2
Kerry L. Jang and Fiona Choi
Where do the personality disorders come from? Are they passed down from parent to child or are they shaped by exposures to formative events, such as a dysfunctional relationship with one’s mother or a bump on the head? The role of nature and nurture on the development of personality and its disorders has been central in the field since personality types were first described (Torgersen, 2009). The history of the genetics of personality is very much the history of psychiatric genetics in general, so much so that triumphs and failures searching for the etiology of schizophrenia or bipolar disorder, are little different from the triumphs and failures searching for the etiology of personality disorder. Indeed, in the context of any history of psychiatric genetics, one can simply exchange the name of the disorder with another and the history is little different.
This state of affairs is perplexing given that several major psychiatric disorders from schizophrenia to autism and the major traits of personality and personality disorder types are among the most heritable of genetically complex medical illnesses. Indeed, one of the most replicable findings in psychiatric genetics are from studies comparing identical to fraternal twin similarities on personality disorders and traits showing that heritable factors account for half of the observed differences. However, 30 years of psychiatric genetics research demonstrate that high heritability does not necessarily facilitate identifying specific genetic causes and to date the history of psychiatric genetics is largely a story of non-replicated discoveries and unrealized expectations.
Why hasn’t high heritability made it easier to find the putative genes? One possible explanation is the “bandwagon effect.” Personality and personality disorder researchers have been quick to adopt the latest advances in genetics methodology, be they twin and adoption studies to estimate heritability, linkage and association studies looking for single gene effects, or now using high throughput single nucleotide pair (SNP) analyses to identify a multitude of genes. Indeed, it would seem that the failure to identify putative genes was very much attributed to the gene finding technique or concepts of gene function rather than the result of overlooking fundamental issues in personality disorder research that are likely behind the inability to identify specific genes. For example, how personality function is measured, either as a typology (e.g., diagnostic category as found in the DSM or ICD systems) or a continuum (e.g., dimensional model embodied in personality trait models), and any imprecision introduced by the chosen personality measure will affect the ability of any genetic association to be found. Current measurement issues and their effect on the hunt for genes has become a bit of a hackneyed subject with some of the key issues being discussed a decade ago and continuing to this day (e.g., Cloninger, 2012; Jang & Vernon, 2018; Livesley, 2008; Reichborn-Kjennerud, 2008).
Indeed, the measurement issue is so alive that the new DSM-5 has attempted to integrate the categorical and dimensional models into a functional hybrid that capitalizes on the strengths of each approach by using the strengths of one approach to mitigate the weakness of the other. Only time will tell if this has been successful. More immediately, with the creation of this hybrid model does this mean that genetics research must stop and retool to use this system, or wait until the validity and functionality of the new system is understood? Quite simply, the answer is “no.” It’s not just a simple question of measurement and waiting for the best available measure. Rather, genetics research can continue but to progress will require a rephrasing of the fundamental questions of what the purpose of genetics research is and where it is best applied. The purpose of the present chapter is to explore this “rephrasing” and to begin by revisiting the raison d’être of genetics research.
What Is the Purpose of Genetics Research?
The primary purpose of genetics research has been to identify putative gene(s) underlying a somewhat vaguely defined group of symptoms. That is a very narrow focus and quite to the contrary, genetics research has a role beyond gene hunting. For example, genetics can help refine how we measure and define personality disorder concepts regarding diagnoses or measurement. For example, determining which personality behaviors are pleiotrophic – that is, influenced by the same or different genetic influences – is useful to refine which behaviors are actually central to a diagnostic category or trait structure. Similarly, determining which personality disorder concepts and other mental illnesses are influenced by a common genetic basis highlights the interrelationships between personality and other conditions such as schizophrenia, and helps us understand why something such as schizotypal personality disorder exists and how it is related to schizophrenia. Perhaps the role of genetic research has nothing to do with genes at all, but rather its most useful function is to highlight the role of the environment, such as the influence of learning, observation, and the impact of environmental conditions that impinge on and shape behavior.
Broadening the fundamental research question also brings to light issues that need to be considered as part of this research, and this chapter discusses four encumbrances that we believe impact the advancement of genetic research on the personality disorders. The focus of the research question will shift as each is discussed. The four encumbrances are (1) the development of grand theories of personality, (2) issues of measurement that continue to beset the field, (3) limitations of genetics research, and (4) the obsession with genetics as a whole. We do not think that dealing with these interrelated issues warrants a wholesale revolution, but rather a shift in perspective that takes the best of the thinking and research extant while opening new avenues of investigation that bypasses the worst of it.
The Problem of Grand Theories
Gordon Allport (1937, p. 48) wrote, “personality is something and personality does something …” that outlined the central task for personality psychologists. Hence personality psychologists have attempted to fulfill his famous dictum by developing an empirically based unified model of personality that integrated all the rich ideas of psychoanalytic theorizing without the introspective methods. The eventual approach adopted was the lexical model that essentially took every word in the English dictionary that describes personality, had people rate themselves and others on these words using a Likert-type scale, and subjected the ratings to factor analysis. Factor analysis of the inter-correlations between these ratings extracted the common variance that defined basic traits, such as neuroticism or extraversion that are considered to exist in every individual, and individual differences were accounted for by the extent to which a person exhibited each of the traits.
However, debate soon ensued over the correct number of traits such as the famous debate over the “Big Five” – Neuroticism, Extraversion, Openness to Experience, Agreeableness, or Conscientiousness (Goldberg, 1990; McCrae & John, 1992) or the “Gigantic Three” – Neuroticism, Extraversion, and Psychoticism (Eysenck, 1994). The debate between the Big Five and Gigantic Three was eventually reconciled when both were shown to be compatible and the models just represented different levels of analysis (Draycott & Kline, 1995). This spurred debate over the existence of a general factor for personality, akin to Spearman’s general intelligence factor, g, extracted from cognitive ability data. Whether or not there is a general factor or only independent personality factors was really an artifact of factor analysis methods used to analyze the data, such as allowing factors to become inter-correlated (oblique versus orthogonal rotations) or factor extraction techniques, such as principal components that seeks to maximize the first factor, for example. The most important thing to recognize in personality research is that the measure used to develop the grand model focused largely on the normal range of function.
Similar issues were mirrored in the personality disorder research with one important difference caused by having its origins in the medical model that preferred to classify behavior into typologies as opposed to traits. Categories of personality disorder were created using prototypical patients whom clinicians agreed exhibited the symptoms indicative of the personality disorder under consideration. This led to a number of categories, such as borderline PD, schizotypal PD, and so on. A problem soon emerged that the diagnostic criteria often overlapped between categories. For example, symptoms of anxiety are a feature of many categories and the degree of overlap on symptoms across categories fueled revisions of the DSM or ICD with the collapsing or creation of new categories. As a result, multiple diagnoses were assigned to patients to cover all of their symptoms and the silliness of it all reached a head when categories such as “personality disorder not otherwise specified” were included to provide a diagnosis for someone who could not be classified. All of this comes as little surprise given that the creation of new categories was a decidedly political affair usually decided by a committee of experts, and ratified by vote at a convention. Moreover, the focus on disorder also defined the primary range of behaviors under study that were clinically significant forms of behavior without clear understanding of when normal behavior became abnormal. Instead, a broad criterion of whether or not a behavior interfered with daily activities was used.
A rapprochement between the two solitudes occurred when new scales of personality disorders began to emerge that embraced the content and dimensionality of personality function in its entirety (Trull & Widiger, 2013; Widiger, 2007). The scales were created using modern psychometric methods and techniques, and whose content was validated and reliability of measurement established by robust research programs in general population and clinical samples. A large body of research also exists that documents their relationship to existing measures of personality such as the NEO-PI-R and EPQ-R, and the well-understood diagnostic categories of the DSM-IV all on clinical and general population samples. Indeed, this body of research began to frame the need for a revision of the DSM-IV categories for the then upcoming DSM-5 – but that is another story (Franić, Borsboom, Dolan, & Boomsma, 2014; Widiger & Lowe, 2008). At last, a unified grand theory began to emerge that could explain what personality is and what it does.
Genetics research, particularly twin research was fundamental in supporting a grand theory because it showed that personality disorder concepts were related to one another because they were influenced by a common set of genetic factors to justify and define the broad concepts and trait domains (see Jang & Vernon, 2018 for a review). This research supported the idea that personality function is best conceptualized as a continuum of normal and extreme range behaviors. Furthermore, heritability analyses supported the hierarchical structure of traits into higher and lower order levels (viz., the Big Five or Gigantic Three debate) by showing that more specific lower order traits show some shared genetic influence and form into fewer broader traits, but a great deal of the variability observed in them remained unique to each facet (Mõttus, Kandler, Bleidorn, Riemann, & McCrae, 2017; Torgersen et al., 2012).
However, in terms of identifying actual loci, the broad nature of each trait – even facet traits that encompassed a range of behaviors – has made gene hunting impossible (Cloninger, 1987; Munafò et al., 2009; Verweij et al., 2010). The usual explanation for the failure included a range of methodological issues including but not limited to small sample size, identification of the wrong neurotransmitter or loci, and/or the use of an inappropriate instrument for measurement. We would argue that the more pressing problem lies with an overly broad and behaviorally complex phenotype. Perhaps the way forward is to move beyond research targeting broad concepts in the search for the grand theory, and instead shifting the focus of genetic research onto highly specific behaviors and emotions related to summative personality constructs of the grand theory. The work in personality nuances captures this idea. For example, two individuals may have the same high score on a measure of sensation seeking. However, one person may engage in skydiving while the other prefers horror movies. What accounts for these differences in expression – skydiving versus horror movies – is what might be central to new genetics research.
Do genes play a role in the differential expression of sensation seeking as opposed to genes underlying sensation seeking per se? Personality nuances represent a meaningful level of the trait hierarchy below facets that correspond roughly to single items (or groups of very similar items) in a facet scale (see McCrae, 2015). For example, bitterness and touchiness may be different nuances of angry hostility, a facet of neuroticism. Nuances may specify either the eliciting situation (e.g., fear of heights as a source of anxiety or inability to accept criticism as source of anger) or the characteristic response to a range of situations (e.g., a nervous tic as an expression of anxiety across different circumstances or feeling offended as a result of criticism of any kind). As such, nuances could be potentially more useful in understanding individuals and their differences.
Indeed, in a sample of twins, personality nuances operationalized from the NEO-PI-R showed good psychometric rank order stability of .72 and validity, and a significant heritable basis on average of 52 percent (Mõttus et al., 2017). Taking this a step further, Mohammad and Kiritchenko showed that fine affect or emotion categories such as excitement, guilt, yearning, and admiration are significant indicators of personality such as the Big Five (Mohammad & Kiritchenko, 2013) and conducting genetic analyses on the emotions associated with each of the main personality traits may be more informative. Finding the genes for what makes a person feel “keen,” “helpless,” “timid,” or “guilty” would be far more informative and clinically significant than the gene for neuroticism which these emotions predict.
The Problem of Measurement
No measure of personality function is perfectly reliable or has a large body of convergent and discriminative validity. These issues are to be distinguished from breadth of personality concepts discussed above, but instead focus on fundamental issues of how personality disorders are measured. Simply put, re-highlight the simple principle learned by all statistics students – “garbage in, garbage out” or GIGO. Research on the mainstream personality measures, be they Eysenck’s EPQ or Costa and McCrae’s NEO-PI-R for example, all converge to some consistent results – that there are three or five major traits, they are related to each other in predictable ways, and that the measures themselves have acceptable levels of reliability, validity, and stability. Such features are less so with personality disorder diagnoses. If the measure of the phenotype has fundamental psychometric problems, it will affect the veracity of any genetic study that it is based upon. This state of affairs is particularly so in the case of the personality disorders whose measurement and conceptualization has been a matter for debate for decades and does not appear to have been resolved with the DSM-5, whose changes remain debatable (Oldham, 2015).
Unfortunately, the long-running issues relevant to personality disorder diagnoses (Jang, Livesley, & Vernon, 1998) culminating with the wholesale and controversial changes made to the classification of personality disorders in the DSM-5 (e.g., Wakefield, 2016) has set back genetic research because those changes were not done solely to enhance reliability and validity. Rather, the decision to continue with suboptimal medical diagnostic categories was to provide health insurance companies with easy to bill conditions. Furthermore, the changes contained in the DSM-5 throw a wrench into new genetics research. A would-be researcher is directly confronted by this problem when deciding what measures to include in the next research grant proposal. Does one include the DSM-5 criteria as primary measures, perhaps include the DSM-IV criteria, a self-report dimensional measure of personality function, and for similar inclusiveness, a measure of the Big Five personality traits as well? The inclusion of measures would be less about including the most reliable and valid measures – those with the best psychometric difficulties – but those that will please the grant reviewer! Genetic researchers will have to decide whether the extent of the changes to the phenotype will mean starting all over again using the new measures, or ignoring these measures. It begs the question of just what do we do the genetics on?
The Future: Back to the Phenotype
Perhaps it is time the genetic research into personality move entirely away from traditional diagnostic approaches or responses to self-report questionnaires. Gottesman and Gould (2003) suggested focusing on “endophenotypes” – a biological marker that may contain a useful link between genetic sequences and behavioral disorders – and that these biological markers can parse behavioral symptoms into more stable phenotypes with a clear genetic connection. The definition of an endophenotype is the ensemble of measurable components in the pathway from distal genotype to psychiatric “disease” that fills the “invisible” gap between them. Individual endophenotypes refer to any one measure that contributes to specifying the pathway from genes to mental disorder. The task ahead is to identify potential endophenotypes for the personality disorders. For example, Siever (2005) suggested that some clinical dimensions of PDs, such as affective instability, impulsivity, aggression, emotional information processing, cognitive disorganization, social deficits, and psychosis, lend themselves to the study of corresponding endophenotypes. The propensity toward aggression can be evaluated by multiple methods including psychometric measures, interview, laboratory paradigms, neurochemical imaging, and pharmacological studies. These suggest that aggression is a measurable trait that may be related to a reduction in serotonergic activity. Hyper-responsiveness of the amygdala and other limbic structures could be related to affective instability, while structural and functional brain alterations underlie the cognitive disorganization in psychotic-like symptoms of schizotypal personality disorder.
Ruocco, Amirthavasagam, and Zakzanis (2012) evaluated whether the magnitude of volume reductions in the amygdala and hippocampus was associated with BPD. Volumetric magnetic resonance imaging results from 11 studies comprising 205 patients with BPD and 222 healthy controls were examined using meta-analytic techniques. Patients showed an average 11 and 13 percent decrease in the size of the hippocampus and amygdala, respectively. No attenuation of volumetric differences was detected in patients being treated with psychotropic medications; and comorbid depression, posttraumatic stress disorder, and substance use disorders were unrelated to volumetric decreases in either structure.
Ruocco and Carcone (2016) reviewed the literature on the neurobiology of borderline personality disorder (BPD), identifying 146 articles in three broad research areas: neuroendocrinology and biological specimens; structural neuroimaging; and functional neuroimaging. Based on the consolidation of results from these studies, they suggest an integrative model to account for interactions among endogenous stress hormones, neurometabolism, and brain structures and circuits involved in emotion and cognition. They concluded that genetics research could profitably incorporate endophenotypes, and gene × environment interaction research that focuses on the expression of genes in response to environmental stressors given that multiple neurobiological systems interact to produce the complex clinical phenotype of the disorder. These include interconnections between hormones, neuropeptides, brain metabolites, neurotransmitter receptors, white matter pathways, gray matter volumes, and neural activity associated with emotion, cognition, and the sense of self (Ruocco & Carcone, 2016).
The studies highlighted above represent a classic approach to finding endophenotypes for personality disorders. However, what is emerging in the literature is the use of “intermediate endophenotypes” such as personality traits as endophenotypes for other major disorders. This trend in the research is occurring because certain personality traits seem to be overrepresented in people with specific disorders. For example, Ersche and colleagues (2012) identified anxious-impulsiveness and studied personality and cognitive dysfunction as endophenotypes for drug dependence. These types of studies are interesting in their attempt to find the genes for another disorder that identifies potential genes underlying a related set of personality traits and functions! It is perhaps within these constellations of traits that the genes may be best identified, as opposed to the previously adopted approach examining traits individually and out of context; or arbitrary groupings that are not observed together in a clinical (i.e., real-world) setting.
A good example of personality traits as the endophenotype for a clinical syndrome is the study by Savitz, Van Der Merwe, and Ramesar (2008). This study used personality endophenotypes for a genetic association analysis of bipolar affective disorder (BPAD). They reasoned that various personality traits are overrepresented in people with BPAD and their unaffected relatives, and these traits may constitute genetically transmitted risk factors or endophenotypes of the illness (Qiu, Akiskal, Kelsoe, & Greenwood, 2017; Savitz et al., 2008). Seven different personality questionnaires comprising 19 subscales were administered to 31 European American families with BPAD (n = 241). Ten of 19 personality traits showed significant evidence of heritability and were selected as candidate endophenotypes. The 3′ untranslated region repeat polymorphism of the dopamine transporter gene (SLC6A3) was associated with scales measuring Self-Directedness and Negative Affect. The short allele of the serotonin transporter gene (SLC6A4) promoter polymorphism showed a trend toward association with higher Harm Avoidance and Negative Affect. The COMT Val158Met polymorphism was weakly associated with Spirituality and Irritable Temperament.
This brief review shows that endophenotypes can take many forms. Where might the search for endophenotypes go next? Kraus (2013) suggests that one direction might be to delve deeper and examine cellular function:
One thing that early gene–personality work overlooked is that a lot has to happen to allow DNA to code for specific hormones/neuropeptides that then have to act at the cellular level to subsequently influence personality. In short, genes need to be expressed at a cellular level in order to influence personality, and so one place where a genetic researcher might want to look to examine gene influences on personality is at this expression – that is, what genes are being unzipped by RNA, so that specific hormones/proteins are produced?
Middeldorp and colleagues (Middeldorp, Ruigrok, Cath, Van Dyck, and Boomsma, 2002) indicated that research on non-human subjects may provide some exciting leads. Research on honeybees is suggestive of the potential of examining RNA (ribonucleic acid) to predict behavior. In this work, mRNA (messenger ribonucleic acid) abundance has been shown to be a significant predictor of behavioral transitions of honeybees from hive workers to foragers (Whitfield, Cziko, & Robinson, 2003). Human work in this domain is an exciting area of future research. However, other investigators such as Paris (2011) remain much less enthusiastic about the utility of endophenotypes. The identification and use of endophenotypes are associated with the assumption that mental processes are reducible to activity at a neuronal level.
The Problem of Behavioral Genetic Methods
The vast majority of the genetic research on personality disorders has largely been based on the analysis of twin similarities. Specifically the comparisons between identical (monozygotic, or MZ) and fraternal (dizygotic, or DZ) twin pairs are used to estimate heritability – the proportion of the observed variability on a measure directly attributable to genetic differences between individuals. The primary reason for the popularity of twin studies is that this method can handle quantitative measures typical of personality research and uses model fitting to test a broad array of questions. However, the largely singular focus on twin research suggests that two or more decades since this research began, the research using twins needs to take some new directions for progress in the field to be made.
Most twin studies use data obtained from reared-together twins because of the relative ease of finding a large representative sample, although there are several variations of the basic design, such as twins reared apart and family-of-twin designs (see Plomin, DeFries, & McClearn, 1990). A correlation coefficient (e.g., Pearson’s r) indexes the similarity of twins. Greater MZ to DZ similarity is directly attributable to the two-fold greater genetic similarity of MZ to DZ twins, assuming all other things being equal. This is because MZ twins share all of their genes, whereas DZ twins share only half on average.
As a simple way to understand the logic of the twin study, if rMZ = .42 and rDZ = .25, the proportion (%) to which the individual differences observed on a measure is due to genetic differences, or the “heritability coefficient,” is estimated as:
Heritability (h2) = 2(rMZ – rDZ) = 2 (.42 – .25) = .34 (100%) = 34%
The heritability coefficient, h2, estimates genetic influences from all sources: additive genetic influences (A or the extent to which genotypes “breed true” from parent to offspring) and genetic dominance (D or genetic effects attributable to the interaction of alleles at the same locus, which results in a phenotype that is not exactly intermediate in expression as would be expected between pure breeding homozygous individuals).
Two environmental effects are also estimated. Common or shared environmental influences (c2 or C) are defined as those that make members of a family similar to one another (i.e., result in familial resemblance), whereas non-shared environmental factors (e2 or E) are those which make members of a family different from one another (i.e., result in differences between family members). It is important to note that it is not that the experience of the environment itself that is shared or not shared, but how these factors influence the resemblance of family members. The non-shared environment term also includes measurement error as this also lowers familial resemblance in a random way. It follows then that:
c2 = 2 rDZ – rMZ = 2(.25) − .45 = .05 (100%) = 5%
and
e2 = 1.0 – h2 – c2 = 1.0 − .34 − .05 = .64 (100%) = 64%
The basic twin method has been translated into path-analytic models (see Neale & Cardon, 1992). Path models are extremely flexible in that they are able to analyze data from different populations and response formats (including diagnostic categories), and have generated an explosion of studies over the past two decades.
Heritability estimates generated using path analytic methods or by the simple equations above are predicated on the same assumptions that imparts imprecision into the estimates. One of the principal assumptions is that greater MZ to DZ similarity on a measured trait is due to genetic factors because MZ twins share all of their genes and DZ twins only half. However, this only holds if the environments of the MZ twins do not cause them to be more similar than DZ twins. That is, the greater MZ twin similarity may not be due to the fact that parents of MZ twin pairs treat their twins (e.g., dressing alike) more similarly than parents of DZ twins. This is known as the “assumption of equal environments” or EEA. Traditionally the test that the assumption holds is by rating the similarity of MZ and DZ twin environments and showing there are no significant differences. For example, when twins are being tested they are asked to complete measures that assess the degree to which they were often dressed alike, went to the same schools, and so on. MZ and DZ agreement or concordance rates on these items are compared and if differences are found (suggesting that the environments of MZ and DZ twins are not the same), then the affected twin similarity variables are correlated with the dependent measure(s) to determine whether they account for a significant proportion of the variance. It has been suggested that environmental similarity between twins does not have much of an impact on trait similarity (Felson, 2014), and that the interaction of sociocognitive variables in response to different environmental conditions such as trauma, being dressed alike, or medical treatments should be the focus (Fosse, Joseph, & Richardson, 2015). A correction for these effects is made by estimating the standardized residual from the regression of the twin similarity on these sociocognitive variables on the personality variables prior to genetic analyses, which may lead to some changes in heritability estimates and our interpretation of them (Fosse et al., 2015).
A second consideration is that twin studies require a relatively large number of twin pairs to have adequate power to detect genetic and environmental influences with any certainty (see Neale, Eaves, & Kendler, 1994). Many of the largest studies have established population-based twin registries where all of the twin pairs in a population are identified by birth records. Other large studies have used volunteer samples drawn from a population. With either method the question remains – are there sufficient numbers of twin pairs with personality disorders to actually study the genetics of personality disorders? Unlike studies of normal personality where everyone is assumed to have some kind of personality – it is difficult to recruit a large sample of twins who have a clinically significant personality disorder. Approximately 32 out 1000 people are twins (about 3 percent of the general population), and if the prevalence of DSM-IV personality disorder itself is 9.1 percent (Lenzenweger, Lane, Loranger, & Kessler, 2007), and the same paper reports prevalence rates for specific diagnoses such as borderline personality disorder at 1.4 percent and antisocial personality disorder 0.6 percent, few affected pairs would be recruited into any study. Twin studies attempt to recruit as large samples as possible using a wide variety of recruitment methods that range from using hospital or church birth records to media advertisements. A recent issue of Twin Research and Human Genetics (Hur & Craig, 2013) lists the world’s major twin studies and despite the size of some of the samples, suitable numbers of twin pairs with personality disorder would not be found.
To get around the problem, the threshold liability model of disease is evoked that assumes that personality function exists on a continuum of normal range and extreme behavior, and that biological and environmental factors are assumed to affect a person’s position on the continuum in a particular way. Under the threshold liability model of disease, the number of individuals in a population falling into each range on a continuum of behavior – normal range, spectrum, and disorder – is determined by the amount, or “dosage,” of genetic and environmental influence. The model is multifactorial in nature and assumes that several genes and environmental effects combine to create an individual’s susceptibility. This suggests that patients differ from non-patients only in the number of pathogenic genetic and/or environmental events or experiences to which they have been exposed.
The threshold liability model is readily modified to explain disorders that exhibit clear discontinuities in the expression of pathology as seen with bimodal distributions of behavior. Under this variant of the model, the same multifactorial causes are still exerting an influence that creates much of the measured variability between individuals, with the addition of one or more significant genetic (e.g., specific gene variant for example) and/or environmental causes (e.g., traumatic experience or exposure) that creates the patient group. The threshold liability model is also important because it explains why the pattern of responses of general population subjects to items assessing PD and symptoms of psychopathology is similar to those of clinical samples (e.g., Jackson & Messnick, 1962; Livesley, Jackson, & Schroeder, 1992; Livesley, Schroeder, Jackson, & Jang, 1998). As such, behavioral genetic research is designed to estimate the extent to which the vulnerability or dosage is attributable to genetic causes by comparing greater genetic similarity of relatives for a phenotype as compared to unrelated individuals.
The other important factor is the importance of the content of a measure used reflecting the continuum of phenotypic expression. If not, the study will be a reflection of the restricted range of behaviors covered by the measures. If a study uses a measure of personality disorders where few indicate that they display behaviors as described in the measure, then the study is not really examining what the measure purports to assess. The solution is to measure personality function in as many ways as possible so that the range of behavior in all its forms and minutiae are covered. With the use of multiple measures, it is imperative that the relationship between the measures is understood – that is, they are related to each other in predictable ways. Moreover, the terms of the threshold liability model of disease in genetics show that genes underlie any observed relationship between the scales as well. In this way, one is assured that the measures are indeed measuring the same constructs in different ways (viz., multi-trait, multi-method matrix) and that the same genetic and environmental factors account for the observed relationships. The degree to which two measures are influenced by a common genetic (pleiotropy) and environmental influences are readily computed from MZ and DZ twin correlations and indexed by the genetic correlation (rG) and environmental correlation (rE), which vary from −1.0 to 1 and are interpreted like any correlation coefficient.
Even a measure that reflects the widest range of the behavior in question does not move research forward because it still requires that sufficient twin pairs fall into the extreme range. Few studies report the range of response or numbers of individuals who fall into the extreme range and thus it is not clear if the model is valid and studies of personality disorder may just be studies of normal personality! Another way to approach the problem is to turn the threshold liability model of genetics upside down. Normally, the threshold liability model begins with the phenotype and makes assumptions about gene dosage. A more useful approach may be to start with the genes themselves to see if they are associated with our measures or conceptions of disorder.
For example, the recent use of genomic-relatedness-matrix residual maximum likelihood (GREML) analysis has been producing new insights into the genetic architecture of personality. GREML works by looking at how very low levels of relatedness, as determined from number of shared variants across the genome, account for similarity in phenotype across traditionally unrelated individuals. In other words, GREML allows the estimation of the total genetic heritability of a trait by taking into account all gene variants available in a data set, without identifying the specific gene variants making up this heritability.
Recent GREML studies of different personality traits have been able to confirm underlying genetic heritability. For example, in a sample of 12,000 unrelated individuals, common single-nucleotide polymorphisms (SNPs) accounted for 6 percent of the variance in neuroticism and 12 percent in the case of extraversion (Vinkhuyzen et al., 2012). The only other study that we are aware of reporting GREML estimates for personality traits found that genetic variants explained 7 percent of the variance in harm avoidance, 10 percent in novelty seeking, and 8 percent in persistence (but no variance in reward dependence) in a sample of 8000 individuals (Verweij et al., 2012). The use of this approach prevents reliance on twins and allows researchers to study directly unrelated individuals with personality disorder diagnoses increasing the chances of obtaining a sufficiently large sample. Moreover, given that certain SNPs are accounting for much lower estimates of heritability on the measures suggests that the method and results can be used to decompose the personality function phenotypes into “nuances” that predict actual behavior as discussed earlier.
The Problem of Not Seeing the Forest Through the Trees
A common answer to the question of “why do we want to find the genes for ______FILL IN THE BLANK HERE____” is because of the general belief that genes cause a behavior and/or an illness. Finding the genes for a condition is important because the liability genes could be manipulated to ameliorate disease. As we stated in the opening of the chapter, this is the raison d’être of genetics research and has become its guiding principle. However, this was done at the cost of ignoring a host of concepts important to human beings such as time, development, critical periods, interactions with genes, and that genes may be a protective factor and not always a liability. This is odd because so much of psychology and psychiatry is about these processes but our estimates of heritability are often single point estimates on measures that summarize behavior over a lifetime. It’s time to seriously incorporate how life transitions, events, and role changes can affect both average slopes in personality traits and individual differences in personality maturation and change (Bleidorn et al., 2010; Bleidorn, Kandler, & Caspi, 2014; Hopwood et al., 2011).
These ideas resonate with Gottlieb’s (1984) themes including: (a) the agency of organisms in constructing their environments (Odling-Smee, Laland, & Feldman, 2003); (b) plasticity of development (West-Eberhard, 2003); (c) the role of phenotypic plasticity in the genesis of evolutionary novelty (Kirschner & Gerhart, 2005); and (d) the deeply contextual character of biological information. These themes not only echo and support many of Gottlieb’s own arguments, but also extend the “developmental point of view” into new domains.
A serious consideration of these issues is vital for the genetics of personality disorders to progress. Recent research on intelligence, actually called the “new genetics of intelligence” can contribute some key ideas to personality disorder research. For example, thousands of single-nucleotide polymorphisms are required to account for the heritability of intelligence because the effect sizes of SNP associations are very small. The new genetics of intelligence relies on the combination of thousands of these SNP effects in a genome-wide polygenic score (GPS), also applicable to complex traits in personality research (Plomin & von Stumm, 2018). Intelligence or cognitive ability have always been considered to run in families with the assumption that this family resemblance was due to nurture, which we defined earlier as “shared family environmental influence.” Siblings were thought to be similar in intelligence because they grew up in the same environment and twin and adoption studies consistently support this assumption, but only until adolescence. After adolescence, the effect of shared family environmental influence on intelligence is negligible, which means that family environments have very little effect on individual differences in the long run. Under the normal range of environmental influence, in the absence of extremes of neglect or abuse, family resemblance for intelligence is due to nature rather than nurture (Briley & Tucker-Drob, 2013).
Kandler and Papendick (2017) have taken these issues into account in studying the genetics of personality disorder by introducing three perspectives of how quantitative behavior genetic modeling can broaden our knowledge about the etiology of personality differences, stability, and change. First, based on the data from 14 cross-sectional, 13 longitudinal, and 3 cross-sequential studies, they illustrate age trends of the genetic and environmental contributions to individual differences in five personality trait dimensions: neuroticism, extraversion, openness, agreeableness, and conscientiousness. Second, they demonstrate estimates of the stability of genetic and environmental differences in personality traits across time and different age groups using the data from 16 longitudinal studies. Finally, they visualize age trends of the genetic and environmental contributions to the stability of personality differences over the life course. In the context of these three perspectives, there is an explanation of the implications that these trends may have for the interplay between genetic and environmental sources during different stages of life and how they can deepen our understanding of personality development across the lifespan (Kandler & Papendick, 2017).
Perhaps the reason why it has been difficult to identify genes is because their effects are only “activated” by exposure to specific environmental triggers, an effect known as “gene × environment interaction.” One of the most investigated genes is the serotonin transporter gene (SLC6A4, also known as 5-HTT), which has been the focus of many personality studies. As far back as 2003, Caspi and colleagues showed that the 5-HTTLPR polymorphism does not show a main effect on depression, but the s-allele increases the risk of depression once an individual is exposed to one or more life events. However, two meta-analyses, including 5 and 14 studies respectively, yielded no evidence for an effect of 5-HTTLPR in interaction with life events on depression (Munafò, Durrant, Lewis, & Flint, 2009; Risch et al., 2009). Subsequently, these meta-analyses were critiqued for having given too much weight to the studies reporting null findings that employed poorer measurement of life events (Caspi, Hariri, Holmes, Uher, & Moffitt, 2010).
Middeldorp and colleagues (Middeldorp et al., 2010) tested for an interaction effect involving 5-HTTLPR on a sample of 1155 twins and their parents and siblings from 438 families, using a detailed measure of life events. They found a significant main effect of number of life events on anxious depression and neuroticism, especially when these were experienced in the past year. No interaction with 5-HTTLPR was found for number of life events either experienced across the lifespan or across the past year, supporting the findings of the meta-analyses. It might be more useful to focus on the joint effect of several genes that are, for example, part of the same biological pathway in interaction with the environment.
The idea of analyzing several polymorphisms simultaneously was tested by Heck and colleagues (2009). They reasoned that previous studies that only examined one or a few polymorphisms within single genes neglected the possibility that the genetic associations might be more complex, comprising several genes or gene regions. As such, they performed an extended genetic association study analyzing 17 serotonergic (SLC6A4, HTR1A, HTR1B, HTR2A, HTR2C, HTR3A, HTR6, MAOA, TPH1, TPH2) and dopaminergic genes (SLC6A3, DRD2, DRD3, DRD4, COMT, MAOA, TH, DBH), which have been previously reported to be implicated with personality traits. One hundred ninety-five SNPs within these genes were genotyped in a sample of 366 general population participants (all European American), and they conducted a replication on an independent sample of a further 335 participants. Personality traits in both samples completed the German version of Cloninger’s Tridimensional Personality Questionnaire (TPQ). From 30 SNPs showing associations at a nominal level of significance, two intronic SNPs, rs2770296 and rs927544, both located in the HTR2A gene, withstood correction for multiple testing. These SNPs were associated with Novelty Seeking. The effect of rs927544 could be replicated for the Novelty Seeking subscale Extravagance, and the same SNP was also associated with Extravagance in the combined samples. Their results show that HTR2A polymorphisms modulate facets of novelty-seeking behavior in healthy adults, suggesting that serotonergic neurotransmission is indeed involved in this phenotype.
Similarly, Derringer and colleagues (Derringer et al., 2010) led a consortium of researchers in an examination of a collection of SNPs associated with dopamine in prior research and subsequently examined associations between this collection of SNPs and sensation-seeking behavior. The findings were promising: Taking into account all the SNPs associated with sensation-seeking behaviors as an aggregate, dopamine genes worked in concert to explain around 6.6 percent of variation in sensation-seeking behavior. This approach is appealing because it involves conceiving of genes and personality not as simple one-to-one relationships, but instead as complex systems of genes that work in concert to express a personality trait.
The importance of including critical periods for development is demonstrated by a number of studies that show when personality traits become more or less stable. Hopwood et al. (2011) investigated the patterns and origins of personality trait changes from ages 17 to 29 using three waves of Multidimensional Personality Questionnaire data provided by twins. Results suggest that (a) trait changes were more profound in the first relative to the second half of the transition to adulthood; (b) traits tend to become more stable during the second half of this transition, with all the traits yielding retest correlations between .74 and .78; (c) negative affectivity declined over time and constraint increased over time; minimal change was observed on agentic or communal aspects of positive affectivity; and (d) both genetic and non-shared environmental factors accounted for personality changes (Hopwood et al., 2011). Overall, these genetically-informed results support a life-course perspective on personality development during the transition to adulthood.
Conclusion
The genetics of personality disorders will remain an active area of research and we hope this targeted chapter will provide some new ideas for the next phase of research. The genetics of personality disorders has been very much a “top-down” approach, where research begins with an imperfect phenotype and the search for genes underlying that imperfect phenotype. Is it any wonder the search for personality disorder liability genes (and environmental causes) have been so unsuccessful.
In this chapter we argue that it is time to switch gears in the study of the genetics of personality disorders by turning how we have currently thought about personality functioning upside down. For starters, the focus must shift from trying to find the genes associated with broad and inclusive psychological concepts and psychiatric typologies such as sensation seeking to the psychological processes of explaining behavioral choices associated with these broad concepts – such as when someone chooses horror movies over skydiving. The nuances of personality as opposed to broad traits and concepts are an avenue of genetics research that may lead to fruitful discoveries. Along these same lines, finding genes underlying the biological process associated with behavior, the so-called “endophenotypes,” is worth continued exploration simply because, like in the case of personality nuances, focus on the broad concepts has not been successful.
Of particular note are new methods that do not rely on traditional twin studies and incorporating considerations around developmental periods as opposed to point estimates is another shift worth following up. The true value of future genetic research lies with how they offer the promise to better understand actual behavior and not simply social constructs that make up current diagnostic systems.
It is an exciting time to be working in personality disorders.
References
Allport, G. (1937). Personality: A Psychological Interpretation. New York: Holt, Rinehart & Winston.
Bleidorn, W., Kandler, C., & Caspi, A. (2014). The behavioural genetics of personality development in adulthood: Classic, contemporary, and future trends. European Journal of Personality, 28(3), 244–255.
Bleidorn, W., Kandler, C., Hülsheger, U. R., Riemann, R., Angleitner, A., & Spinath, F. M. (2010). Nature and nurture of the interplay between personality traits and major life goals. Journal of Personality and Social Psychology, 99(2), 366–379.
Briley, D. A., & Tucker-Drob, E. M. (2013). Explaining the increasing heritability of cognitive ability across development. Psychological Science, 24(9), 1704–1713.
Caspi, A., Hariri, A. R., Holmes, A., Uher, R., & Moffitt, T. E. (2010). Genetic sensitivity to the environment: The case of the serotonin transporter gene and its implications for studying complex diseases and traits. American Journal of Psychiatry, 167, 509–527.
Cloninger, C. R. (1987). A systematic method for clinical description and classification of personality variants. Archives of General Psychiatry, 44, 573–588.
Cloninger, R. (2012). Genetics of personality disorders. In J. I. Nurnberger & W. H. Berrettini (Eds.), Principles of Psychiatric Genetics (pp. 316–323). Cambridge University Press.
Derringer, J., Krueger, R. F., Dick, D. M., Saccone, S., Grucza, R. A., Agrawal, A., … Schuckit, M. A. (2010). Predicting sensation seeking from dopamine genes: A candidate-system approach. Psychological Science, 21(9), 1282–1290.
Draycott, S. G., & Kline, P. (1995). The big three or the big five—the EPQ-R vs the NEO-PI: A research note, replication and elaboration. Personality and Individual Differences, 18(6), 801–804.
Ersche, K. D., Turton, A. J., Chamberlain, S. R., Müller, U., Bullmore, E. T., & Robbins, T. W. (2012). Cognitive dysfunction and anxious-impulsive personality traits are endophenotypes for drug dependence. American Journal of Psychiatry,169, 926–936.
Eysenck, H. J. (1994). The big five or giant three: Criteria for a paradigm. In C. F. Halverson, Jr., G. A. Kohnstamm, & R. P. Martin (Eds.), The Developing Structure of Temperament and Personality from Infancy to Adulthood (pp. 37–51). Hillsdale, NJ: Lawrence Erlbaum Associates.
Felson, J. (2014). What can we learn from twin studies? A comprehensive evaluation of the equal environments assumption. Social Science Research, 43, 184–199.
Fosse, R., Joseph, J., & Richardson, K. (2015). A critical assessment of the equal-environment assumption of the twin method for schizophrenia. Frontiers in Psychiatry, 6, 62.
Franić, S., Borsboom, D., Dolan, C. V., & Boomsma, D. I. (2014). The big five personality traits: Psychological entities or statistical constructs? Behavior Genetics, 44(6), 591–604.
Goldberg, L. R. (1990). An alternative “description of personality”: The big-five factor structure. Journal of Personality and Social Psychology, 59(6), 1216–1229.
Gottesman, I. I., & Gould, T. D. (2003). The endophenotype concept in psychiatry: Etymology and strategic intentions. American Journal of Psychiatry, 160, 636–645.
Gottlieb, G. (1984). Evolutionary trends and evolutionary origins: Relevance to theory in comparative psychology. Psychological Review, 91(4), 448–456.
Heck, A., Lieb, R., Ellgas, A., Pfister, H., Lucae, S., Roeske, D., … Ising, M. (2009). Investigation of 17 candidate genes for personality traits confirms effects of the HTR2A gene on novelty seeking. Genes, Brain and Behavior, 8, 464–472.
Hopwood, C. J., Donnellan, M. B., Blonigen, D. M., Krueger, R. F., McGue, M., Iacono, W. G., & Burt, S. A. (2011). Genetic and environmental influences on personality trait stability and growth during the transition to adulthood: A three-wave longitudinal study. Journal of Personality and Social Psychology, 100(3), 545–556.
Hur, Y. M., & Craig, J. M. (2013). Twin registries worldwide: An important resource for scientific research. Twin Research and Human Genetics, 16(1), 1–12.
Jackson, D. N., & Messnick, S. (1962). Response styles on the MMPI: Comparison of clinical and normal samples. Journal of Abnormal and Social Psychology, 65, 285–299.
Jang, K. L., Livesley, W. J., & Vernon, P. A. (1998). A twin study of genetic and environmental contributions to gender differences in traits delineating personality disorder. European Journal of Personality, 12(5), 331–344.
Jang, K. L., & Vernon, P. A. (2018). Genetics. In W. J. Livesley (Ed.), Handbook of Personality Disorders: Theory, Research, and Treatment (pp. 177–195). New York: Guilford Press.
Kandler, C., & Papendick, M. (2017). Behavior genetics and personality development: a methodological and meta-analytic review. In J. Specht (Ed.), Personality Development Across the Lifespan (pp. 473–495). London: Academic Press.
Kirschner, M. W., & Gerhart, J. C. (2005). The Plausibility of Life: Resolving Darwin’s Dilemma. New Haven: Yale University Press.
Kraus, M. W. (2013). Do genes influence personality? Psychology Today, July 11. Retrieved from www.psychologytoday.com/gb/blog/under-the-influence/201307/do-genes-influence-personality.
Lenzenweger, M. F., Lane, M. C., Loranger, A. W., & Kessler, R. C. (2007). DSM-IV personality disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 62(6), 553–564.
Livesley, W. J. (2008). Research trends and directions in the study of personality disorder. Psychiatric Clinics of North America, 31(3), 545–559.
Livesley, W. J., Jackson, D. N., & Schroeder, M. L. (1992). Factorial structure of traits delineating personality disorders in clinical and general population samples. Journal of Abnormal Psychology, 101, 432–440.
Livesley, W. J., Schroeder, M. L., Jackson, D. N., & Jang, K. L. (1994). Categorical distinctions in the study of personality disorder: Implications for classification. Journal of Abnormal Psychology, 103(1), 6–17.
McCrae, R. R. (2015). A more nuanced view of reliability: Specificity in the trait hierarchy. Personality & Social Psychology Review, 19, 97–112.
McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60(2), 175–215.
Middeldorp, C. M., Ruigrok, P., Cath, D. C., Van Dyck, R., and Boomsma, D. I. (2002). Candidate genes for mood disorders in humans: A literature review. American Journal of Human Genetics, 114, 768.
Middeldorp, C. M., Slof-Op, M. C. T., Landt, S. O., Medland, S. E., van Beijsterveldt, C. E., Bartels, M., … Boomsma, D. I. (2010). Anxiety and depression in children and adults: Influence of serotonergic and neurotrophic genes? Genes, Brain and Behavior, 9, 808–816.
Mohammad, S. M., & Kiritchenko, S. (2013). Using nuances of emotion to identify personality. Paper presented at the ICWSM Workshop on Computational Personality Recognition. Retrieved from http://arxiv.org/abs/1309.6352.
Mõttus, R., Kandler, C., Bleidorn, W., Riemann, R., & McCrae, R. R. (2017). Personality traits below facets: The consensual validity, longitudinal stability, heritability, and utility of personality nuances. Journal of Personality and Social Psychology, 112(3), 474–490.
Munafò, M. R., Durrant, C., Lewis, G., & Flint, J. (2009). Gene–environment interactions at the serotonin transporter locus. Biological Psychiatry, 65, 211–219.
Munafò, M. R., Freimer, N. B., Ng, W., Ophoff, R., Veijola, J., Miettunen, J., … Flint, J. (2009). 5-HTTLPR genotype and anxiety-related personality traits: A meta-analysis and new data. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 150B(2), 271–281.
Neale, M. C., & Cardon, L. R. (1992). Methodology for Genetic Studies of Twins and Families. London: Kluwer.
Neale, M. C., Eaves, L. J., & Kendler, K. S. (1994). The power of the classical twin study to resolve variation in threshold traits. Behavior Genetics, 24(3), 239–258.
Odling-Smee, F. J., Laland, K. N., & Feldman, M. W. (2003). Niche Construction: The Neglected Process in Evolution. Monographs in Population Biology, vol. 37. Princeton: Princeton University Press.
Oldham, J. M. (2015). The alternative DSM-5 model for personality disorders. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 14(2), 234–236.
Paris, J. (2011). Endophenotypes and the diagnosis of personality disorders. Journal of Personality Disorders, 25, 260–268.
Plomin, R., DeFries, J. C., & McClearn, G. E. (1990). Behavioral Genetics: A Primer (2nd ed.). New York: Freeman.
Plomin, R., & von Stumm, S. (2018). The new genetics of intelligence. Nature Reviews Genetics, 19, 148–159.
Qiu, F., Akiskal, H. S., Kelsoe, J. R., & Greenwood, T. A. (2017). Factor analysis of temperament and personality traits in bipolar patients: Correlates with comorbidity and disorder severity. Journal of Affective Disorders, 207, 282–290.
Reichborn-Kjennerud, T. (2008). Genetics of personality disorders. Psychiatric Clinics of North America, 31(3), 421–440.
Risch, N., Herrell, R., Lehner, T., Liang, K. Y., Eaves, L., Hoh, J. … Merikangas, K. R. (2009). Interaction between the serotonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: A meta-analysis. Journal of the American Medical Association, 301, 2462–2471.
Ruocco, A. C., Amirthavasagam, S., & Zakzanis, K. K. (2012). Amygdala and hippocampal volume reductions as candidate endophenotypes for borderline personality disorder: A meta-analysis of magnetic resonance imaging studies. Psychiatry Research, 201(3), 245–252.
Ruocco, A. C., & Carcone, D. (2016). A neurobiological model of borderline personality disorder. Harvard Review of Psychiatry, 24(5), 311–329.
Savitz, J., Van Der Merwe, L., & Ramesar, R. (2008). Personality endophenotypes for bipolar affective disorder: a family‐based genetic association analysis. Genes, Brain and Behavior, 7(8), 869–876.
Siever, L. J. (2005). Endophenotypes in the personality disorders. Dialogues in Clinical Neuroscience, 7(2), 139–151.
Torgersen, S. (2009). The nature (and nurture) of personality disorders. Scandinavian Journal of Psychology, 50(6), 624–632.
Torgersen, S., Myers, J., Reichborn-Kjennerud, T., Røysamb, E., Kubarych, T. S., & Kendler, K. S. (2012). The heritability of Cluster B personality disorders assessed both by personal interview and questionnaire. Journal of Personality Disorders, 26(6), 848–866.
Trull, T. J., & Widiger, T. A. (2013). Dimensional models of personality: The five-factor model and the DSM-5. Dialogues in Clinical Neuroscience, 15(2), 135–146.
Verweij, K. J., Zietsch, B. P., Medland, S. E., Gordon, S. D., Benyamin, B., Nyholt, D. R., … Wray, N. R. (2010). A genome-wide association study of Cloninger’s Temperament scales: Implications for the evolutionary genetics of personality. Biological Psychology, 85(2), 306–317.
Verweij, K. J., Yang, J., Lahti, J., Veijola, J., Hintsanen, M., Pulkki-Råback, L., … Zietsch, B. P. (2012). Maintenance of genetic variation in human personality: Testing evolutionary models by estimating heritability due to common causal variants and investigating the effect of distant inbreeding. Evolution: International Journal of Organic Evolution, 66(10), 3238–3251.
Vinkhuyzen, A. A. E., Pedersen, N. L., Yang, J., Lee, S. H., Magnusson, P. K. E., Iacono, W. G., … Wray, N. R. (2012). Common SNPs explain some of the variation in the personality dimensions of neuroticism and extraversion. Translational Psychiatry, 2(4), e102.
Wakefield, J. (2016). Diagnostic issues and controversies in DSM-5: Return of the false positives problem. Annual Review of Clinical Psychology, 12, 105–132.
West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. New York: Oxford University Press.
Whitfield, C. W., Cziko, A.-M., & Robinson, G. E. (2003). Gene expression profiles in the brain predict behavior in individual honey bees. Science, 302(5643), 296–299.
Widiger, T. A. (2007). Dimensional models of personality disorder. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 6(2), 79–83.
Widiger, T. A., & Lowe, J. R. (2008). A dimensional model of personality disorder: Proposal for DSM-V. Psychiatric Clinics of North America, 31(3), 363–378.