of neurotransmitters and hormones on personality have been written by Netter (2004),
Hennig (2004), and Zuckerman (2005). Many inferences about personality can be drawn from
the study of nonhuman animals (see chapter 14, this volume), and consistency with nonhuman
analogs is a hallmark of good theory in personality neuroscience, but we limit our review to
human methods.
Influential Theories in Personality Neuroscience

We briefly describe the influential models of Eysenck, Gray, Zuckerman, Cloninger,
Depue, and Panksepp and relate each model to the Big Five, with the aim of translating results
from different systems into a single common language. This approach is readily justified by the
fact that these models have been demonstrated to fall within the same factor structure as the Big
Five (e.g., Markon et al., 2005; Angleitner, Riemann, & Spinath, 2004). Most of the theorists
have revised their models substantially over time; in the interest of space we discuss only the
latest version of each.

Eysenck assigned traits to three “superfactors,” Extraversion, Neuroticism, and
Psychoticism (Eysenck & Eysenck, 1985). Extraversion and Neuroticism are nearly identical in

Personality Neuroscience 7
Eysenck’s system and the Big Five, whereas the unfortunately-labeled Psychoticism reflects a
roughly equal blend of low Conscientiousness and low Agreeableness (Goldberg & Rosolack,
1994). In his biological theorizing, Eysenck (1967; Eysenck & Eysenck, 1985) relied heavily on
the functions of the brain’s ascending reticular activating system, associating Extraversion with
the reticulo-cortical circuit and Neuroticism with the reticulo-limbic circuit. Eysenck
hypothesized (a) that extraverts have lower baseline levels of cortical arousal than introverts and
therefore choose more arousing activities in order to achieve their preferred level of arousal, and
(b) that extraverts may have higher preferred or optimal levels of arousal. He hypothesized that
neurotics are more easily aroused by emotion-inducing stimuli than are emotionally stable
people. Eysenck did not develop as well-specified a biological model of Psychoticism, but at
different times he hypothesized that Psychoticism was negatively associated with serotonergic
function (Eysenck, 1992) and positively associated with dopaminergic function (Eysenck, 1997).

Jeffrey Gray, who was Eysenck’s student, focused more heavily on neurobiology than on
personality, with an emphasis on the development of a “conceptual nervous system” describing
functional systems that could be mapped onto brain systems. The main components of this
conceptual nervous system are the Behavioral Approach System (BAS), which responds to cues
for reward, and the Fight-Flight-Freezing System (FFFS) and Behavioral Inhibition System
(BIS), which respond to two distinct classes of threatening stimuli (Gray & McNaughton, 2000;
Pickering & Gray, 1999). Immediately threatening, punishing, or frustrating stimuli activate the
FFFS, which produces active avoidance (panic and flight) or attempted elimination (anger and
attack). Stimuli that one needs or desires to approach but that also contain potential threat (thus
creating an approach-avoidance conflict) activate the BIS, which produces vigilance, rumination,
and passive avoidance, as well as anxiety and even potentially depression (Gray & McNaughton,

Personality Neuroscience 8
2000). (Approach-approach or avoidance-avoidance conflicts are less common but can also
activate the BIS, which responds specifically to any conflict between goals.) Biologically, Gray
linked the BAS to the dopaminergic system, the BIS primarily to the septo-hippocampal system
but also to the amygdala, and the FFFS to the amygdala, hypothalamus, and periaqueductal gray.

Gray’s model of personality, Reinforcement Sensitivity Theory (RST; see Chapter 18,
this volume), describes personality traits as a function of individual differences in the
sensitivities of BIS, BAS, and FFFS. Gray (1982) originally described two dimensions of
personality associated with BIS sensitivity and BAS sensitivity, which he labeled Anxiety and
Impulsivity respectively. Gray viewed Anxiety and Impulsivity as 30˚ rotations from
Neuroticism and Extraversion, respectively. Gray and McNaughton (2000) noted, however, that
questionnaire measures of Anxiety or BIS sensitivity are, in practice, difficult to distinguish from
Neuroticism. Further, they described Neuroticism as a general sensitivity to threat produced
jointly by FFFS and BIS. Additionally, several of Gray’s colleagues have recently suggested that
measures of Extraversion (not Impulsivity) may be the best measures of BAS sensitivity
(Pickering, 2004; Smillie, Pickering & Jackson, 2006). These parallels are consistent with
research showing that measures of BIS and BAS sensitivity tap the same latent constructs as
measures of Neuroticism and Extraversion (Elliot & Thrash, 2002; Zelenski & Larsen 1999).

Zuckerman (2005) has provided the most extensive review of personality neuroscience to
date, in the second edition of his book, Psychobiology of Personality. This book is organized
around a hybrid of the Big Five and Zuckerman’s own model of personality, the Alternative
Five, which are Sociability, Neuroticism-Anxiety, Aggression-Hostility, Impulsive Sensation
Seeking, and Activity. Zuckerman identified the first four of these with Extraversion,
Neuroticism, Agreeableness (reversed), and Conscientiousness (reversed), respectively, in the

Personality Neuroscience 9
Big Five. Factor analyses largely bear out these associations (Aluja, Garcia, & Garcia, 2002,
2004; Angleitner et al., 2004; Zuckerman, Kuhlman, Joireman, Teta, & Kraft, 1993), though the
situation is somewhat complicated for Impulsive Sensation Seeking, which is sometimes as
strongly associated with Extraversion and Openness/Intellect as with Conscientiousness and also
tends to show a moderate negative association with Agreeableness. Aditionally, these factor
analyses show that Zuckerman’s Activity scale serves as a marker of Extraversion,
Conscientiousness, or both.

We can not hope to summarize all of the biological research relevant to the Alternative
Five that is synthesized in Zuckerman’s (2005) Psychobiology of Personality (some of it will be
discussed below). In brief, Zuckerman linked personality traits to underlying behavioral
mechanisms, which he in turn linked to the brain functions of various neurotransmitters,
hormones, and enzymes. One notable feature of his theorizing is that he described behavioral
mechanisms (including approach, inhibition, and arousal) as determined by multiple biological
systems and as contributing to multiple traits. In his model, for example, approach is influenced
by both dopamine and testosterone and contributes to both Sociability and Impulsive Sensation

In contrast, Cloninger (1987) developed a model of personality traits based on the
premise that individual neurotransmitter systems might be related uniquely to specific traits.
Cloninger hypothesized that the dopaminergic system was linked to a trait of Novelty Seeking,
the serotonergic system to Harm Avoidance, and the norepinephrine system to Reward
Dependence. Cloninger’s latest model includes these three traits plus four others: Persistence,
Self-Directedness, Cooperativeness, and Self-Transcendence (Cloninger, Svrakic, & Przybeck,
1993). He hypothesized that the original three traits and Persistence reflect dimensions of

Personality Neuroscience 10
temperament, meaning that they should be evident early in ontogeny and strongly genetically
determined. In contrast, he hypothesized that Self-Directedness, Cooperativeness, and Self-
Transcendence reflect dimensions of character, meaning that they should develop later, being
determined by experience during development rather than primarily by genes.

Research has demonstrated several problems with Cloninger’s model. First, a simple
distinction between temperament and character appears untenable. The character traits show
much the same levels of heritability as the temperament traits (Ando et al., 2004; Gillespie,
Cloninger, Heath, & Martin, 2003). Second, evidence has accumulated to contradict the idea that
single neurotransmitter systems are responsible for Novelty Seeking, Harm Avoidance, and
Reward Dependence (Paris, 2005). Finally, Cloninger’s seven-factor structure has not proven
consistently replicable. Factor analyses have demonstrated (a) that the scales Cloninger
developed do not group together in the manner that he assigned them to his seven traits (Ando et
al., 2004; Ball, Tennen, & Kranzler, 1999; Herbst, Zonderman, McCrae, & Costa, 2000), and (b)
that his instrument is best described by the five factor structure of the Big Five (Markon et al.,
2005; Ramaniah, Rielage, & Cheng, 2002). Harm Avoidance and Self-Determination (reversed)
are both markers of Neuroticism. Cooperativeness, Persistence, and Self-Transcendence are
markers of Agreeableness, Conscientiousness, and Openness/Intellect, respectively. Reward
Dependence combines Agreeableness and Extraversion. Finally, Novelty Seeking shows a
pattern similar to Zuckerman’s Impulsive Sensation Seeking: it is most strongly associated with
Conscientiousness (reversed), but also consistently loads positively on Extraversion as well as
sometimes negatively on Agreeableness and positively on Openness.

Depue describes five trait dimensions: Agentic Extraversion, Affiliation, Anxiety, Fear,
and Nonaffective Constraint (Depue & Lenzenweger, 2005). Depue and Collins (1999) proposed