Disinhibition and Detachment in adolescence: A developmental cognitive neuroscience perspective on the Alternative Model for Personality Disorders

Abstract

Personality pathology often emerges during adolescence, but attempts to understand its neurocognitive basis have traditionally been undermined by problems associated with the categorical classification of personality disorders. In contrast, dimensional models of personality pathology, such as the Alternative Model for Personality Disorders in DSM-5, may provide a stronger foundation for neurobiological investigations of maladaptive individual differences in personality. As an example, we review studies of the adolescent development of reward processing and cognitive control and connect these systems to the normal personality hierarchy and to two dimensions included in the AMPD—Detachment and Disinhibition. We argue that by linking developmental changes in these systems to the AMPD, researchers will be better positioned to understand the relationship between neurocognitive development and the expression of personality pathology in adolescence and early adulthood.

Personality pathology often emerges and is most severe during adolescence [1]. In parallel, adolescence is characterized by widespread neurocognitive changes that influence reward processing, risk preferences, social behavior, and self-control [2]. Unfortunately, attempts to understand the links between personality dysfunction and brain development during adolescence have been stifled by the categorical classification of personality disorders (PDs) in the Diagnostic and Statistical Manual of Mental Disorders (APA, 2013). Problems with categorical PDs are well-known, including the use of arbitrary boundaries to distinguish between normal and abnormal functioning, excessive comorbidity, and within-disorder heterogeneity [4].

The Alternative Model for Personality Disorders (AMPD), codified in Section III of DSM-5 [3] provides a dimensional alternative to the existing categorical PD nosology. Within the AMPD, PD diagnoses are made on the basis of a) impairments in self and interpersonal functioning, and b) ratings on 25 maladaptive personality traits, which can be arranged hierarchically to resemble pathological variants of the Big Five [5].

The merits and limitations of the AMPD have been articulated elsewhere [6–8], so we do not review them here. Instead, we summarize specific findings from developmental cognitive neuroscience and examine their relevance for understanding normal personality dimensions and the AMPD. Furthermore, we argue that relative to PD categories, the AMPD provides a better foundation for interpreting and conducting neuroscientific studies of personality pathology, for several reasons. First, the AMPD removes arbitrary boundaries between normal and abnormal functioning, building on robust evidence that personality pathology reflects maladaptive variation in the same trait dimensions as normal personality [9]. Moreover, major traits in the normal personality hierarchy and the AMPD are conserved across species [10] and reflect neurobehavioral systems that have evolved to meet specific adaptive challenges [11]. Consequently, the AMPD may aid neuroscientific research by providing phenotypic targets that are more directly related to the organization of neural systems underlying normal and abnormal personality.

Another key feature of the AMPD is that it does not discourage the assessment of personality pathology symptoms in youth. By contrast, the DSM’s categorical model advises against pediatric PD diagnoses except in “relatively unusual instances” (p. 647, APA, 2013), even though numerous developmental studies have provided compelling evidence that personality pathology does not suddenly express itself at age 18 [12]. This language likely contributes to the underdiagnosis of PDs in youth and hinders research on developmental antecedents [13]. Personality pathology is the product of complex developmental pathways that begin early in life and influence subsequent adaptation and maladaptation [14,15]. By advancing a trait model that has reasonable continuity over the lifespan, the AMPD can facilitate neuroscientific research examining how biological processes (e.g., puberty) interact with environmental and sociocultural factors to shape the development of personality.

Finally, consistent with a developmental psychopathology perspective, the AMPD acknowledges that healthy and disordered personality functioning are largely governed by the same processes and mechanisms [16]. As a result, research on the biological and environmental processes that drive typical development can help to uncover the origins of personality pathology, just as studies of personality pathology can yield insights into healthy adaptation.

Below, we review contemporary studies of reward processing and cognitive control in adolescence and consider their points of convergence with the personality neuroscience and personality development literatures. We then consider how maladaptive or extreme variation in typical adolescent neurocognitive development might be conceptualized within the AMPD, focusing specifically on Detachment and Disinhibition.

Neurocognitive development in adolescence: Reward processing and cognitive control

During adolescence, youth navigate an array of new developmental tasks, from pursuing intense social and romantic relationships to exploring nascent autonomy [17]. Such experiences, coupled with neurocognitive development, contribute to an important period of plasticity in personality development [18]. For example, risk-taking increases in late adolescence, including drug and alcohol use, as well as risky sexual behavior [19]. Adolescents also become more sensitive to social evaluation [20]. For some, the changes that accompany adolescence can reveal or exacerbate existing vulnerabilities. Indeed, rates of personality pathology and other forms of psychopathology are elevated in adolescence [1,21], as are other serious outcomes including non-suicidal self-injury and suicide attempts [22].

Although several accounts of adolescent changes in personality and behavior have been proposed, dual systems models are especially popular [23]. There are important variations in dual systems models, but all share a common emphasis on the development and relative balance of two neurocognitive systems: a socioemotional reward processing system that peaks during adolescence and a cognitive control system that strengthens gradually with age.

Adolescent Reward Sensitivity: Linking Dual Systems Accounts to Normal Personality

Within the normal personality literature, individual differences in reward sensitivity typically fall within the domain of Extraversion, which describes individuals who are enthusiastic, gregarious, assertive, and sensation-seeking. Maladaptive forms of Extraversion are included within the AMPD under the Detachment domain, which encompasses pathologically low positive emotionality (i.e., anhedonia) and introversion (i.e. intimacy, social avoidance) [24].

A key tenet of dual systems models is that reward sensitivity increases with development, reaching a peak in adolescence before declining or potentially plateauing into adulthood [23]. Longitudinal research has not typically found age-related increases in Extraversion during adolescence, with several studies reporting a decline (e.g., Soto, 2016). However, it is important to recall that personality traits have a hierarchical structure, such that broader traits account for patterns of covariation among more specific traits. Specific facets of Extraversion show divergent developmental trajectories [26], such that those most closely related to dopaminergic functioning, including sensation-seeking and social dominance, actually increase during the adolescent years [27–29].

Findings from fMRI studies typically support the dual systems perspective. For example, ventral striatum activation during reward processing peaks in adolescence and is positively associated with risk-taking behavior [30,31]. Likewise, a broader valuation network encompassing the ventral striatum, ventromedial prefrontal cortex (vmPFC), and medial prefrontal cortex is more responsive to reward receipt among late adolescents than children or adults [32].

Reward-related striatal activation partly reflects dopaminergic inputs from the midbrain ventral tegmental area (VTA) [33,34]. In animal studies, mesolimbic dopamine activity increases following unpredicted rewards and decreases following unpredicted omissions of reward [35,36]. This pattern of activity has been called a reward prediction error (RPE) signal, which helps one to learn the subjective value of different stimuli and choose behavioral responses accordingly [37]. In humans, striatal RPEs elicited during a reward task peak during adolescence [38], though not all studies find this, potentially due to task differences or sample size [39]. Pre-adolescent and adolescent participants also show heightened functional coupling between the VTA and ventral striatum within incentivized contexts. This coupling declines with age as a function of decreasing cue-related activation of the orbitofrontal cortex [40], a region that encodes reward value and stimulates dopaminergic transmission via direct inputs to the VTA [41]. Taken together, this evidence points to a temporary increase in dopaminergic neurotransmission during adolescence, which may contribute to enhanced reward sensitivity.

Inputs from other cortical and subcortical areas to the striatum are also involved in modulating adolescent reward sensitivity. A recent meta-analysis of reward processing found that relative to adults, adolescents show greater activation in the ventral and dorsal striatum, as well as functionally connected regions, including the insula, amygdala, orbitofrontal cortex, and the anterior cingulate cortex (ACC) [31]. Developmental changes in the connectivity between these regions and the striatum appear to mirror changes in reward sensitivity. For example, connectivity between the ventral striatum and vmPFC declines with age, and mediates developmental reductions in pleasure ratings following reward [42]. Likewise, the structural integrity of limbic inputs to the striatum declines during the transition to adulthood, coinciding with age-related reductions in the enhancement of cognitive performance by rewards [43].

Neurobiological research on Extraversion mostly converges with the dual systems literature. Specifically, Extraversion moderates the efficacy of pharmacological manipulations of the dopamine system, is associated with heightened activity in the ventral striatum during reward anticipation, and has been linked to increased functional connectivity between the dopaminergic midbrain and striatum [26,44]. Furthermore, increases in the volume and activity of the ventral striatum are prospectively associated with increases in Extraversion during adolescence [28,45]. Finally, several studies have shown that Extraversion is associated with the reward-related positivity (RewP), an event-related potential that spikes following feedback about an outcome and likely encodes an RPE produced by midbrain dopaminergic inputs to the striatum and ACC [46,47]. Interestingly, the amplitude of the RewP increases in response to gains, but not losses, as children transition into adolescence, consistent with the prediction of dual systems models [48].

Maturation of Inhibitory Control and Concomitant Increases in Conscientiousness

Cognitive control involves the capacity to guide one’s behavior flexibly and voluntarily, which requires both working memory and inhibitory control (or the ability to suppress disruptive impulses) [49]. Within the personality literature, cognitive control is closely related to Conscientiousness, a trait that describes those who are persistent, planful, and orderly¹. Much like cognitive control, Conscientiousness has been hypothesized to facilitate the pursuit of long-term goals and promote rule-based behavior, both of which rely on the ability to avoid distractions and suppress disruptive impulses [50]. Maladaptively low Conscientiousness is represented within the Disinhibition domain of the AMPD [51]. Disinhibition has been linked to a variability in both behavioral and neurophysiological measures of error processing and inhibitory control [52]. For example, Disinhibition is associated with a reduction in the amplitude of the P3 event-related potential (ERP), a positive-going waveform that occurs following presentation of a salient stimulus in cognitive control tasks. Interestingly, P3 indices can be combined with behavioral performance measures and self-report ratings to model a cross-modal, latent Disinhibition factor [53].

From a dual systems perspective, adolescent risk-taking is conceptualized as a developmental imbalance in which the influence of the reward system outweighs a slower-developing cognitive control system. Behavioral studies have found that while many aspects of cognitive control are present early in life, they improve gradually over the course of development, remaining more variable in adolescence than in adulthood [54]. This mirrors the developmental trajectory of Conscientiousness, which tends to increase substantially in adolescence before rising more slowly through adulthood (though, it may also dip early in adolescence) [55]. Development of cognitive control depends on functional changes in the activity and connectivity of distributed regions, including dorsolateral prefrontal cortex (DLPFC), posterior parietal cortex, insula, and ACC [56]. For example, children engage the DLPFC more during inhibitory control tasks, with activation steadily declining into adulthood and mirroring performance improvements [57,58]. Interestingly, deactivating the DLPFC using transcranial magnetic stimulation (TMS) leads to increased risk-taking on a gambling task [59], whereas activating DLPFC using electrical stimulation has the opposite effect [60]. Focal damage to the area has also been linked to lower self-reported Conscientiousness [61].

The dorsal ACC is another key region in the development of cognitive control. It tends to be more heavily recruited with age, particularly in response to errors during inhibitory control tasks [57,58]. Age-related increases in dorsal ACC activation may reflect developmental improvements in one’s ability to effortfully attend to current task demands and adjust behavior accordingly as those demands change (perhaps via the recruitment of other control regions such as DLPFC) [62]. Conscientiousness is positively associated with dorsal ACC activation during inhibitory control tasks [63]. Likewise, Disinhibition has been linked to a reduced amplitude of the error-related negativity (ERN) on go/no-go tasks; the ERN is an event-related potential involved in error processing that has been localized to the ACC [64]. Changes in Conscientiousness also mediates the relationship between cortical thinning of the dorsal ACC during adolescence and measures of depression, anxiety, aggression, and risk-taking, consistent with the idea that poor inhibitory control is a transdiagnostic risk factor [65].

Changes in cognitive control over the course of development are tightly linked to changes in a set of highly overlapping functional brain networks—including the salience network (SN), cingulo-opercular network, and Rothbart and Posner’s executive attention network [66]—all of which include the ACC and DLPFC. The basic architecture of these networks is in place early in development, but between-network connectivity strengthens throughout development, supporting improvements in cognitive control [67]. Consistent with this, network analyses show that prefrontal cognitive control regions break off into a distinct module from subcortical regions (including those involved in reward sensitivity) in individuals low in Conscientiousness [68]. The SN, whose central nodes include the dorsal ACC and anterior insula, has been linked to the detection of behaviorally relevant stimuli and the initiation of cognitive control [69]. Previous studies have shown that developmental increases in SN recruitment during inhibitory control tasks mediate age-related performance improvements [32]. Conscientiousness has been hypothesized to be related to functionally overlapping components of the salience and ventral attention networks [44], which Rueter and colleagues (2018) collectively refer to as the “goal priority network.” In a recent resting-state fMRI study, these authors found that Conscientiousness was associated with greater connectivity within part of the goal priority network that included the DLPFC, anterior insula, and dorsal ACC. Conscientiousness was also associated with greater coupling of these regions with functionally distinct subnetworks [70].

Disinhibition is also negatively associated with functional connectivity between the insula and DLPFC during inhibitory control [71], which suggests that the coordination of control-related regions may be critical to suppressing impulsive responses. Interestingly, cortical thickness of the insula is negatively correlated with both age and Conscientiousness, consistent with the idea that greater synaptic plasticity in the insula facilitates improved deliberation with age [72]. Conscientiousness may also be supported by better structural connectivity of the insula with limbic and prefrontal regions. For example, Lewis and colleagues (2016) found that Conscientiousness was positively associated with higher fractional anisotropy (a measure of structural connectivity) in the left uncinate fasciculus, a white matter tract that connects the insula, orbitofrontal cortex, anterior temporal cortex, and amygdala [73]. Taken together, empirical studies consistently find that the cognitive control system develops gradually with age, is related to widespread changes in control-related brain circuits, and may be a key source of individual differences in Conscientiousness and Disinhibition. Thus, heterogeneity in the development of cognitive control and reward processing neurobehavioral systems likely contributes to normative age-related risk-taking. At the extremes, variation in the development of these systems may underlie more problematic impulsive behaviors that are typically observed in personality pathology.

Linking Adolescent Neurocognitive Development to Dimensional Personality Pathology

Thus far, we have reviewed evidence supporting a dual systems account of adolescent neurocognitive development and have related individual differences in these circuits to normative personality traits, namely Extraversion and Conscientiousness. To date, however, very few studies have explicitly linked the emergence of personality pathology with prevailing accounts of neurocognitive development. By removing the boundary between normal and abnormal functioning, the AMPD provides a useful foundation from which researchers can develop and test such accounts.

How might the AMPD capture maladaptive development of the reward processing and cognitive control systems? At least with respect to reward processing, this is a complicated question. The normal personality literature distinguishes between two interrelated facets of Extraversion: an affiliative component reflecting hedonic reward consumption (aka “liking”), and an agentic component that is more closely related to incentive motivation (aka “wanting”) [34,44]. Recent evidence has shown that the Detachment domain of the AMPD is a slightly stronger marker of affiliative Extraversion (reversed) than agentic Extraversion [74], pointing to a potential gap in coverage that may require further consideration as the system is revised over time. On the other hand, there may also be substantive reasons for this gap. While some components of agentic Extraversion are robustly linked to risk for psychopathology, others show minimal associations [75]. Sensation-seeking is a lower-order personality trait that equally blends Conscientiousness and agentic Extraversion, and it is robustly associated with externalizing problems [75]. This suggests incentive motivation may be more problematic when accompanied by at least some difficulties with top-down inhibitory control [76].

Interestingly, sensation-seeking exhibits a U-shaped developmental trajectory, peaking in adolescence in much the same way as ventral striatal activity on reward processing tasks [29]. Deviations from the typical U-shaped reward processing trajectory may be associated with the onset of externalizing problems in adolescence. For instance, one study found that smaller ventral striatum volumes in mid-adolescence predicted an increased likelihood of initiating substance use in a two-year follow-up period [77]. In the same study, increases in reward sensitivity were associated with a greater likelihood of substance use initiation, and with more self-reported alcohol use during the follow-up period [77].

Disruptions in reward circuitry during adolescence may also affect a host of other problematic behaviors that are more indirectly related to reward-seeking. For instance, activity in the ventral striatum following provocation predicts the magnitude of one’s subsequent retaliation, suggesting that social status behaviors partly depend on reward-related circuitry [78]. Consistent with this, larger striatal volume in early adolescence is positively associated with aggression [79], potentially reflecting a disruption in the typical development of circuits that mediate reward-related aggression. Overall, these findings highlight the need for longitudinal cognitive neuroscience studies to address questions about the developmental trajectory of personality pathology in youth. For instance, do some adolescents exhibit time-limited personality pathology that remits once cognitive control systems exert a greater influence on behavior, while reward processing systems either plateau or decline? Does more persistent sensitivity of reward circuits predict maladaptive expressions of Extraversion in other youth? These are questions that can only be answered via longitudinal studies, preferably in high risk developmental samples that capture the full range of maladaptive trait dimensions.

Contextualizing maladaptive cognitive control within the AMPD is more straightforward, as Disinhibition is closely related to both neural and behavioral indices of inhibitory control [53]. One AMPD facet particularly relevant to dual systems accounts is risk-taking, which was originally assigned to the Disinhibition domain but has a nearly equal negative cross-loading on Detachment (suggesting it may be the AMPD equivalent to sensation-seeking) [80]. Youth high on risk-taking may struggle to balance the lure of immediately rewarding opportunities (e.g., going to a party, having sex, trying drugs) in adolescence with the sustained effort required to attain longer-term goals, including academic and occupational achievement, stable familial and romantic relationships, and maintaining a healthy lifestyle. Given its pattern of cross-loadings, personality neuroscience researchers should consider how risk-taking depends on circuits related to both reward and cognitive control. For instance, connectivity between the dorsal striatum and inferior frontal gyrus (which is part of the goal priority network linked to Conscientiousness) is associated with an improved capacity to suppress approach responses to appetitive cues in teens and adults compared to children; only in adolescents, however, this frontostriatal coupling is also accompanied by greater ventral striatal activity [81]. This might suggest that risk-taking in adolescence reflects a biasing of cognitive control circuits related to Conscientiousness by reward signals that are encoded in the ventral striatum and more closely linked to Extraversion.

Of course, other circuits and traits that we did not cover here undoubtedly play a role in how adolescents navigate developmental challenges. For instance, another AMPD trait, Antagonism, has been linked to components of the default mode network implicated in mentalizing abilities [82], and may contribute to a devaluing of stable relationships during adolescence, in favor of short-term rewards (e.g., being more willing to manipulate or exploit a partner for immediate gain). In contrast, high scores on the AMPD’s Negative Affect, coupled with risk-taking, could promote coping strategies meant to gain immediate relief from distress (e.g., self-harm; suicide attempts). The role of these traits in the development of personality pathology likely depends on the plasticity of functionally distinct networks during adolescence. For instance, networks supporting mentalizing abilities and social cognition continue to develop during adolescence [20], and extreme variations in those systems may play an important role in more interpersonal forms of personality pathology (e.g., rejection sensitivity, suspiciousness, or manipulativeness).

Conclusions

The dissemination of the AMPD, and the broader shift toward dimensional conceptualizations of psychopathology in recent years, are promising developments for researchers interested in the neurocognitive basis of personality pathology. By removing an arbitrary normal/abnormal boundary and reframing personality pathology in terms of empirically derived traits, the AMPD helps to build a bridge between developmental studies of neurobehavioral systems and dimensional studies of traits. Recent evidence suggests that the AMPD has clinical utility, improving communication between patients, caregivers, and providers, easing diagnostic assessment for clinicians, and helping to guide clinical decision-making [83,84]. Integrating developmental neuroscientific evidence into the AMPD can facilitate new insights into the mechanisms that fuel personality pathology and potentially leads to new innovations in assessment and intervention. Here, we reviewed how two traits in the AMPD, Detachment and Disinhibition, can be used to connect the emergence of adolescent personality pathology with the development of reward and cognitive control systems. Moving forward, we hope that this kind of integration can be used to highlight areas in the AMPD that are in need for further elaboration and refinement. Incorporating evidence from multiple levels of analysis should improve the rigor and comprehensiveness of the AMPD, making it a more clinically useful nosology.