Borderline personality disorder (BPD) is a severe and complex disorder characterized by instability in interpersonal relations, behavior, and emotions (American Psychiatric Association, 2013). Findings have suggested that the defining and associated features of BPD are best characterized by a phenotype consisting of emotion dysregulation (ED), relational dysfunction (often referred to as relational disturbance or instability), and impulsivity (Gunderson, Reference Gunderson2007; Links, Helsegrave, & van Reekum, Reference Links, Heslegrave and van Reekum1999; McCloskey et al., Reference McCloskey, New, Siever, Goodman, Koenigsberg, Flory and Coccaro2009). Among these features, ED, involving difficulty modulating emotional responding, appears particularly central to many of the serious behavioral difficulties associated with BPD, including nonsuicidal self-injury (NSSI), suicidal behavior, and self-damaging impulsive behaviors (e.g., substance use, risky sexual behavior, and recklessness). Therefore, understanding how ED develops and crystallizes into a core feature of BPD should be considered a high priority.
A clearer understanding of the developmental processes contributing to ED could illuminate possible targets for preventative and efficient, targeted treatment interventions. Prevention, in turn, has the potential to forestall the escalation of serious behavioral problems, functional impairments, and heavy utilization of health and mental health resources associated with BPD. Within later adolescence and earlier adulthood, when BPD is established, understanding the processes contributing to and maintaining ED could help facilitate the development of more efficient, targeted treatments.
Over the past few decades, considerable progress has been made in our understanding of the developmental processes contributing to BPD and ED, as well as in the treatment of BPD and related clinical problem areas among children, adolescents, and adults. Consistent with a developmental psychopathology perspective on the development of personality disorders (Cicchetti, Reference Cicchetti2014), this review addresses theory and research on the possible biopsychosocial process operating at various levels of development to contribute to the development of ED and BPD. Contemporary developmental psychopathology models of BPD and ED have suggested that key underlying biologically based, heritable vulnerabilities (e.g., trait impulsivity) manifest in a variety of ways throughout development (e.g., attention-deficit/hyperactivity disorder, early conduct, and mental health problems) and confer vulnerability to key environmental risk factors (e.g., childhood maltreatment), which exacerbate the development of ED and eventual BPD (e.g., see Beauchaine, Hinshaw, & Bridge, Reference Beauchaine, Hinshaw and Bridge2019; Crowell, Beauchaine, & Linehan, Reference Crowell, Beauchaine and Linehan2009; Crowell, Kaufman & Beauchaine, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014).
BPD
Many of the defining features of BPD either include or are related to ED. The symptoms of BPD, represented in both the DSM-V (American Psychiatric Association, 2013) and the tenth edition of the ICD-10 (World Health Organization, 2016), represent marked instability in several life domains. According to the DSM-V, five of nine criteria are required for the diagnosis; these criteria largely capture difficulties in relationships, impulsive and self-damaging behavior, identity and mood disturbance, and affective instability. Similarly, in the ICD-10, BPD symptoms fall within the category “emotionally unstable personality disorder” and include disturbances in emotions, cognition, relationships, and behavior. ED or instability are central features of BPD within both diagnostic systems.
BPD is associated with considerable emotional suffering, severe behavioral problems, heavy utilization of health and mental health resources, and functional impairment. Approximately 70% to 80% of people with BPD have a history of NSSI and/or suicide attempts, and approximately 10% die by suicide (Black, Blum, Pfohl, & Hale, Reference Black, Blum, Pfohl and Hale2004; Gunderson, Reference Gunderson2001; Paris & Zweig-Frank, Reference Paris and Zweig-Frank2001; Skodol et al., Reference Skodol, Siever, Livesley, Gunderson, Pfohl and Widiger2002). The high prevalence of co-occurring disorders further underscores the considerable suffering associated with BPD and complicates treatment (as clinicians must be reasonably proficient in many area). The vast majority of patients with BPD have at least one additional diagnosis (Fyer, Frances, Sullivan, Hurt, & Clarkin, Reference Fyer, Frances, Sullivan, Hurt and Clarkin1988; Zanarini et al., Reference Zanarini, Frankenberg, Duo, Sickel, Trisha, Levin and Reynolds1998). Studies have suggested that up to two-thirds of patients with BPD have substance use problems, over half have PTSD, and well over half have a co-occurring anxiety disorder (up to 88% in one study of inpatients; Zanarini et al., Reference Zanarini, Frankenberg, Duo, Sickel, Trisha, Levin and Reynolds1998; Zarini, Frankenburg, Hennen, Reich, & Silk, Reference Zanarini, Frankenburg, Hennen, Reich and Silk2004). Major depressive disorder is particularly common, affecting approximately 80% of people with BPD (Lieb, Zanarini, Schmahl, Linehan, & Bohus Reference Lieb, Zanarini, Schmahl, Linehan and Bohus2004; Zanarini et al., Reference Zanarini, Frankenberg, Duo, Sickel, Trisha, Levin and Reynolds1998, Reference Zanarini, Frankenburg, Hennen, Reich and Silk2004). Consistent with this complex clinical picture and high prevalence of self-damaging behaviors, people with BPD heavily utilize emergency, inpatient, outpatient psychiatric, and primary care services, costing an average of $12,000–$30,000 per patient per year (in USD; Comtois, Elwood, Holdcraft, Smith, & Simpson, Reference Comtois, Elwood, Holdcraft, Smith and Simpson2007; Linehan & Heard, Reference Linehan, Heard, Miller and Magruder1999). The prevalence of BPD (1.4% to 5.9%) also is comparable to or higher than that of other costly disorders, such as schizophrenia and bipolar disorder (Lenzenweger, Lane, Loranger, & Kessler, Reference Lenzenweger, Lane, Loranger and Kessler2007).
Although rates of remission and recovery (approximately 60% at 16 years after hospital discharge; Zanarini, Frankenburg, Reich, & Fitzmaurice, Reference Zanarini, Frankenburg, Reich and Fitzmaurice2012) suggest a prognosis that is more hopeful than previously thought, substantial, and persistent functional impairment persists among a majority of BPD patients (Biskin, Reference Biskin2015). A sizable proportion of BPD patients experience impairments in social and occupational functioning requiring public assistance, including psychiatric disability support and Social Security disability income benefits (Bateman & Fonagy, Reference Bateman and Fonagy2003; Bateman & Fonagy, Reference Bateman and Fonagy2008; Grant et al., Reference Grant, Chou, Goldstein, Huang, Stinson, Saha and Ruan2008; Jørgensen et al., Reference Jørgensen, Kjølbye, Freund, Bøye, Jordet and Andersen2009; Koons et al., Reference Koons, Chapman, Betts, O'Rourke, Morse and Robins2006; Soloff & Chiappetta, Reference Soloff and Chiappetta2017; Zanarini, Jacoby, Frankenburg, Reich, & Fitzmaurice, Reference Zanarini, Jacoby, Frankenburg, Reich and Fitzmaurice2009). In one longitudinal study of patients with BPD, 69.4% fell within the range of poor psychosocial functioning at baseline assessment, and this proportion only reduced to 51.2% at the 8-year mark of the study (Soloff & Chiappetta, Reference Soloff and Chiappetta2017). In another study examining long-term follow-up data following mentalization-based treatment, 54% of patients remained within the range of poor psychosocial functioning 5 years following treatment, despite findings supporting the possible efficacy of mentalization-based treatment (Bateman & Fonagy, Reference Bateman and Fonagy2008). Given the likely role of ED in many of the behavioral and functional difficulties associated with BPD, understanding the developmental trajectory of ED will hopefully illuminate ways to curtail persistent functional impairment and suffering.
Emotion Regulation and Dysregulation
Understanding the specific characteristics and development of ED in BPD requires a broader framework in which to conceptualize emotion regulation and dysregulation. Definitions of emotion tend to emphasize that emotions (a) have several components, including cognitive, physiological, and behavioral features; (b) evolved over millennia to motivate adaptive behavior and facilitate social cohesion in a manner that maximized survival; and (c) generally arise in motivationally relevant contexts, such as contexts that are relevant to an individual's goals (e.g., the goal of survival, connecting with others, maintaining social standing, and so forth; Gross, Reference Gross and Gross2014; Lazarus, Reference Lazarus1991; Mauss, Levenson, McCarter, Wilhelm, & Gross, Reference Mauss, Levinson, McCarter, Wilhelm and Gross2005; Scherer, Schorr, & Johnstone, Reference Scherer, Schorr and Johstone2001). When a friend cancels plans, for example, an individual with BPD who is particularly sensitive to rejection or abandonment might respond with intense fear, consisting of physiological changes (e.g., alterations to heart rate and hormonal, neurotransmitter, and sympathetic nervous system responding), appraisals of the situation as disastrous or threatening (e.g., “He hates me.” “He doesn't want to be my friend.”), and actions functioning to prevent perceived negative outcomes (e.g., texting or calling the friend repeatedly, seeking reassurance). The situation may be especially relevant to the individual's goals if he is afraid to be alone, highly values the relationship, or historically has experienced severe rejection or abandonment.
Emotion regulation entails alterations to these different aspects of the multicomponent emotion system, involving changes in the form, frequency, experience, or expression of emotions (Beauchaine, Reference Beauchaine2015; Gross, Reference Gross1998, Reference Gross and Gross2014; Thompson, Reference Thompson1994). Emotion regulation also can occur at various stages of the emotion generative process. The process model of emotion regulation (Gross, Reference Gross1998), for example, indicates which types of emotion regulation processes occur at different points in the sequence of events related to an emotional response. The individual could engage in situation selection by deciding not to put himself in a position where a friend might cancel plans by avoiding the friend or mitigating the situation by asking if they can reschedule for the next evening. If the situation arises, he could decide to modify his attention by distracting from the situation or related emotions and focusing on other activities (e.g., binge watching a favorite show) or focusing on less emotionally evocative aspects of the situation (e.g., by paying attention when the friend says he wants to get together net week). He could also engage in cognitive reappraisal (e.g., “He doesn't hate me; he's just busy. I have a great opportunity to catch up on my show.”). Later in the chain of events, he might use response modulation strategies (e.g., diaphragmatic breathing or progressive muscle relaxation) to dampen the physiological arousal associated with fear of abandonment. He also might modify his emotional expressive behavior by changing his facial expression or body language. Other options are to avoid reaching out repeatedly or asking for reassurance and engaging in behavior that is less compatible with or opposite to fears of loneliness (e.g., “opposite action”; Linehan, Reference Linehan1993b, Reference Linehan2015).
An extended version of this process model (Gross, Reference Gross and Gross2014; Sheppes, Suri, & Gross, Reference Sheppes, Suri and Gross2015) encapsulates the sequence of emotion regulation into three key processes: (a) identification, involving the recognition and identification of an emotional response; (b) selection, involving the selection of a strategy to regulate emotions; and (c) implementation, involving the implementation of an emotion regulation strategy. Within each stage, emotion regulation problems can be characterized as aberrations in the perception or valuation (cost-benefit analysis) of emotional events, experiences, or potential regulation strategies. Emotion regulation difficulties could also come in the form of problems with the selection or implementation of regulation strategies.
Some researchers have operationalized ED in terms of the characteristics of emotional responding (e.g., overly intense or prolonged responses), whereas others have defined ED primarily in terms of deficits in the ability to use explicit emotion regulation strategies. Beauchaine (Reference Beauchaine2015), for example, has distinguished emotion regulation from dysregulation, defining ED as the presence of emotional responses that are too prolonged, intense, labile, or situationally inappropriate, or maladaptive behaviors related to emotional states. Within this framework, evidence for ED would consist of measures of the intensity, duration, or changeability of emotional responses in particular contexts. The client who feels intensely distraught for the whole evening after a friend has canceled plans would be considered emotionally dysregulated by virtue of maladaptively intense and prolonged emotional responding. Other indications of ED might include the client repeatedly texting the friend or repeatedly seeking reassurance, expressing hostility or anger, or self-injuring. Gratz and Roemer (Reference Gratz and Roemer2004), in contrast, have framed ED as difficulty with the use of key, explicit emotion regulation tasks or behaviors. Within this model, ED consists of difficulties in one or more of the following areas: emotional awareness and clarity, the ability to accept emotional experiences and inhibit impulsive behavior, and the capacity to regulate emotions. Similarly, another conceptualization of ED, advanced in the context of dialectical behavior therapy (DBT), is “the inability, even when one's best efforts are applied, to change in a desired way emotional cues, experiences, actions, verbal responses, and/or nonverbal expressions under normative conditions” (Linehan, Bohus, & Lynch, Reference Linehan, Bohus, Lynch and Gross2007, p. 583). Similar to Beauchaine's (Reference Beauchaine2015) definition, this conceptualization emphasizes the importance of goals (i.e., “in a desired way”), and similar to Gratz and Roemer (Reference Gratz and Roemer2004), emphasis is placed on particular emotion regulation capacities or skills.
For the purposes of the present paper, I conceptualize ED broadly as a deficit in the implicit (e.g., through physiological homeostatic mechanisms) or explicit (e.g., through effortful emotion regulation) regulation of the emotion system that hampers the individual's ability to engage in effective goal- or value-oriented behavior. Persons with BPD, therefore, may evidence ED through indicators of heightened or prolonged emotional responding, if such responding impedes value- or goal-directed action or contributes to suffering. Other indicators of ED would include deficits in the skills needed to select or implement effective emotion regulation strategies (i.e., deficits in explicit emotion regulation).
The distinction between an overly intense or prolonged emotional responding and an individual's capacity to effectively regulate emotions has important developmental implications and is likely to continue to be a subject of debate. One model is that ED in BPD is essentially a behavioral skill deficit fostered by the invalidating overlaid on an emotionally vulnerable temperament (as suggested by Linehan's original theory; Linehan, Reference Linehan1993a). Alternatively, the situation may be more complex, with ED consisting of both characteristics of emotional responding and a deficit in the capacities needed to regulate emotions (as suggested by Beauchaine, Reference Beauchaine2015; Crowell et al., Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014). In both cases, developmental processes and interventions that encourage the development of emotion regulation capacities or skills would be important. With the more complex model, however, it would also be important to examine how various psychosocial and biological processes shape how emotional response characteristics (e.g., high reactivity and sensitivity, or slow return to baseline), rather than simply coping or emotion regulation behaviors, develop over time.
Biosocial Developmental Theories of BPD
Among theories of the development of BPD, Linehan's (Reference Linehan1993a) biosocial theory most strongly emphasizes ED. In her early treatment development efforts, Linehan (Reference Linehan1993a) discovered that a substantial proportion of highly suicidal patients met criteria for BPD. Many of the hallmark interpersonal, emotional, and behavioral problems accompanying BPD presented during treatment, making it difficult to proceed with standard cognitive or behavioral interventions, which largely focused on changing thoughts and behavior and building skills. Patients with BPD (who frequently had a history of pervasive invalidation or maltreatment) often experienced the change-oriented focus of cognitive–behavioral therapy as invalidating, became emotionally dysregulated, and withdrew; became critical or hostile toward the therapist; or dropped out. To address these problems, Linehan (Reference Linehan1993a) systematically incorporated into this emerging treatment the balance and synthesis of acceptance and change-oriented therapeutic styles and strategies within a dialectical theoretical framework. She also used existing theory and research on BPD to assemble a working model, the biosocial theory, to better understand the development and maintenance of BPD. Along with dialectical and behavioral theory, the biosocial theory became the basis for a comprehensive cognitive–behavioral treatment approach (Linehan, Reference Linehan1993a, Reference Linehan1993b, Reference Linehan2015) emphasizing the role of ED in the development and maintenance of BPD.
Within the biosocial theory, the transaction of a biologically based vulnerability to emotions (emotion vulnerability) with an invalidating rearing environment contributes to the development of ED. Emotion vulnerability consists of a combination of a low threshold for emotional elicitation (emotional sensitivity), intense emotional responses (emotional reactivity), and a prolonged return to baseline emotional responding (slow return to baseline). The invalidating environment consists of the indiscriminate rejection of the child's communication of internal experiences (emotions, thoughts, and sensations), intermittent reinforcement of emotional escalation (e.g., inconsistent provision of attention and support that is sometimes contingent on extreme emotional expression), and the oversimplification of the ease of problem solving. The biosocial model is a transactional model, whereby the child's emotional temperament influences the invalidating environment, and vice versa. High emotionality elicits invalidation, and invalidation exacerbates high emotionality and does not provide the type of emotional socialization needed to develop core emotion regulation capacities. Over time, the individual develops ED, and various behavioral problems characteristic of BPD emerge (e.g., NSSI, suicide attempts, substance use, and harmful impulsive behaviors) and are negatively reinforced and maintained by the reduction of intolerable emotional arousal and/or changes in the social environment. Within this framework, many of the serious, harmful behaviors (e.g., NSSI, suicidal behavior, reckless impulsivity, drug and alcohol use, etc.) associated with BPD occur in response to or function to reduce ED (Linehan, Reference Linehan1993a). For example, NSSI and other problem behaviors characteristic of BPD often result in negative or positive reinforcement through the up- or downregulation of emotions (Brown, Comtois, & Linehan, Reference Brown, Comtois and Linehan2002; Chapman, Gratz, & Brown, Reference Chapman, Gratz and Brown2006; Kleindienst et al., Reference Kleindienst, Bohus, Ludaescher, Limberger, Kuenkele, Ebner-Priemer and Schmahl2008; Linehan, Reference Linehan1993a; Nock & Prinstein, Reference Nock and Prinstein2004). These behaviors also further contribute to disruptions in functioning, relationships, and so forth.
At the time of the publication of Linehan's (Reference Linehan1993a) biosocial theory, the literature on emotions and emotion regulation in BPD was scant. Even the broader literature on emotions and emotion regulation was in its infancy until the early 1990s (Gross, Reference Gross1998, Reference Gross and Gross2014; Thompson, Reference Thompson1994). Much of the developmental literature on psychopathology focused on syndromes of disinhibition (Patterson & Newman, Reference Patterson and Newman1993), such as attention-deficit/hyperactivity disorder (ADHD), and the developmental factors influencing the progression from child conduct problems to adult antisocial personality disorder (e.g., coercive behavioral processes; Patterson, DeBaryshe, & Ramsey, Reference Patterson, DeBaryshe and Ramsey1989). The biosocial theory was, therefore, based primarily on clinical observations, findings on the enduring effects of childhood maltreatment, and existing emotion and behavioral science. Since then, the literature on the development of psychopathology more broadly, and BPD and ED more specifically, has proliferated, providing an expanded picture of the potential factors contributing to ED in BPD.
Crowell et al.’s (Reference Crowell, Beauchaine and Linehan2009) biosocial developmental model of BPD
Crowell et al. (Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) have proposed an expanded biosocial developmental model of BPD, incorporating developments in research on developmental psychopathology as well as biological and psychosocial factors associated with BPD. Compared with Linehan's (Reference Linehan1993a) biosocial theory, which proposes that emotion vulnerability is the key biologically based vulnerability factor, Crowell et al. (Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) more strongly emphasize the role of trait impulsivity, a heritable genetic predisposition that confers vulnerability to various childhood behavioral disorders (e.g., oppositional defiant disorder and ADHD). Crowell et al. (Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014; see also Beauchaine, Hinshaw, et al., Reference Beauchaine, Hinshaw and Bridge2019) propose that trait impulsivity moderates the effects of invalidating or adverse rearing contexts, such that those with high trait impulsivity are particularly vulnerable to developing ED and related behavioral problems (e.g., NSSI and suicide attempts) in these contexts. Instead of necessarily serving as a biologically based vulnerability factor, within Crowell et al.’s (Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) model, emotion vulnerability and dysregulation develop in the context of the invalidating rearing environment, primarily through coercive conflict-escalation processes, the intermittent reinforcement of emotional lability, abuse, neglect, and other adverse circumstances.
This expanded biosocial developmental model has five key tenets, some of which have garnered additional support since the model was articulated about 10 years ago. First, as mentioned, trait impulsivity is a key biologically based and heritable vulnerability factor for the development of BPD and other externalizing behavioral problems sometimes referred to as syndromes of disinhibition (e.g., oppositional defiant disorder, ADHD, conduct disorder, eventual adult antisocial personality disorder, etc.). Second, adverse, invalidating environmental contexts shape the development of emotional vulnerability (referred to as “emotional lability”), which can be considered both biologically based and highly susceptible to environmental influence. Third, trait impulsivity continues to interact with environmental risk factors to exacerbate ED and disrupt normative social and emotional development. Among youth who are higher in impulsivity, various environmental risks may have a more potent effect on their development of behavioral, social, and emotional difficulties, compared with those who are lower in trait impulsivity. Fourth, in adolescence, at-risk individuals begin to engage in maladaptive coping and problematic behaviors that pose risk for BPD, further disrupt the development of emotion regulation capacities, and carry their own negative interpersonal and intrapersonal consequences. In particular, NSSI and suicide attempts are considered key precursors to the development of adult BPD. These and other behaviors commonly associated with BPD (e.g., drug and alcohol use, reckless, and self-damaging behavior) become entrenched due to repeated reinforcement in the form of emotion regulation and/or changes in the social environment. Fifth, in adolescence, the presence of ED more generally and NSSI and suicidal behavior more specifically continue to contribute to risk for the development of adult BPD.
The Development of ED in BPD
As described by Crowell et al. (Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014), the development of BPD more broadly and ED more specifically, likely follows a complex, heterotypic trajectory (Belsky et al., Reference Belsky, Caspi, Arseneault, Bleidorn, Fonagy, Goodman and Moffitt2012). Key heritable factors (e.g., trait impulsivity and possibly neuroticism) appear to elevate the risk of a range of childhood emotional and behavioral problems and may contribute to vulnerability to the adverse environmental circumstances that foster the development of ED. Key signs of worsening ED often emerge in adolescence, where underlying dispositional factors (such as impulsivity) may increase vulnerability to alterations in the development of the social cognitive capacities needed to effectively regulate emotions. During this period, worsening ED and related interpersonal and behavioral problems, along with various peer and other environmental influences fostering maladaptive coping (e.g., through peer and possibly parental modeling processes and influences), elevate the risk of NSSI and suicidal behavior, which further contribute to the development of BPD. In adulthood, where BPD tends to show more temporal stability (compared with adolescence and childhood), findings have shown evidence of a variety of biological, cognitive, and behavioral characteristics of ED that reflect some of these earlier developmental processes.
Key genetic and heritable vulnerability factors
Heritable factors may play an important role in the development of BPD and ED, predisposing individuals to be vulnerable to the effects of adverse environments in childhood and adolescence. In turn, environmental risk factors may influence the expression of particular genetic and biological propensities (e.g., see Amad, Ramoz, Thomas, Jardri, & Gorwood, Reference Amad, Ramoz, Thomas, Jardri and Gorwood2014, regarding a “plasticity” theory). As mentioned, Crowell et al. (Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) have suggested that impulsive children are particularly likely to develop behavioral and emotional difficulties in adverse or invalidating environments. Heritable factors also may operate in a transactional manner, whereby the heritable trait and its manifestations influence and are exacerbated by environmental factors. Highly impulsive children may be particularly difficult to parent, presenting behavioral challenges through inattentiveness, hyperactivity, noncompliance, and so forth. Parents who share with their children a propensity toward impulsivity may be prone to use coercive tactics to control their children's behavior as well as to erratically and intermittently reinforce emotional escalation. Invalidation, abuse, or neglect, along with parental modeling of maladaptive behaviors (e.g., poor impulse control and emotional lability) may further exacerbate the child's existing propensities toward heightened emotionality and impulsivity. Heritable factors may also influence alterations in the development of social cognitive capacities needed to regulate emotions. Some have suggested, for example, that impulsivity plays a key role in the development of neuropsychological deficits in executive functioning that influence later conduct problems (Caspi, Henry, McGee, Moffit, & Silva, Reference Caspi, Henry, McGee, Moffitt and Silva1995).
Findings from several studies have suggested an important role of heritable factors in the propensity to develop BPD and ED. First-degree relatives of individuals with BPD are several times more likely to have BPD than the general population (Baron, Gruen, Asnis, & Lord, Reference Baron, Gruen, Asnis and Lord1985; Links, Steiner, & Huxley, Reference Links, Steiner and Huxley1988; Zanarini, Gunderson, Marino, Schwartz, & Frankenburg, Reference Zanarini, Gunderson, Marino, Schwartz and Frankenburg1988), and twin studies have suggested moderate heritability for BPD (.40 to around .70). Torgersen et al. (Reference Torgersen, Lygren, Oien, Skyre, Onstad, Edvardsen and Kringlen2000), for example, found a heritability for BPD of .69 among 221 twin pairs assessed for the presence of personality disorders. Another study of 5,000 twins in the Netherlands, Belgium, and Australia found consistent heritability estimates of .42 using a self-report measure of BPD features (Distel et al., Reference Distel, Trull, Derom, Thiery, Grimmer, Martin and Boomsma2008). Another study, focused on Norwegian twins, found evidence that BPD features are best accounted for by a model including a moderately heritable overarching latent “BPD” factor (Reichborn-Kjennerud et al., Reference Reichborn-Kjennerud, Ystrom, Neale, Aggen, Mazzeo, Knudsen and Kendler2013), with heritable and unique environmental factors accounting for 55% and 45% of this factor, respectively. Within this model, two subfactors included an interpersonal factor, which had low heritability (2.2%), and an affective instability factor, with heritability of 29.3%. In addition, the BPD impulsivity criterion, consisting of self-inflicted injuries and impulsive, self-damaging behaviors, received the majority of its genetic influence (which was 44.6%) from what the authors referred to as “criterion-specific” factors not accounted for by the heritable influences related to the overall BPD factor. These findings support the existence of a highly heritable impulsivity dimension. In contrast, unique environmental factors seem to account for the majority of the variability in the interpersonal and affective symptoms of BPD. Together, these findings provide at least indirect support for the notion that impulsivity is a key heritable risk factor for BPD, and that aspects of the rearing environment substantially influence the development of ED and interpersonal dysfunction.
Studies also have examined potential heritable vulnerabilities to BPD within the framework of normal personality theory, largely anchored in the Five-Factor Model (FFM) of personality, consisting of openness, conscientiousness, extraversion, agreeableness, and neuroticism. Distel et al. (Reference Distel, Trull, Willemsen, Vink, Devon, Lynsey and Boomsma2009), for example, examined FFM traits in relation to BPD in a large study of twins from the Netherlands, Belgium, and Australia. The overall heritability estimate for BPD was .45, and findings indicated that all of the genetic variance in BPD was shared by FFM traits. Among associations of FFM traits with BPD, neuroticism had the strongest association, accounting for 45% of the variability in BPD traits, and the additive genetic correlation of BPD with neuroticism was the highest among the FFM traits (r = .95). It is interesting to note that nearly one-third of the variance in unique environmental effects on BPD was not shared with the FFM traits. The authors concluded that heritable vulnerabilities conferred by high neuroticism and low conscientiousness and agreeableness might make people vulnerable to some of the unique environmental factors that play a role in the development of BPD.
These findings, in conjunction with findings on the heritability and association of trait impulsivity with BPD and possible precursors to BPD, suggest that heritable vulnerability to BPD and ED may consist of a combination of trait impulsivity and neuroticism. Findings have indicated that both neuroticism and impulsivity are at least moderately heritable. One study examining heritability for different forms of impulsivity found heritability of 45%–50% for impulsivity assessed by trait measures and 28%–55% for behavioral indicators of impulsivity (e.g., delay discounting, risk taking, and behavioral disinhibition; Bevilacqua & Goldman, Reference Bevilacqua and Goldman2013). Similarly, findings on the heritability of the FFM traits have generally indicated significant and moderate heritability estimates for neuroticism (ranging from 15% to 50%; Power & Pluess, Reference Power and Pluess2015; Vukasović & Bratko, Reference Vukasović and Bratko2015). Children with high trait impulsivity and neuroticism may be particularly vulnerable to the types of environmental influences that shape the development of ED and eventual BPD. Further, the high co-occurrence of mood and anxiety disorders with BPD suggests a heterotypic model in that neuroticism may manifest in different ways at various stages of development, culminating (with the right mix of experiences) in disorders characterized by anxiety and mood disturbance. Therefore, it seems likely that, when it comes to underlying heritable vulnerabilities and their manifestations throughout development, BPD straddles the internalizing and externalizing dimensions.
Adverse rearing environments
Adverse childhood environments have received considerable attention as key environmental factors that influence the development of BPD and ED (Calkins & Hill, Reference Calkins, Hill and Gross2007; Eisenberg et al. Reference Eisenberg, Fabes, Shepard, Guthrie, Murphy and Reiser1999). The literature on childhood maltreatment in relation to BPD is substantial, so I will highlight a few key findings here. Adults with BPD tend to report high rates of childhood adversity, suggesting an important role for trauma or abuse, at least in some cases. Rates of reported past childhood sexual abuse among persons with BPD are high (Zanarini et al., Reference Zanarini, Yong, Frankenburg, Hennen, Reich, Marino and Vujanovic2002). Findings also have suggested that persons with BPD report a higher prevalence of childhood sexual (Paris, Zweig-Frank, & Guzder, Reference Paris, Zweig-Frank and Guzder1994; Yen et al., Reference Yen, Shea, Battle, Johnson, Zlotnick, Dolan-Sewell and Zanarini2002) and physical abuse (Paris et al., Reference Paris, Zweig-Frank and Guzder1994) than those with other personality disorders. A chart-review study found a prevalence of past sexual abuse of 50% among inpatient adolescents with BPD (Westen, Ludolph, Misle, Ruffins, & Block, Reference Westen, Ludolph, Misle, Ruffins and Block1990). Given these findings, some have speculated that BPD represents complex PTSD, suggesting that past trauma drives the ED characteristic of BPD (Cloitre, Garvert, Weiss, Carlson, & Bryant, Reference Cloitre, Garvert, Weiss, Carlson and Bryant2014; Ford & Courtois, Reference Ford and Courtois2014; Trippany, Helm, & Simpson, Reference Trippany, Helm and Simpson2006; van der Kolk, Pelcovitz, Roth, & Mandel, Reference van der Kolk, Pelcovitz, Roth and Mandel1996). Other researchers have suggested that past abuse or trauma are neither necessary nor sufficient for a BPD diagnosis (Zanarini et al., Reference Zanarini, Williams, Lewis, Reich, Vera, Marino and Frankenburg1997). Although it appears that childhood adversity is unusually common in BPD, up to 50% of adults with BPD do not report childhood sexual abuse, and up to 54% do not meet criteria for PTSD (Lieb et al., Reference Lieb, Zanarini, Schmahl, Linehan and Bohus2004). It is likely, therefore, that trauma and severe childhood adversity plays an important role for some people, and that characteristics of childhood environments likely transact with biological and other factors in complex ways to influence the development of ED and BPD (Chapman, Hope, & Turner, Reference Chapman, Hope, Turner, Lejuez and Gratzin press; Crowell et al., Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014).
Some research has examined the types of childhood maltreatment that might be particularly associated with BPD. One study, for example, examined BPD features among maltreated and nonmaltreated children (Hecht, Cicchetti, Rogosch, & Crick, Reference Hecht, Cicchetti, Rogosch and Crick2014). Findings indicated that greater chronicity of maltreatment was associated with more severe BPD features. Further, the number of developmental periods in which maltreatment occurred predicted whether individuals fell within a high-risk for the BPD group (those with high scores on three or more categories of symptoms; Hecht et al., Reference Hecht, Cicchetti, Rogosch and Crick2014). Maltreatment status was associated with higher scores on all categories of BPD features, suggesting that maltreatment may not uniquely influence the development of particular types of symptoms, such as emotional instability or ED-related features.
Some findings also have suggested that ED may play an important mediating role in the association between childhood experiences and the later development of BPD. In a study examining the influence of mothers’ personality disorder symptoms on child psychopathology, Kaufman, Puzia, Mead, Crowell, and Beauchaine (Reference Kaufman, Puzia, Mead, Crowell and Beauchaine2017) found that children's difficulty regulating emotions partially mediated the association of a composite of maternal BPD/antisocial personality disorder symptoms on internalizing and externalizing symptoms a year later. In another study examining the association of childhood parental emotion socialization (reported retrospectively) with ED and self-harm among college students, emotional invalidation (specifically, punishment and neglect) predicted self-harm, and difficulty evaluating emotions (an aspect of difficulties with emotion regulation) mediated the relationship between these parental responses and self-harm (Buckholdt, Parra, & Jobe-Shields, Reference Buckholdt, Parra and Jobe-Shields2009). Another study examined the association of reports of childhood parental emotion socialization and trauma with current BPD features among college students, finding that difficulties regulating emotions mediated the association of emotion socialization and parental psychological control (but not childhood trauma) with BPD features (Hope & Chapman, Reference Hope and Chapmanin press). Some of this literature on parenting, trauma, abuse, and other childhood experiences is limited by its focus on adult reports of childhood experiences, and lack of examination of interactions of heritable and environmental factors.
Addressing some of these limitations, Belsky et al. (Reference Belsky, Caspi, Arseneault, Bleidorn, Fonagy, Goodman and Moffitt2012) examined the potential role of environmental and heritable factors in the development of BPD in a longitudinal cohort study, following twins (N = 1,116) from birth to age 12. Findings indicated that maltreatment in childhood more strongly predicted BPD features at 12 years among children with family histories of psychiatric disorder. Findings from this study painted a complex picture of the development of BPD and indicated that some of the unique environmental influences accounting for a proportion of the variance in BPD features (at least among those with high levels of BPD features) may include high maternal negative expressed emotion.
Some research on the neurobiological correlates of adverse childhood experiences has suggested potential mechanisms underlying a link between childhood adversity, ED, and BPD. As discussed later, adults with BPD appear to be prone to poor frontal regulation of the limbic system, whereby the connectivity of prefrontal regions with the amygdala appears reduced in BPD patients. Typically, increases in the activity of certain frontal areas (e.g., the ventrolateral and medial prefrontal cortex in particular) are associated with reductions in the activity of the amygdala (Lieberman et al., Reference Lieberman, Eisenberger, Crockett, Tom, Pfeifer and Way2007). Some findings, however, have suggested that persons exposed to significant early life stress may show a positive association of ventrolateral prefrontal and amygdala activation (Taylor, Eisenberger, Saxbe, Lehman, & Lieberman, Reference Taylor, Eisenberger, Saxbe, Lehman and Lieberman2006). If the prefrontal cortex typically acts as the “brakes” on the amygdala in normally developing individuals, the opposite may occur among those exposed to adverse childhood events. Further, animal research has suggested that the prefrontal cortex is particularly dense with cortisol receptors (Sanchez, Young, Plotsky, & Insel, Reference Sanchez, Young, Plotsky and Insel2000), and that stress increases dopamine activity in the prefrontal cortex (Roth, Tam, Ida, Yang, & Deutch, Reference Roth, Tam, Ida, Yang and Deutch1988) and may disrupt the communication between the prefrontal cortex and other brain regions (Liston, McEwen, & Casey, Reference Liston, McEwen and Casey2009). These findings indicate possible mechanisms for the lingering effects of childhood adversity on the emotion regulation system (Diamond, Reference Diamond2018). Studies examining brain structure and functioning in BPD, therefore, should incorporate measures of childhood adversity and clinical control groups with PTSD or past child abuse.
Emerging BPD and related behavioral problems in adolescence
Adolescence appears to be a particularly important period marking increased severity of ED and the emergence of some of the key behavioral problem areas associated with BPD (e.g., NSSI and suicide attempts). Increasingly complex social dynamics, academic demands, conflicts between autonomy and dependence, among other factors, likely test and shape emotion regulation capabilities during adolescence (Ahmed, Bittencourt-Hewitt, & Sebastian, Reference Ahmed, Bittencourt-Hewitt and Sebastian2015; Casey et al., Reference Casey, Jones and Hare2008). In an excellent review of the development of neurocognitive foundations of emotion regulation, Ahmed et al. have argued that the amygdala, various regions of the prefrontal cortex, and the connectivity of prefrontal and limbic regions all mature during adolescence (with some of these regions continuing to develop into young adulthood) and are directly relevant to the development of emotion regulation capacities (Ahmed et al., Reference Ahmed, Bittencourt-Hewitt and Sebastian2015). This pattern of development is particularly relevant to ED and BPD.
Findings have indicated that some of the cognitive and social cognitive capacities that facilitate emotion regulation show particularly steep development during adolescence (Blakemore & Robbins, Reference Blakemore and Robbins2012; Shaw et al., Reference Shaw, Kabani, Lerch, Eckstrand, Lenroot, Gogtay and Wise2008). Adults with BPD show deficits in a variety of neurocognitive domains, some of which are particularly important for effective emotion regulation (e.g., attention, cognitive flexibility, and planning; Ruocco, Reference Ruocco2005; Unoka & Richman, Reference Unoka and Richman2016). The onset of various internalizing and externalizing behavioral problems in adolescence (Lancefield, Raudino, Downs, & Laurens, Reference Lancefield, Raudino, Downs and Laurens2016) could indicate difficulties in the development of emotion regulation capacities, perhaps stemming from earlier and ongoing transactions of a vulnerable temperament with aspects of the rearing environment. In particular, the development of certain types of behavioral problems, such as NSSI, suicidal behavior, and substance use problems, might at least indirectly indicate that ED is beginning to crystallize. Therefore, it appears that (a) basic capacities and characteristics of brain functioning necessary for emotion regulation develop differently in those with BPD, and (b) such developmental differences begin to become more apparent in adolescence.
Findings on the trajectory of BPD symptoms in childhood, adolescence, and adulthood have further suggested continuity in these symptoms and indicated that adolescence may be a key period in which to identify and try to prevent the further development of ED. A study of the association of childhood psychopathology and the later development of BPD (the Pittsburgh Girls Study; N = 1,233; Stepp, Burke, Hipwell, & Loeber, Reference Stepp, Burke, Hipwell and Loeber2012) found that childhood ADHD and oppositional defiant disorder symptoms predicted poor impulse control, problems with self-regulation, and BPD symptoms in adolescence. These findings support the biosocial developmental model's (Crowell et al., Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) hypothesis that early impulse control problems are key indicators of a developmental trajectory toward increasing ED and BPD.
BPD symptoms in adolescence also predict later behavioral problems in early adulthood. The Children in the Community Study, for example, followed a diverse group of 800 children in various areas of New York over a 20-year period, examining the trajectory of adolescents experiencing symptoms of BPD (Cohen, Crawford, Johnson, & Kasen, Reference Cohen, Crawford, Johnson and Kasen2005). Some of the findings suggested that childhood sexual abuse predicted BPD but not other personality disorder symptoms in adolescence. BPD symptoms in early adolescence (mean age = 14) predicted BPD symptoms as well as lower academic and occupational attainment and greater psychosocial impairment in early adulthood (mean age = 33; e.g., lower global assessment of functioning and lower role function). Consistent with other research (Zanarini et al., Reference Zanarini, Frankenburg, Hennen, Reich and Silk2004), the Children in the Community Study found age-related declines in some BPD symptoms (Winograd, Cohen, & Chen, Reference Winograd, Cohen and Chen2008).
Some of these behavioral problems likely are influenced by and influence the development of ED and BPD. Adolescents with developing BPD symptoms, for example, may be particularly prone to substance use problems. Substance use, in turn, may have direct effects on brain development, further exacerbate psychosocial stressors (e.g., impairing functioning and contributing to academic problems and relationship discord), and worsen the trajectory of BPD symptoms (Bornovalova et al., Reference Bornovalova, Verhulst, Webber, McGue, Iacono and Hicks2018; Zanarini et al., Reference Zanarini, Frankenburg, Hennen, Reich and Silk2004). A recent study of female twins investigated genetic and environmental influences on the association of BPD features, major depressive disorder, and alcohol use disorders (AUD; Bornovalova et al., Reference Bornovalova, Verhulst, Webber, McGue, Iacono and Hicks2017). Following adolescents from age 14 through early adulthood (age 24), the findings of this study suggested that genetic and unique environmental factors accounted for the covariation of BPD with major depressive disorder and AUD over time, and that AUD predicted lower decline in BPD features. These findings suggest that at least some of the behavioral problems co-occurring with BPD may worsen the BPD symptom trajectory, possibly through their effects on brain development and exacerbation of ED and contextual factors (e.g., interpersonal problems and stress). Similarly, other longitudinal research with adults has indicated that the presence of SUDs, PTSD, and mood and anxiety disorders reduces the likelihood of remission from BPD (Zanarini et al., Reference Zanarini, Frankenburg, Hennen, Reich and Silk2004). Another recent longitudinal study examined self-harm and suicidal behaviors as predictors of the development of BPD 5 years later among adolescent psychiatric inpatients (Homan, Sim, Fargo, & Twohig, Reference Homan, Sim, Fargo and Twohig2017). Findings suggested that only suicide threats uniquely predicted BPD 5 years later (Homan et al., Reference Homan, Sim, Fargo and Twohig2017). As one of the key functions of emotions is to communicate, suicide threats may indicate a combination of high ED and deficits in communication skills, along with a learning history reinforcing emotional escalation.
Some studies have investigated whether some of the differences in the volume of certain brain areas that have been observed among adults with BPD are also observable among adolescents at risk of BPD (e.g., those engaging in self-injury) or with an early diagnosis of BPD. One study, for example, found that self-injuring adolescent girls demonstrated reduced volume in the insular cortex and right inferior frontal gyrus, compared with non-self-injuring controls (Beauchaine, Sauder, Derbridget, & Uyeji, Reference Beauchaine, Sauder, Derbidge and Uyeji2019). Another investigation compared the volume of prefrontal and orbitofrontal gray matter among adolescents with BPD versus psychiatric and nonpsychiatric controls (Brunner et al., Reference Brunner, Henze, Parzer, Kramer, Feigl, Lutz and Stieltjes2010). Findings indicated that adolescents with BPD showed reduced volume in the dorsolateral frontal gyrus and left orbitofrontal gyrus compared with nonpsychiatric controls, but that there were no differences in the volume of these or other regions between clinical controls and BPD participants (Brunner et al., Reference Brunner, Henze, Parzer, Kramer, Feigl, Lutz and Stieltjes2010). Takahashi et al. (Reference Takahashi, Chanen, Wood, Yücel, Tanino, Suzuki and McGorry2009) also examined insular cortex volume among adolescent participants with BPD and healthy controls and found no differences in volume between these groups. BPD participants, however, with a history of violent behavior over the past 6 months demonstrated a smaller bilateral insular volume, compared with BPD participants with no such history (Takahashi et al., Reference Takahashi, Chanen, Wood, Yücel, Tanino, Suzuki and McGorry2009). Taken together, these findings suggest that some of the brain differences associated with BPD may begin to emerge in a narrower fashion (differences in volume in fewer areas compared with adulthood) in adolescence compared with adulthood, that these brain differences may not be unique to BPD, and that the association of violence and brain volume and functioning should be further explored. Violence may be a more extreme indicator of high trait impulsivity and ED among adolescents at risk for BPD.
Self-injury and suicidal behavior
Models of the development of ED, BPD, and related problems (e.g., NSSI and suicidal behavior) have emphasized important transitions and transactions occurring in adolescence. The development of NSSI and suicidal behavior in early to middle adolescence, in particular, has received particular attention. Beauchaine, Hinshaw, et al. (Reference Beauchaine, Hinshaw and Bridge2019) have proposed a similar model as the biosocial developmental model of BPD to explain the development of NSSI and suicidal behavior among adolescents. In particular, children with impulsive temperaments are especially vulnerable to the effects of maltreatment, and are likely to have interpersonal difficulties and to be involved in peer contexts that support the development of severely maladaptive coping behaviors in early adolescence, such as NSSI (i.e., through modeling and contagion effects; Derbidge & Beauchaine, Reference Derbidge, Beauchaine, Lewis and Rudolph2014). NSSI emerges as an indicator of worsening ED and a way to regulate ED. In support of this model, findings from one study suggested that adolescents with histories of recent NSSI or suicide attempts (at least one episode in the past 6 months) may be particularly reactive to maternal behavior during discussions of conflictual topics (Crowell et al., Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014). In this study, adolescents with NSSI or suicide attempt histories demonstrated both behavioral and physiological reactivity to mothers’ behavior, but depressed adolescents only showed behavioral reactivity, suggesting a potentially unique biological vulnerability among self-injuring or suicidal adolescents. Findings from other research have suggested that children with ADHD are particularly vulnerable to developing self-inflected injury if they have had a history of childhood maltreatment (see Beauchaine, Hinshaw, et al., Reference Beauchaine, Hinshaw and Bridge2019, for a review; see also Derbidge & Beauchaine, Reference Derbidge, Beauchaine, Lewis and Rudolph2014; Guendelman, Owens, Galan, Gard, & Hinshaw, Reference Guendelman, Owens, Galan, Gard and Hinshaw2015; Swanson, Owens, & Hinshaw, Reference Swanson, Owens and Hinshaw2014).
Similarly, Miller and Prinstein (Reference Miller and Prinstein2019) have proposed that abnormal development of the biological stress response system influences the emergence of NSSI and suicidal behavior in adolescence. Within this model, normative changes in the biological stress response system, such as changes in the hypothalamic–pituitary–adrenal axis and sympathetic and parasympathetic nervous system reactivity, result in a period of enhanced sensitivity to stressors among adolescents. This may be particularly the case for interpersonal stressors, given the importance of peer relationships in adolescence and evidence of increased sensitivity to social rewards and punishments (Casey, Jones, & Hare, Reference Casey, Jones and Hare2008) and prefrontal brain response to social evaluation (Sommerville, Reference Somerville2013). Miller and Prinstein (Reference Miller and Prinstein2019) have suggested that abnormalities in the development of the stress response system (e.g., reduced connectivity of the prefrontal and limbic systems) may begin to appear in adolescence and influence the development of NSSI and suicidal behavior. Some studies, accordingly, have found that youth with suicide histories have demonstrated reduced amygdala and PFC connectivity in response to emotional stimuli, a finding that parallels some of the research on adult BPD (discussed later in this review; Johnston et al., Reference Johnston, Wang, Liu, Blond, Wallace, Liu and Blumberg2017). As mentioned earlier, structural brain differences associated with self-injury and BPD may begin to emerge in adolescence as well, albeit in a narrower fashion than among adults patients with BPD.
The Course of BPD
In this section, I briefly summarize some of the research on the course of BPD and BPD symptoms. As many of the studies have begun following participants with BPD at a mean age corresponding to young adulthood, this research is probably best viewed as a snapshot of how BPD progresses through young adulthood. In addition, although there have been many studies of the course of BPD, I will focus primarily on findings from the McLean Study of Adult Development, a large study of BPD inpatients, followed posthospitalization for several years (Zanarini, Frankenburg, Hennen, & Silk, Reference Zanarini, Frankenburg, Hennen and Silk2003). Findings from this research have suggested a fairly hopeful prognosis for many people with BPD, with findings from the 16-year point indicating that 99% of patients achieved at least 2 years of remission from the diagnosis of BPD, 60% of patients recovered, and 5% died by suicide (Zanarini et al., Reference Zanarini, Frankenburg, Reich and Fitzmaurice2012). Earlier studies based on this data set examining patterns of symptom remission have indicated that the dysregulated behavioral patterns that are likely related to ED tend to subside over time, but that enduring emotional suffering in the form of mood symptoms often persists. Some of the most persistent symptoms, even among people who no longer meet diagnostic criteria, included mood and emotional disturbances (Zanarini et al., Reference Zanarini, Frankenburg, Hennen and Silk2003). Although NSSI, suicidality, affective instability, and serious interpersonal instability subsided relatively rapidly, emotional symptoms such as chronic dysphoria and intense anger tended to remit more slowly (Zanarini et al., Reference Zanarini, Frankenburg, Reich, Silk, Hudson and McSweeney2007). Moreover, general impulsivity and recklessness also tended to remit more slowly over a 10-year period. This finding suggested that, while the severe, emotion-driven behaviors such as NSSI and suicide attempts appear to resolve fairly quickly (Zanarini et al., Reference Zanarini, Frankenburg, Reich, Silk, Hudson and McSweeney2007), more enduring tendencies toward impulsivity may persist and appear in various forms, again consistent with heterotypic models highlighting trait impulsivity as an enduring predisposition to BPD (e.g., Crowell et al., Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014).
Characteristics of ED in Adults With BPD
With these findings on the developmental factors and periods in mind, it is helpful to consider what characterizes ED among adults with BPD. What is the net result of the potential genetic, biological, and psychosocial factors operating at various points in development? Findings across several studies over the past couple of decades have begun to clarify the nature of ED in BPD. The summary below is based on the conceptualization of emotions as systemic responses to motivationally relevant situations, consisting of cognitive, physiological, and behavioral components. ED, therefore, consists of aberrations in cognitive, physiological, or behavioral responding that impede the individual's ability to achieve goals or engage in effective action in particular contexts. I will focus here on cognitive, behavioral, and physiological factors that are indicative of ED in BPD, rather than on the literature on emotional responding in BPD, which has been reviewed a few times in the past several years (Carpenter & Trull, Reference Carpenter and Trull2013; Dixon-Gordon, Haliczer, & Conkey, Reference Dixon-Gordon, Haliczer, Conkey, Beauchaine and Crowellin press; Rosenthal et al., Reference Rosenthal, Gratz, Kosson, Cheavens, Lejuez and Lynch2008).
Cognitive aspects of ED
Cognitively, persons with BPD tend to have difficulty with emotional clarity and awareness (Leible & Snell, Reference Leible and Snell2004), identifying and describing emotions (Wolff, Stiglmayr, Bretz, Lammers, & Auckenthaler, Reference Wolff, Stiglmayr, Bretz, Lammers and Auckenthaler2007; Derks, Westerhof, & Bohlmeijer, Reference Derks, Westerhof and Bohlmeijer2016), and a tendency toward less specific differentiation among negative emotions and lower emotional granularity (i.e., the tendency to represent emotions in a nonspecific manner; Suvak et al., Reference Suvak, Litz, Sloan, Zanarini, Barrett and Hofmann2011). Effective emotion regulation often requires and is facilitated by clarity and awareness of an individual's emotional state. Not knowing what specific emotion one is feeling would make it difficult to select adaptive emotion regulation strategies. Persons with BPD also may be especially prone to appraise certain situations or emotional experiences as excessively negative or intolerable or as having disastrous consequences. Linehan (Reference Linehan1993a) has suggested that some persons with BPD are phobic of emotions. Clinically, persons with BPD often describe their emotions as overwhelming, out of control, and intolerable.
Other cognitive characteristics of ED include difficulty regulating attention to emotionally evocative events. Linehan (Reference Linehan1993a) has proposed that persons with BPD may have difficulty disengaging their attention from emotional material. Some research indicating that people with BPD are particularly prone to rumination, for example, suggests difficulty disengaging from emotional events (Abela, Payne, & Moussaly, Reference Abela, Payne and Moussaly2003; Smith, Grandin, Alloy, & Abramson, Reference Smith, Grandin, Alloy and Abramson2006), and the cascade model of BPD posits that NSSI is particularly likely when rumination has amplified negative emotional experiences to an intolerable level (Selby & Joiner, Reference Selby and Joiner2009). Findings from some studies have suggested the presence of an attentional bias toward negative and BPD-related stimuli (words; Kaiser, Jacobs, Domes, & Arntz, Reference Kaiser, Jacob, Domes and Arntz2017); this possible bias may amplify emotional responding and disrupt the ability of persons with BPD to disengage or distance themselves from emotional events. The findings on attentional biases in BPD, however, are complex, sometimes inconsistent, and may depend on the laboratory paradigms and stimuli used (Kaiser et al., Reference Kaiser, Jacob, Domes and Arntz2017).
Physiological aspects of ED
Findings on physiological regulation of emotions in BPD paint a fairly complex picture. Activity of the parasympathetic nervous system, as indexed by respiratory sinus arrhythmia (RSA), has been theorized to be an indication of emotion regulation (Porges, Reference Porges2001), with higher RSA indicating greater vagal influence on heart rate variability. Research on parasympathetic reactivity in BPD remains fairly limited. Some studies have found decreasing RSA among BPD patients in response to emotionally evocative films relative to nonpatient (Austin, Riniolo, & Porges, Reference Austin, Riniolo and Porges2007) or clinical controls (Fitzpatrick & Kuo, Reference Fitzpatrick and Kuo2015). Another study did not find differences in RSA activity among BPD participants in response to a laboratory emotional stressor, compared with nonclinical controls (Weinberg, Klonsky, & Hajcak, Reference Weinberg, Klonsky and Hajcak2009).
Brain imaging research has examined the connectivity of activity in the prefrontal and limbic regions, as deficient inhibitory influence of certain prefrontal areas on the amygdala could be considered an aspect or mechanism of ED. Some findings have indicated that participants with BPD show a combination of heightened activation of the amygdala in response to emotional stimuli (Donegan et al., Reference Donegan, Sanislow, Blumberg, Fullbright, Lacadie, Skudlarski and Wexler2003; Hazlett et al., Reference Hazlett, Zhang, New, Zelmanova, Goldstein, Haznedar and Chu2012; particularly the left amygdala) with lower activation of the dorsolateral prefrontal cortex (Schulze, Schmahl, & Niedtfeld, Reference Schulze, Schmahl and Niedtfeld2016). In contrast, a meta analysis of 11 studies comparing BPD patients to healthy controls showed greater activation in the insula and posterior cingulate cortex and lower activation in the amygdala, subgenual anterior cingulate, and dorsolateral prefrontal cortex (Ruocco, Amirthavasagam, Choi-Kain, & McMain, Reference Ruocco, Amirthavasagan, Choi-Kain and McMain2013). Extant findings, therefore, are mixed with respect to brain-related characteristics of ED in BPD. Further, the research generally has compared participants with BPD to healthy controls in terms of responses to unpleasant, emotionally evocative images; thus, more research is needed to determine whether these patterns of brain activity are specific to BPD or accounted for by elevated psychopathology, or relevant to emotional reactions to other situations or stimuli.
Some evidence has suggested that differences in brain activity during the implementation of certain emotion regulation strategies may hamper the effectiveness of these strategies among those with BPD. One study, for example, examined the effects of cognitive reappraisal among patients with BPD and past trauma, compared with healthy controls with and without past trauma (Lang et al., Reference Lang, Katchoubey, Frick, Spitzer, Grabe and Barnow2012). Compared with non-trauma-exposed healthy controls, those with BPD and past trauma showed less activation in the prefrontal cortex during upregulation of emotions and less activation of the anterior cingulate during downregulation (Lang et al., Reference Lang, Katchoubey, Frick, Spitzer, Grabe and Barnow2012). These patterns of brain activity may be due primarily to past trauma exposure, as there were no differences between trauma-exposed BPD participants and trauma-exposed healthy controls without PTSD (Lang et al., Reference Lang, Katchoubey, Frick, Spitzer, Grabe and Barnow2012). In another study, BPD patients and healthy controls viewed pictures depicting social interactions and either used cognitive distancing or simply viewed the pictures (Koenigsberg et al., Reference Koenigsberg, Fan, Ochsner, Liu, Guise, Pizzarello and Siever2010). Findings indicated that, while healthy controls and BPD patients similarly showed increased activation of the dorsolateral and ventrolateral prefrontal cortex and posterior cingulate when distancing, BPD patients showed less deactivation in the amygdala, among other areas (Koenigsberg et al., Reference Koenigsberg, Fan, Ochsner, Liu, Guise, Pizzarello and Siever2010). Similarly, a recent study compared the functional connectivity of various frontal regions with the amygdala during an effortful emotion regulation task (using emotional pictures), finding that BPD patients showed weaker increases in functional connectivity compared with nonpatient controls, but not compared with clinical controls with Cluster C personality disorders (Baczkowski et al., Reference Baczkowski, van Zutphen, Siep, Jacob, Domes, Maier and Arntz2017). Taken together, these findings suggest that brain functioning that normally facilitates emotion regulation may be impaired in BPD, but it remains unclear whether these impairments are unique to BPD.
In contrast, maladaptive emotion regulation strategies appear to temporarily result in greater connectivity of the amygdala and prefrontal regions (i.e., prefrontal inhibition of amygdala activity) among BPD patients. In a study modeling the effects of NSSI on brain functioning, Reitz et al. (Reference Reitz, Kluetsch, Niedtfeld, Knorz, Lis, Paret and Schmahl2015) randomly assigned self-injuring women with BPD (and healthy controls) to an incision or sham condition following a stress induction. The incision condition involved a shallow incision administered by an experimenter with a sharp plastic device, designed as a laboratory analogue for NSSI. Whereas healthy controls did not demonstrate evidence of physiological stress-reduction or increases in prefrontal connectivity in the incision condition, women with BPD reported decreased aversive tension (distress), decreased amygdala activation, and increased prefrontal connectivity (to the superior frontal gyrus) in the incision versus the sham condition. Findings from other research examining the effect of painful stimulation on brain activity among BPD patients also have suggested that pain increases the functional connectivity of prefrontal regions with the amygdala during emotional stressors (Niedtfeld et al., Reference Niedtfeld, Kirsch, Schulze, Herpertz, Bohus and Schmahl2012). Together with findings that adaptive strategies such as cognitive reappraisal under-recruit prefrontal regions associated with emotion regulation, these findings echo what self-injuring patients with BPD often say about how letting go of NSSI and learning more adaptive strategies is sometimes an uphill battle.
Behavioral aspects of ED or heightened BPD features
Other studies have examined the behavioral aspects of emotion regulation in BPD. BPD is associated with limited access to emotion regulation strategies (Glenn & Klonsky, Reference Glenn and Klonsky2009; Salsman & Linehan, Reference Salsman and Linehan2012), and this would suggest that persons with BPD might rely primarily on maladaptive strategies that produce quick relief. Some findings, however, have suggested that rather than simply relying on maladaptive emotion regulation strategies, people with BPD may tend to engage in greater effort to their emotions across both adaptive and maladaptive strategies (Dixon-Gordon, Aldao, & De Los Reyes, Reference Dixon-Gordon, Aldao and De Los Reyes2015), with the possible exception that they tend to use less acceptance as an emotion regulation strategy (Chapman, Dixon-Gordon, & Walters, Reference Chapman, Dixon-Gordon and Walters2013). There also is some evidence that people with heightened BPD features as well as those with a diagnosis of BPD might experience and report fewer emotional benefits from acceptance versus suppression and avoidance, both in the laboratory and in daily life (Chapman, Rosenthal, Dixon-Gordon, Turner, & Kuppens, Reference Chapman, Rosenthal, Dixon-Gordon, Turner and Kuppens2017). Compared with healthy controls, however, BPD patients find certain adaptive emotion regulation strategies, such as cognitive reappraisal and distraction, to be less effective (Sauer et al., Reference Sauer, Sheppes, Lackner, Arens, Tarrasch and Barnow2016). This could be due to lack of skillful use of the strategies, lesser impact of the strategies on emotional distress (e.g., as suggested by the research reviewed above on brain functioning in relation to cognitive reappraisal), problems with the validity of measures of emotion regulation strategies, or other factors. Overall, though, this pattern of findings tentatively suggests that persons with BPD may (a) put forth a fair amount of effort to regulate their emotions, and (b) find this effort largely unsuccessful.
Other behavioral aspects of emotion dyregulation include emotion-related disruptions in behavior and engagement in mood-dependent actions. Consistent with clinical observations and research suggesting that persons with BPD tend to engage in maladaptive behaviors when they experience emotional distress (e.g., NSSI), some laboratory research has suggested that people with heightened BPD features or a diagnosis of BPD might have difficulty inhibiting behavior or solving social problems when they are emotionally distressed (Dixon-Gordon, Chapman, Lovasz, & Walters, Reference Dixon-Gordon, Chapman, Lovasz and Walters2011). In addition, a recent study also found that participants with BPD demonstrated a decreased ability to learn from punishment following an emotional induction (Dixon-Gordon, Hackel, Tull, & Gratz, Reference Dixon-Gordon, Hackel, Tull and Gratz2018). More research is needed, however, to determine whether problematic behavioral patterns in BPD are primarily mood dependent. Furthermore, high rates of behaviors appearing to serve emotion regulation functions (e.g., NSSI, suicide attempts, substance use, or harmful forms of impulsivity), along with a strong association of BPD features with avoidant coping, emotion suppression, and experiential avoidance more broadly (Beblo et al., Reference Beblo, Fernando, Kamper, Griepenstroh, Aschenbrenner, Pastuszak and Driessen2013; Chapman, Specht, & Cellucci, Reference Chapman, Specht and Cellucci2005), could also be considered behavioral indicators of ED. Findings also have suggested that individuals with BPD have difficulty tolerating distress in the service of goal-directed behavior (Gratz, Frances, Sullivan, Hurt, & Clarkin, Reference Gratz, Rosenthal, Tull, Lejuez and Gunderson2006), and dissociation in response to stress is a defining feature of BPD and has been observed clinically and in laboratory and experience sampling research (Ebner-Primer et al., Reference Ebner-Priemer, Badeck, Beckmann, Wagner, Feige, Weiss and Bohus2005).
Conclusions
Although the developmental processes contributing to the development of ED and BPD are complex, advances in research and theory have begun to illuminate these processes. Influential theories of the development of BPD have strongly emphasized the role of ED. Linehan's (Reference Linehan1993a) original biosocial theory proposed the transaction of an emotionally vulnerable temperament and an invalidating environment contributes to the development of ED and BPD. More recent extensions of this model (Crowell et al., Reference Crowell, Beauchaine and Linehan2009, Reference Crowell, Kaufman, Beauchaine, Sharp and Tackett2014) have proposed that ED emerges from the interaction between an impulsive temperament and an invalidating environment. Findings have suggested that trait impulsivity is a strong candidate for a heritable vulnerability to BPD, and other findings have also suggested a role for neuroticism. It seems likely that both factors at play confer vulnerability to the adverse environmental contexts that often contribute to the development of ED and BPD. Broadly, an invalidating rearing environment may play an important role in socializing the vulnerable child's emotion regulation tendencies, shaping increasing emotional lability and dysregulation, culminating in the development of maladaptive coping behaviors that often begin to appear in adolescence (e.g., NSSI, suicidal behavior, and substance use). In addition, inadequate or delayed development of various social cognitive capacities and adaptive stress response may further contribute to ED and serious behavioral problems among adolescents. These behavioral problems, in turn, contribute to and predict the development of BPD. In adulthood, studies have highlighted various key cognitive, behavioral, and physiological components of ED, including possible differences in brain structure and functioning that are likely rooted in earlier developmental processes.
One challenge moving forward in the research is to determine where ED stops and BPD begins. Aspects of ED (emotional lability in particular) are built into the criteria for BPD. It is hard to imagine the construct of BPD without ED. Is ED a causal, transdiagnostic contributor to the array of features and behaviors observed among those with BPD, or is it best characterized as a core feature of BPD? In addition, based on the notion that ED consists of dysregulated emotional responding that interferes with important, goal-oriented behavior, many of the problematic behaviors occurring among those with BPD could be considered at least indirect indicators of ED. Considered in this way, ED could quickly devolve into a circular construct. The breadth and reach of the construct of ED is both its major strength and limitation, and this problem is not unique to BPD.
Invalidation has received considerable attention as an important socialization factor contributing to ED and BPD, and more research is needed to define and clarify the role of invalidation. Invalidation can take many forms that do not involve trauma or abuse, and recent reviews have suggested that studies thus far have failed to adequately measure and capture the construct of invalidation (Boucher et al., Reference Boucher, Pugliese, Allard-Chapais, Lecours, Ahoundova, Chouinard and Gaham2017; Musser, Zalewski, Stepp, & Lewis, Reference Musser, Zalewski, Stepp and Lewis2018). Further, it is very difficult to ascertain the specific influence of parenting behaviors on the development of psychopathology. There are many potential explanations for any association of parenting with negative outcomes, from shared genetic factors, to proactive, evocative, and other person–environment transactions. More research is needed to refine and measure the construct of invalidation, and more longitudinal research is needed to determine when (e.g., what developmental period), how, and for whom invalidation influences the development of ED and BPD.
The emerging picture of the complex developmental processes contributing to BPD and ED have important implications for prevention and treatment. Beauchaine, Hinshaw, et al. (Reference Beauchaine, Sauder, Derbidge and Uyeji2019) have outlined several key steps for the prevention and treatment of adolescent NSSI and suicidal behavior, arguing in particular that prevention should target the invalidating environment through evidence-based parenting interventions, and that treatment should address ED (which may be more influenced by environmental factors and easier to address) rather than trait impulsivity (which could be a fairly stable tendency). Beauchaine, HInshaw, et al. (Reference Beauchaine, Sauder, Derbidge and Uyeji2019) have argued that we know enough about key precursors and risk factors for NSSI and suicidal behavior that preventative interventions or treatments should target these precursors before the behaviors begin to emerge.
Furthermore, findings from recent research has supported this approach. A recent randomized clinical trial examined an adapted version of DBT (DBT-C; consisting of individual parent training and skills training sessions with parents and their children, and the availability of the therapist for phone coaching calls) for preadolescent children (N = 43; Perepletchikova et al., Reference Perepletchikova, Nathanson, Aexlrod, Merrill, Walker, Grossman and Walkup2017). The focus was on children with disruptive mood dysregulation disorder (DMRD), as DMRD is a key indicator of the risk for future childhood conduct problems and may indicate a trajectory toward self-injury and BPD for some children. Findings indicated a much lower dropout rate in DBT-C compared with treatment as usual, as well as a higher rate of positive treatment response and a higher remission rate for DMRD. It remains to be seen, however, whether this type of treatment might curtail the later development of NSSI, suicidal behavior, or BPD.
Theory and research suggest that, among adolescents, NSSI should be a high-priority treatment target. NSSI likely emerges earlier for most adolescents, compared with suicidal behavior, and should be a key target for treatment. Prominent models, for example, suggest that suicide attempts result from a combination of suicidal desire and the capability to enact lethal self-injury (i.e., acquired capability; Joiner, Reference Joiner2005; Klonsky & May, Reference Klonsky and May2015; Witte, Gordon, Smith, & Van Orden, Reference Witte, Gordon, Smith and Van Orden2012). Within these models, repeated NSSI increases an individual's ability to enact suicide plans (Hamza, Stewart, & Willoughby, Reference Hamza, Stewart and Willoughby2012; May & Victor, Reference May and Victor2018; Willoughby, Heffer, & Hamza, Reference Willoughby, Heffer and Hamza2015). The integrated model (Hamza et al., Reference Hamza, Stewart and Willoughby2012), for example, proposes three potential explanations for the NSSI–suicide attempt relation: (a) NSSI and suicide are related by virtue of their shared relation with a third variable (e.g., BPD); (b) NSSI is a “gateway” behavior that facilitates suicide attempts; and (c) based on Joiner's (Reference Joiner2005) interpersonal-psychological theory of suicide, NSSI is a painful/provocative event that increases an individual's capability to enact suicidal behavior (i.e., acquired capability) by increasing fearlessness of death and habituation to pain.
Along with preventative interventions targeting at-risk groups before the emergence of NSSI, treatment addressing NSSI when it emerges (and ideally beforehand) in adolescence would be expected to have a major impact on eventual risk for suicide and the development of problems characteristic of BPD. Findings from two large studies of DBT for adolescents have supported the efficacy of this treatment for reducing suicide attempts, self-injury, suicidal ideation, and depression among self-injuring adolescents (McCauley et al., Reference McCauley, Berk, Asarnow, Adrian, Cohen, Korslund and Linehan2018; Mehlum et al., Reference Mehlum, Tørmoen, Ramberg, Haga, Diep, Laberg and Sund2014). Further research is needed to determine if broader-scale preventative interventions, such as emotion regulation training in schools, might help to curb the development and exacerbation of ED and future BPD among adolescents.
Financial support
Work on this chapter was supported by a Michael Smith Foundation for Health Research Career Investigator Award and a Simon Fraser University FASS Dean's Research Grant.
