The Emotionally Sensitive Child-Adverse Parenting Experiences-Allostatic (Over)Load (ESCAPE-AL) Model for the Development of Secondary Psychopathic Traits

This model asserts that the primary mechanisms explaining why individuals who experience prolonged extreme adverse parenting develop secondary CU/psychopathic traits involve: a predisposition to heightened emotion sensitivity in the child that causes them to respond to these experiences with intense emotional and physiological arousal. These responses produce stress levels that are toxic because they are inadequately buffered due to problematic (i.e., disorganized) parent–child attachment and the parent’s failure to provide emotionally sensitive, responsive, and predictable care, and instead being the source of this stress (see Fig. 1). Over repeated increasing aversive parent–child interactions across time, this transactional process damages the child’s developing stress response system, which produces the callous and chronically antisocial, aggressive, and impulsive phenotype that mimics primary CU/psychopathy. The sections below outline the theoretical and empirical bases for this model. They provide a more detailed account of the proposed developmental processes involved in secondary CU/psychopathy and identify future research directions for testing this Emotionally Sensitive Child-Adverse Parenting Experiences-Allostatic (Over)Load (ESCAPE-AL) model.

The ESCAPE-AL model argues that the extreme levels of anxiety, depression, and anger/hostility (i.e., “neuroses”) seen in individuals with secondary CU/psychopathic traits are underpinned by a dispositional vulnerability involving heightened “emotional sensitivity” to environmental stress. The emotionally sensitive individual has a low threshold for reacting to events in their environment, even those which may be inconsequential to others, with a rapid and extreme emotional reaction from which they are slow to recover. Emotional sensitivity is thought to be evident from birth, manifesting in infancy as hypersensitivity and intense negative affectivity (i.e., excessive crying and difficulty soothing) to environmental stimulation (e.g., novel stimuli such as the presence of unfamiliar adults, others’ emotional expressions, sounds) (Crowell et al., 2009; Linehan, 1993). Models of infant temperament capture this behavioral pattern within dimensions of negative affect/reactivity (i.e., fear, distress to limitations, frustration, anger, soothability; Putnam et al., 2008) and emotionality (i.e., fear, anger, general distress; Buss & Plomin, 1975), which is one of the first in life to emerge and which demonstrates homotypic continuity in predicting distress and negative affectivity at preschool age (Komsi et al., 2006; Putnam et al., 2008; Rothbart, 1989). Temperamental negative affectivity in infancy—particularly involving frustration (i.e., interruption of ongoing tasks/goal blocking) and sadness (i.e., exposure to suffering, disappointment, object loss) subcomponents—predicts co-occurring internalizing and externalizing problems in preschoolers with low effortful control (i.e., emotional and behavioral self-regulation; Gartstein et al., 2012). Similarly, Linehan (1993) explained that the “[emotionally] sensitive child reacts emotionally to even slight frustration or disapproval… annoyance may turn to rage” and “…partings may precipitate very intense and painful grief” (p. 44).

Heterotypic continuity of this heightened emotional sensitivity expresses as difficult temperament in the second year of life and from early childhood as angry-irritable symptoms. Anger-irritability is a subdimension of ODD involving excessive reactivity to negative emotional stimuli, which among other ODD dimensions, uniquely predicts later life internalizing problems (Beachaine & Tackett, 2020; Stringaris, 2011). Extended to adulthood, heightened emotional sensitivity is captured within neuroticism (i.e., the tendency to experience negative affect) from adult personality models (e.g., Costa & McCrae, 1992; Evans & Rothbart, 2007; Eysenck & Eysenck, 1985; Guarino et al., 2007); within the Internalizing (fear, distress subfactors) spectrum of the Hierarchical Taxonomy of Psychopathology (HiTOP, Kotov et al., 2017), and high Acute Threat (“fear”) and Potential Threat (“anxiety”) biobehavioral constructs of the Research Domain Criteria (RDoC, Insel et al., 2010) alternative dimensional frameworks to major classification systems for mental disorders. However, emotional sensitivity is a related, broader construct that is not restricted to threatening stimuli. Rather, adults who suffer from it describe emotional sensitivity as involving frequent, erratic, intense, persistent, and unwanted negative affectivity in reaction to external stimuli, even that which is non-consequential, non-emotional, or unknown to the individual (Wall et al., 2018). Such general situational physiological reactivity (e.g., general corrugator muscle tension) is, however, robustly associated with Acute Threat (i.e., assessed using self-report and multiple physiological indicators of negative emotional reactivity) and anxiety disorder symptoms (Yancey et al., 2016).

There is empirical evidence supporting heightened emotional sensitivity in secondary CU/psychopathy across the lifespan. First, high temperamental negative reactivity to frustration and to novelty at 6-months-old predicted greater mother-reported CU traits at first grade within a highly disadvantaged birth cohort (Wagner et al., 2017). Second, children characterized as secondary CU variants at age 3 were significantly more likely to have a difficult temperament assessed in the first 6 months of life, relative to control children, but not those with primary CU or CP-only (Fanti & Kimonis, 2017). Third, childhood CU traits and internalizing problems are both predicted by earlier levels of angry-irritable ODD symptoms (Barker & Salekin, 2012; Stringaris & Goodman, 2009). Finally, secondary CU variants show heightened attention and arousal to both socioemotional and non-emotional stimuli and heightened sensitivity to threat (i.e., high anxiety or fearful arousal; e.g., Fanti et al., 2018), which contrasts against the reduced pattern that is hypothesized to explain the development of (primary) CU traits in Dadds and Frick’s (2019) Responsiveness, Emotional Attention, and Learning (REAL) model and Waller and Wagner’s (2019) Sensitivity to Threat and Affiliative Reward (STAR) model.

Several studies report that individuals classified as secondary CU/psychopathy variants do not show the same neurobiological and neurocognitive correlates related to the processing of fear and distress stimuli that are found in primary variants (Dadds et al., 2018). For example, incarcerated adolescent boys classified as secondary CU variants showed significantly greater attentional orienting to distress cues on a dot-probe task relative to primary CU variants that showed reduced attentional orienting to distress (Kimonis et al., 2012a), a pattern that is robustly observed in CU-type conduct problems (e.g., Kimonis et al., 2018). Another study found that incarcerated aggressive male adolescents with secondary CU traits showed increased augmented startle reflexes following visual threat prime—a physiological indicator of Acute Threat—compared with those with primary CU traits who showed the characteristic reduced augmentation of the startle reflex found in association with psychopathic traits (Kimonis et al., 2017a, 2017b; Patrick et al., 1993). In one of the few fMRI studies, Meffert et al. (2018) found that trauma exposure moderated the association between CU traits and right amygdala responsiveness to fearful facial expressions (see also Sethi et al., 2018). That is, youth with secondary CU traits (high maltreatment) showed greater right amygdala reactivity to fear faces relative to those with primary CU traits (low maltreatment) who showed the typical pattern of reduced amygdala responses to negative stimuli found in psychopathic youth and adults (Blair, 2013; Marsh et al., 2008; Viding et al., 2012). Another study showed that juvenile variants differed in the functional connectivity of the amygdala (Dugré & Potvin, 2023). These results are consistent with the ESCAPE-AL model in finding that heightened emotional sensitivity is an important risk factor for secondary CU traits.

While patterns of emotional attention and arousal indicative of heightened emotional sensitivity clearly differentiate secondary from primary CU variants, there has been relatively less research comparing CU variants against youth with CP-only for whom heightened emotional arousal in the context of real or perceived threat is central to developmental models (Dodge, 2009; Lochman et al., 2000). The few notable studies conducted to date compared variant groups to incarcerated adolescents scoring low on psychopathic traits, finding some support for greater emotional sensitivity in secondary CU variants relative to CP-only youth. First, Kimonis et al., (2017a) found that secondary CU variants showed enhanced startle potentiation to aversive images relative to youth with CP-only, both those with a history of maltreatment (Cohen’s d = 0.65, p = 0.06) and without (d = 0.58, p < 0.05). Second, in a different incarcerated sample Kimonis et al., (2012a) found that juvenile secondary psychopathic variants showed greater facilitation to distress stimuli on a dot-probe task relative to boys low on psychopathic traits, although this moderate effect (d = 0.50) did not achieve statistical significance. Thus, relative to youth with CP-only, secondary CU variants are both more emotionally sensitive and show a callous phenotype.

This callous phenotype also distinguishes secondary CU traits from BPD, a construct for which emotional sensitivity is central to its major developmental models (i.e., Linehan’s biosocial theory: Crowell et al., 2009). There is a long-standing debate over whether BPD is a female phenotypic expression of psychopathy (Blackburn, 1998; Sprague et al., 2012). Several studies find that mixed sex adults and youth classified as secondary CU/psychopathy variants show higher levels of BPD symptoms than primary variants (Falkenbach et al., 2014; Goulter et al., 2017; Skeem et al., 2007). These two populations also share several overlapping characteristics, including high rates of early adverse childhood experiences, including maltreatment, trauma, parent–child attachment problems, and poor quality of care (Lyons-Ruth, 2008; Porter et al., 2020); a complex comorbid clinical profile involving anxiety and mood disorders, pathological anger, PTSD symptoms, substance-related disorders, and disorders of impulse including ADHD (Zanarini et al., 1998); and an attentional bias toward negative and threatening stimuli (Jovev et al., 2012; Kaiser et al., 2016). The literature provides less clarity around characteristics distinguishing between these two clinical populations, highlighting the need for studies comparing youth with secondary CU traits and borderline pathology as an important future research direction.

While their shared characteristics leave open the possibility of etiological overlap between BPD and secondary CU traits, these psychopathologies diverge in several key ways. First, unlike secondary CU traits that can be reliably identified in early childhood, BPD is rarely diagnosed before adulthood despite increasing evidence that it emerges earlier than once thought with adult symptoms evident from late childhood and reliably identified during adolescence (Miller et al., 2008). Second, developmental models of BPD and the ESCAPE-AL model diverge in the specific environmental influences giving rise to psychopathology in a biologically vulnerable child. Major etiological models specify that BPD develops for the emotionally sensitive child within an invalidating early family environment in which caregiver(s) are emotionally “unavailable” and intolerant of the child’s expression of negative emotions, particularly those unlinked to observable events (Linehan, 1993). While such emotionally dysfunctional interactions are more likely to occur within abusive or neglectful contexts, current consensus is that child maltreatment is neither necessary nor sufficient to develop BPD (see Crowell et al., 2009). In contrast, the ESCAPE-AL model argues that repeated transactions between the emotionally sensitive child and caregiver(s) that are a source of threat are central to the etiology of secondary CU/psychopathic traits. While it stands to reason that childhood secondary CU traits may be a risk factor for later BPD, or that emotionally sensitive young children who do not experience sufficiently adverse parenting to precipitate allostatic overload later develop BPD and not secondary CU traits, these are empirical questions that future prospective, longitudinal research is needed to test. The ESCAPE-AL model argues that a third key difference between these clinical groups is that stress-induced physiological damage produces the callous phenotype that distinguishes secondary variants from other forms of psychopathology that share similar symptoms and correlates (i.e., BPD, CP-only).

An interesting paradox is that individuals with secondary CU traits show this classic “cool” and callous phenotype while also displaying heightened emotional sensitivity that is evident in their attentional, physiological, and neurocognitive responses. Theory explains that these individuals learn to emotionally detach, despite having the capacity for a full range of emotional experience, in order to escape the pain, distress, and fear associated with trauma exposure (Kerig & Becker, 2010; Porter, 1996). Their callous façade has been described as a type of posttraumatic response that serves a self-preservation function as a form of “survival coping” (p. 17, Ford et al., 2006). This pattern of avoiding, controlling, and repressing unwanted emotions, as in dissociation, involves overmodulation of emotions and is characteristic of one PTSD subtype linked to trauma occurring within close interpersonal relationships, relative to an undermodulation subtype that conversely shows hyperarousal and emotional reactivity (Lanius et al., 2010). Mozley et al. (2018) proposed that posttraumatic overmodulation occurring in association with exposure to interpersonal trauma is a risk factor for the development of CU traits, supported by findings that incarcerated youth with secondary CU traits showed both high rates of interpersonal trauma (v. non-interpersonal trauma, e.g., accidents, illness, disaster; Kerig et al., 2012) and greater self-reported emotional numbing (specific to fear and sadness), nonacceptance, and lack of clarity around their own emotional states than primary CU variants (Bennett & Kerig, 2014). Other research reports elevated symptoms of alexithymia in secondary CU variants, indicative of an impaired ability to identify and describe their own emotional experiences despite experiencing emotional arousal in distressing situations (Cecil et al., 2018). In contrast to the perspective that overmodulation symptoms are learned in response to interpersonal trauma, the ESCAPE-AL model argues that the mechanism underpinning them involves extreme, unmitigated, and neurotoxic levels of stress experienced in response to chronic severely adverse early caregiving, which exceed the young dispositionally at-risk child’s coping resources and damages their developing physiological stress response system.

There is robust evidence that secondary CU variants experience extreme levels of adverse childhood experiences and particularly dysfunctional parenting (Craig et al., 2021a). The nature of their adversity is broad, ranging from low maternal warmth/sensitivity (Craig et al., 2021a; Fanti & Kimonis, 2017; Kaouar et al., 2023) and affectionless and rejecting parenting (Karpman, 1941), to parenting that is harsh (Goulter et al., 2017; cf. Bégin et al., 2021; Craig et al., 2021a; Humayun et al., 2014), hostile and verbally aggressive (Craig et al., 2021a), to childhood maltreatment (i.e., sexual or physical abuse or neglect; Bégin et al., 2021; Goulter et al., 2019; Kahn et al., 2013; Kimonis et al., 2011, 2012a, 2017a, 2017b; Porter, 1996), to experiencing interpersonal traumatic events including witnessing intimate partner violence and other violence exposure (Bennett & Kerig, 2014; Docherty et al., 2016; Kahn et al., 2013; Sharf et al., 2014; Tatar et al., 2012), to broader cumulative environmental risk across early childhood (Cecil et al., 2014). For example, secondary CU variants identified at age 3 experienced significantly lower levels of maternal sensitivity between 6- and 24-months old, relative to control children and those with CP-only (Fanti & Kimonis, 2017). Relatedly, meta-analytic research (k = 29 studies, N = 9894) reported a significant moderate (r = 0.23) association between childhood maltreatment and CU traits, which was stronger at higher levels of anxiety (Todorov et al., 2023; see also de Ruiter et al., 2022).

The limited longitudinal research conducted in this area finds that adverse parenting increases over time for children with secondary CU traits. For example, longitudinal research uncovered that parents of children with secondary CU traits show increasingly harsh, punitive, and physically abusive punishment, and/or avoidant, cold, rejecting, and emotionally neglectful behaviors over time, relative to primary CU variants (Bégin et al., 2021; Craig et al., 2021b). This escalation in adverse parenting may be driven in part by parents’ maladaptive cognitions. Primary caregiver(s) describe the frequent, persistent, intense, and erratic negative affectivity of children with an emotionally sensitive disposition as producing a feeling of “walking on eggshells” (Linehan, 1993). Similarly, at the University of New South Wales Parent–Child Clinic where we assess and treat many young children with conduct problems and secondary CU traits, parents often describe their child in this way. Parents of these young children with secondary CU traits have more negative cognitions about their relationship with the child that are characterized by low warmth and greater feelings of anger/hostility, resentment, contempt, active avoidance and/or a desire to do harm to their child, relative to parents of primary CU variants and CP-only children (Kaouar et al., 2023; see also Reijman et al., 2016). This finding demonstrates the need for future research investigating more nuanced parenting factors associated with CU variants to advance understanding of their development and inform tailoring of parenting interventions.

Studies examining micro-components of parenting behaviors in early childhood found that maternal fright during pregnancy, disinterest during infant feeding (Mendoza Diaz, 2018), low parent–child emotional closeness (i.e., low eye contact, positive and approving facial/ vocal affect, physical proximity during interactions) and emotion communication were uniquely associated with CU traits (Koh et al., 2023), although these studies did not disaggregate variants. Attachment research suggests that frightened and frightening parenting behaviors may originate from parental trauma exposure, leading to an inability to cope with intense negative emotions displayed by their child and contributing to disorganized attachment (Main & Hesse, 1990a, 1990b). Indeed, parents of children with secondary CU traits show higher rates of intimate partner violence and stress-related psychopathology, which are associated with an increased risk of engaging in hostile, rejecting, detached, unresponsive parenting and maltreatment (Cecil et al., 2014; Cohn et al., 1986; Fanti & Kimonis, 2017; van Ijzendoorn et al., 2020). Another potentially important parental risk factor for engaging in these adverse behaviors and warranting future research investigation is parental psychopathic traits, which are associated with child CU traits generally (Mendoza Diaz et al., 2018).

Repeated severe prolonged childhood maltreatment is a precipitating traumatic event within ICD-11’s Complex PTSD diagnosis (CPTSD). Also, exposure to traumatic interpersonal childhood victimization (e.g., emotional/physical abuse, family and community violence) occurring together with disrupted primary caregiver attachment (e.g., caregiver separation, impairment) is central to Developmental Trauma Disorder (DTD), which was proposed as a new and separate diagnosis to PTSD in DSM-5 but ultimately rejected due to limited empirical evidence. CPTSD and DTD share several symptoms in common with secondary CU traits, including impaired interpersonal empathy (DTD), reactive aggression (DTD), emotional and behavioral dysregulation associated with impaired effortful control (DTD, CPTSD), emotional overmodulation (DTD, CPTSD, e.g., dissociation, emotional numbing), threat-related attentional bias (DTD), and attachment problems (DTD), but are also distinct in their requirement that the individual also meet diagnostic requirements for PTSD (CPTSD) that was found to present in the majority (69%) of DTD cases (Spinazzola et al., 2021). Also, relative to children diagnosed with PTSD alone, those meeting criteria for DTD alone had the highest rates of ODD and those with DTD + PTSD had extensive psychiatric comorbidity (Ford et al., 2022). Critically, impaired prosocial emotions co-occurring with high levels of both externalizing and internalizing disorder symptoms are core to secondary CU traits, none of which are required criteria for diagnoses of DTD or CPTSD. Future research is needed to elucidate whether these are separate diagnostic constructs involving distinct etiological pathways or are manifestations of a more complex underlying form of psychopathology that is rooted in the cumulative physiological effects of exposure to traumatic stress whose source is the young child’s primary caregiver(s).

Chronic adverse parenting experiences repeatedly activate the child’s physiological stress response system, producing cumulative wear-and-tear on the body that is termed “allostatic load” (Danese & McEwen, 2012; McEwen & Stellar, 1993). Relative to CP-only children who also experience elevated dysfunctional parenting practices (i.e., harsh and coercive parenting, Dodge et al., 1990; Patterson et al., 1984), the environmental challenges (i.e., increasing frequent and severe angry/hostile, harsh/punitive, cold/rejecting, insensitive parenting) of children with secondary CU traits (a) are more extreme and prolonged; (b) are experienced with more intense emotional and physiological arousal because of the child’s emotionally sensitive disposition; (c) are inadequately buffered; and (d) exceed the developing young child’s physiological ability to cope (Guidi et al., 2021). A child’s ability to adaptively respond to extremely challenging situations and other changing conditions by maintaining physiological equilibrium is termed “allostasis” (Sterling & Eyer, 1988). Allostasis is facilitated in securely attached parent–child dyads by the warm, sensitive, and responsive caregiver who has a buffering effect on the child’s stress levels (Gunnar, 2017; Hostinar et al., 2014). In contrast, for children with secondary CU traits who show the highest rates of disorganized and avoidant attachment styles, even compared against primary CU variants who were predominately securely attached (Cecil et al., 2018), the primary caregiver(s) both exacerbates their stress levels as a central source of fear and threat and at the same time fails to buffer the harmful effects of the chronic, cumulative, and toxic levels of stress they produce in their emotionally sensitized young child (Carroll et al., 2013; Main & Hesse, 1990a, 1990b; van Ijzendoorn et al., 1999).

When a child’s stress response system is continuously activated and buffering factors are inadequate, as in cases of disorganized parent–child attachment and low maternal responsiveness (Ellis & Del Giudice, 2014; Evans et al., 2007), this can produce an extreme state in which allostasis reaches a tipping point known as “allostatic overload” (Fava et al., 2019; McEwen & Stellar, 1993). Allostatic overload resulting from prolonged early adverse childhood experiences dramatically affects brain development, leading to structural and functional abnormalities in several brain areas (i.e., hippocampus, prefrontal and orbitofrontal cortex, amygdala) and altering the maturation and responsiveness of psychophysiological stress response systems (Danese & McEwen, 2012; Weiss & Wagner, 1998). Chronic stress impairs the ability of the hypothalamic–pituitary–adrenal (HPA) axis, the central component of the stress system, to maintain homeostasis and results in dysregulation that is evident in atypical arousal and stress hormone (e.g., cortisol) levels (Danese & McEwen, 2012). For example, relative to non-maltreated children, maltreated children showed greater variability in their afternoon basal cortisol concentrations, which attenuated over time for those with higher initial levels (Doom et al., 2014; Trickett et al., 2010). This altered stress response system may phenotypically express as the callous, antisocial, disinhibited, and comorbid clinical profile that characterizes individuals with secondary CU/psychopathic traits. Allostatic over/load is the hypothesized mediator of chronic stress on such psychological and physical disorders (Danese & McEwen, 2012), and is measurable in the hormonal outputs of the child’s physiological stress response system (Chen et al., 2012).

There are several biomarkers of this HPA axis dysfunction, one of which is an unbalanced ratio between the stress hormone cortisol and the most abundant steroid in the human body, dehydroepiandrosterone (DHEA), which are co-released under stressful conditions. According to the cortisol-DHEA ratio hypothesis, DHEA has a protective function by helping to return the stress response system to homeostasis. In doing so, it aids in buffering the neurotoxic effects of prolonged cortisol exposure on the HPA axis and hippocampus (Kimonides et al., 1999; Young et al., 2002). A high cortisol-to-DHEA ratio in the context of adverse childhood experiences is an indicator of increased chronic stress, HPA axis dysregulation, and is associated with poor mental and physical health (Goodyer et al., 2001). One study found that cortisol-to-DHEA ratios were higher among incarcerated adolescent offenders (N = 232, M age = 16.75, SD = 1.15) classified as secondary CU variants, relative to all other antisocial subgroups (Kimonis et al., 2017a, 2017b). In this study, secondary CU variants also self-reported the highest levels of childhood maltreatment, stressful life events, and PTSD symptoms.

Findings for cortisol alone are mixed, likely due to methodological differences between studies. One study found lower morning basal cortisol concentrations in 3-year-old secondary CU variants relative to primary CU variants and controls, but not CP-only children (Fanti & Kimonis, 2017), while another found no differences in afternoon basal concentrations between adolescent antisocial subgroups (Kimonis et al., 2017a, 2017b). A third study found that higher basal cortisol levels in 15-month-olds was associated with more high-intensity negative emotional reactivity during a fear-inducing task relative to CP-only and low CP control groups, and predicted membership to a high CP/high CU group at first grade (Mills-Koonce et al., 2015). Infants in the same dataset with higher cortisol reactivity also showed greater antisocial behavior in first grade when they experienced lower maternal sensitivity, but not infants with lower cortisol reactivity (Wagner et al., 2017). Although the latter studies did not disaggregate primary and secondary CU variants, all children were recruited from a birth cohort in areas characterized by high social adversity (i.e., child poverty). Taken together, these findings provide preliminary support for the role of allostatic overload in the development of secondary CU traits, but require replication in other samples and using multi-system indices (e.g., RDoC’s Sustained Threat [“chronic stress”] construct; Goulter et al., 2023a reviews neuroendocrine and inflammatory correlates of CU traits).