Adversity and hyperresponsivity

Studies examining associations between early adversity and physiological reactivity have typically shown that in the contexts of early insensitive or abusive parenting, children develop heightened physiological reactivity to stress (Cicchetti & Rogosch, Reference Cicchetti and Rogosch2001a, Reference Cicchetti and Rogosch2001b; Essex, Klein, Cho, & Kalin, Reference Essex, Klein, Cho and Kalin2002; Hill-Soderlund et al., Reference Hill-Soderlund, Mills-Koonce, Propper, Calkins, Granger and Moore2008; Tarullo & Gunnar, Reference Tarullo and Gunnar2006). Early experiences of fear may sensitize developing corticolimbic pathways to react more readily to future threatening situations by upregulating neurobiological stress responses (Gunnar & Vazquez, Reference Gunnar and Vazquez2006; Heim & Nemeroff, Reference Heim and Nemeroff2001). Exaggerated HPAA responses have been noted in rhesus monkeys who have been separated from their biological mothers at birth and raised in peer-only groups (Dettmer, Novak, Suomi, & Meyer, Reference Dettmer, Novak, Suomi and Meyer2012; Suomi, Reference Suomi1997). However, increased HPAA activity has also been linked to more moderate experiences of adversity, such as attending full-time, low-quality, center-based care (Dettling, Parker, Lane, Sebanc, & Gunnar, Reference Dettling, Parker, Lane, Sebanc and Gunnar2000).

In addition to proximal indices of stress, chronic exposure to more distal environmental and socioeconomic risks has been linked with both elevated HPAA and ANS levels (Brady & Matthews, Reference Brady and Matthews2006; Evans & English, Reference Evans and English2002; Evans & Kim, Reference Evans and Kim2007; Wilson, Kliewer, Plybon, & Sica, Reference Wilson, Kliewer, Plybon and Sica2000). Different sources of proximal and distal risk may also interact to predict children's physiological reactivity. In a study of traumatized 1.5- to 6-year-olds who had experienced a life-threatening event or witnessed an event that threatened the life of an immediate relative, children who had high levels of posttraumatic stress disorder symptoms and were also exposed to low levels of positive parental discipline showed the highest PNS reactivity in response to recalling a traumatic personal event (Scheeringa, Zeanah, Myers, & Putnam, Reference Scheeringa, Zeanah, Myers and Putnam2004).

Heightened physiological responses in turn have been linked to increased vulnerability to psychopathology, especially internalizing symptoms (Boyce et al., Reference Boyce, Quas, Alkon, Smider, Essex and Kupfer2001; Cicchetti & Rogosch, Reference Cicchetti and Rogosch2001b; Schiefelbein & Susman, Reference Schiefelbein and Susman2006). However, not all children exposed to early adverse experience will develop an upregulated physiological phenotype. The next step for resilience researchers is to investigate whether differences in children's adaptive functioning following adversity exposure are mediated by changes in their physiological reactivity. It will also be important to examine how these processes may be moderated by the presence or absence of established protective factors. For example, children's gender and temperament have been reported to moderate the effects of stressful context on physiological reactivity (Fox, Henderson, Rubin, Calkins, & Schmidt, Reference Fox, Henderson, Rubin, Calkins and Schmidt2001; Gunnar, Kryzer, Van Ryzin, & Phillips, Reference Gunnar, Kryzer, Van Ryzin and Phillips2010). Similarly, sensitive caregivers can buffer children from exhibiting heightened physiological responses to stressful situations (Gilissen, Koolstra, van IJzendoorn, Bakermans-Kranenburg, & Van der Veer, Reference Gilissen, Koolstra, van IJzendoorn, Bakermans-Kranenburg and Van der Veer2007; Gunnar & Donzella, Reference Gunnar and Donzella2002). Researchers need to determine whether the emergence of resilience can be traced back to differential effects of adversity exposure on physiological reactivity or whether other factors enable resilient children to overcome vulnerabilities associated with heightened stress arousal.

Stress inoculation

For decades, researchers have argued that early exposure to psychological or physiological stress can actually be beneficial for adaptation by preparing children to deal with future experiences of stress and adversity. Various terms, such as “steeling effect” (Rutter, Reference Rutter1987, Reference Rutter2006), “inoculation effect” (Eysenck, Reference Eysenck and Cooper1983), and “psychophysiological toughness” (Dienstbier, Reference Dienstbier1989), have been used to describe this phenomenon. Children exposed to limited stress early in life are expected to develop dampened physiological reactivity over time, which makes them less physiologically vulnerable to subsequent stressful experiences. The stress-inoculation hypothesis adopts the metaphor of a vaccine to describe how early stressful experiences can inoculate children against future experiences of adversity.

The most rigorous empirical support for this hypothesis comes from animal research inspired by Levine's seminal work with rodents (Levine, Reference Levine1957). For example, postnatal handling involving very brief, daily separation of rodent pups from their mothers yields attenuated neuroendocrine responses to stress in adulthood (Meaney, Aitken, Bhatnagar, Van Berkel, & Sapolsky, Reference Meaney, Aitken, Bhatnagar, Van Berkel and Sapolsky1988; Meaney et al., Reference Meaney, Diorio, Francis, Widdowson, LaPlante and Caldji1996).

Similarly, Parker, Lyons, and colleagues showed that although early intermittent separation of squirrel monkeys from their natal group concurrently elicited elevated stress response, separated monkeys at 9 months of age showed lower levels of HPAA response to novel test environment than their nonseparated age mates (Parker, Buckmaster, Schatzberg, & Lyons, Reference Parker, Buckmaster, Schatzberg and Lyons2004). Monkeys that experienced early separations also displayed lower levels of anxiety symptoms, higher levels of cognitive control, and greater exploration of novel situations in comparison to the control group (Lyons, Parker, & Schatzberg, Reference Lyons, Parker and Schatzberg2010; Parker, Buckmaster, Sundlass, Schatzberg, & Lyons, Reference Parker, Buckmaster, Sundlass, Schatzberg and Lyons2006; Parker et al., Reference Parker, Rainwater, Buckmaster, Schatzberg, Lindley and Lyons2007). In contrast to rodent studies, which show that the effect of brief separations on stress reactivity is mediated by maternal increase in licking and grooming after reunion (Francis & Meaney, Reference Francis and Meaney1999; Liu et al., Reference Liu, Diorio, Tannenbaum, Caldji, Francis and Freedman1997), Parker and colleagues (Reference Parker, Buckmaster, Sundlass, Schatzberg and Lyons2006) found that the effect of stress inoculation on monkeys' resilience was not explained by temporary changes in maternal caregiving behaviors. Although these studies provide a compelling argument that early experiences of adversity can directly contribute to a more resilient physiological response to future stressors, the stress inoculation hypothesis needs to be prospectively examined in human children.

DiCorcia and Tronick (Reference DiCorcia and Tronick2011) have recently proposed the everyday stress resilience hypothesis, suggesting that resilience emerges from infants' everyday experiences with normative stressors rather than from experiences of overcoming extreme adversity. They argue that infant's interactions with caregivers generate numerous microstressors, for example, when parents miss or misread infant cues and demands. Such stressors provide the child with valuable experiences of dyadic mismatches generated by parents' failures to regulate infant physiology, affect, and behavior, as well as experiences of reparation that lead to successful dyadic coregulation. DiCorcia and Tronick (Reference DiCorcia and Tronick2011) use the metaphor of a runner who trains for a marathon by running short distances to describe the process by which infants develop resilience, or the increased capacity to regulate future stressors. Although studies show initial heightened physiological reactivity followed by recovery in infants exposed to laboratory simulation of dyadic mismatches using the face-to-face Still Face paradigm (Feldman, Singer, & Zagoory, Reference Feldman, Singer and Zagoory2010; Ham & Tronick, Reference Ham and Tronick2006; Moore et al., Reference Moore, Hill Soderlund, Propper, Calkins, Mills Koonce and Cox2009), it is unclear how these experiences directly contribute to individual differences in children's physiological reactivity to future adversity and whether infants whose caregivers lack the necessary amount of “reparatory sensitivity” can later develop resilience.

Furthermore, it is important to determine whether more severe experiences of stress and adversity also promote better adaptation over time. As reviewed by Bonnano, Brewin, Kaniasty, and Greca (Reference Bonanno, Brewin, Kaniasty and Greca2010), a few prospective studies show that adults who have previous experiences with disaster (e.g., flood, earthquake) are less negatively affected by new disaster experiences of a similar nature. Although there is a lack of similar studies in children, Wolmer, Hamiel, and Laor (Reference Wolmer, Hamiel and Laor2011) recently tested a school-based stress inoculation training intervention with fourth and fifth graders living in conflict areas of Israel. In comparison to the control group, children who were exposed to the intervention, designed to prepare them for potential trauma, showed significantly lower levels of posttraumatic symptoms following extended rocket attacks that occurred 3 months later.

One major impediment to a more comprehensive understanding of stress inoculation is the lack of common definitions for low, moderate, and high levels of adversity exposure across different developmental periods. For example, although 15-min separation of rodent pups from their mothers yields attenuated stress response later in life, maternal deprivation as indexed by hours of separation has been linked to heightened stress reactivity (van Oers, de Kloet, Li, & Levine, Reference van Oers, de Kloet, Li and Levine1998). Resilience researchers have historically regarded adversity exposure as unequivocally harmful to children's development. If we are to tease apart how and when experiences of early stressors induce physiological resilience to subsequent adversity, it is important to identify optimal levels of stress and arousal across development. Although DiCorcia and Tronick (Reference DiCorcia and Tronick2011) propose that midrange caregivers' “reparatory sensitivity” provides children with the right balance of mismatches and reparation to develop coping and regulatory skills, it remains unclear what midrange sensitivity represents in an absolute sense. Further, given that maternal sensitivity has been found to moderate infants' reactivity and recovery to stressful dyadic regulatory mismatches (Feldman et al., Reference Feldman, Singer and Zagoory2010; Moore et al., Reference Moore, Hill Soderlund, Propper, Calkins, Mills Koonce and Cox2009), researchers need to investigate whether experiences of early stress (analogous to primate models) or experiences of maternal sensitivity (analogous to rodent models) promote physiological resilience. Finally, it will be important to examine stress inoculation processes across different timeframes, ranging from seconds and minutes to weeks, months, and years. Such research would greatly benefit from standardized measures of adversity exposure that would enable comparison of findings across different samples.

Curvilinear models

If high adversity in early childhood relates to heightened physiological reactivity, whereas moderate adversity leads to a buffered physiological profile, it does not necessarily follow that that relation of contextual influences and physiological reactivity must be linear. Boyce and Ellis (Reference Boyce and Ellis2005) have used evolutionary principles to suggest that distribution of high biological sensitivity phenotypes is conditional on the quality of early childrearing context. According to their biological sensitivity to context theory, children who grow up in nurturing environments are likely to develop high biological sensitivity in order to better engage with the stimulating influences that surround them. Children who grow up in threatening environments are also likely to develop high biological sensitivity, in their case, as a protective strategy that helps maintain the vigilance necessary for survival. In contrast, the majority of children, whose early environments fall between these two extremes, neither nurturing nor threatening, are likely to develop lower biological sensitivity. Thus, Boyce and Ellis (Reference Boyce and Ellis2005) hypothesize that the relation between early experience and high biological sensitivity is curvilinear and U-shaped, with high sensitivity phenotypes prevalent in the contexts of low and high adversity exposure.

Although various studies provide indirect support for different segments of this U-shaped function, direct tests of the model have been rare, as empirical studies tend to measure a limited range of environmental influences or focus exclusively on linear associations between context and physiology. Using signal detection analyses with two separate samples, Ellis, Essex, and Boyce (Reference Ellis, Essex and Boyce2005) provided initial evidence of a positive curvilinear relation between the quality of contextual experiences and ANS reactivity. Recent studies reveal more complex nonlinear associations between adversity and physiological reactivity. Gunnar, Frenn, Wewerka, and Van Ryzen (Reference Gunnar, Frenn, Wewerka and Van Ryzin2009) showed that children who were adopted early from foster care homes (moderate adversity) showed lower levels of HPAA reactivity than both children who were raised in intact, affluent families (low adversity), and children who were adopted late from orphanages (high adversity). However, Engert and colleagues (Reference Engert, Efanov, Dedovic, Duchesne, Dagher and Pruessner2010) found the opposite pattern in adults. Adults who retrospectively reported medium levels of maternal care in childhood showed higher cortisol reactivity than those in low- and high-maternal care groups. Finally, my colleagues and I have explicitly tested quadratic relations between family context and children's HPAA activity, rather than employing group-based comparisons (Bush, Obradović, Adler, & Boyce, Reference Bush, Obradović, Adler and Boyce2011). Family socioeconomic status (SES) disadvantage, as indexed by parental income and education, and family adversity, as indexed by exposure to stressful life experiences such as parental depression, harsh parenting, and marital conflict, had opposite effects on physiology across the kindergarten school year. The relation between family SES risk and cortisol diurnal output was positively quadratic in fall, whereas the relation between family adversity and cortisol diurnal output was negatively quadratic in spring. Moreover, results show that these processes may be further influenced by children's ethnic background. It is essential that researchers continue investigating how nonlinear associations between contextual influences and physiological reactivity emerge and why they vary across different adversity contexts, individual differences, and developmental periods.

According to the biological sensitivity to context theory, a U-shaped distribution of high biological sensitivity phenotypes should produce the best survival odds for the entire population. However, studies still need to demonstrate how nonlinear relations between adversity and physiological phenotypes translate to overall distribution of adaptive functioning. For example, Seery, Holman, and Silver (Reference Seery, Holman and Silver2010) have recently reported significant quadratic relations between cumulative lifetime adversity and indices of adaptive functioning in a nationally representative sample of adults. Adults who were exposed to some adversity, but less than the sample's average, reported the lowest levels of global distress, functional impairment, and posttraumatic stress symptoms, and the highest levels of life satisfaction when compared to those who reported no adversity or high levels of adversity. More developmental studies of nonlinear associations among adversity exposure, adaptation, and physiological response are needed to better understand how the theorized U-shaped function between adversity and physiological reactivity contributes to the inverted U-shaped function between adversity and adaptation. Moreover, researchers need to examine variability in adaptation that is not explained by these curvilinear functions, as resilient individuals often defy population trends.

Adversity and hyporesponsivity

In sharp contrast to the work reviewed up to this point, a growing number of studies link adversity exposure with diminished HPAA activity, which may index overall desensitization of the stress response system to chronic negative experiences. For example, avoidant attachment classification in infancy predicted lower heart rate when children were 4 years old (Burgess, Marshall, Rubin, & Fox, Reference Burgess, Marshall, Rubin and Fox2003). Similarly, living with depressed mothers in low-income, urban areas of Mexico was related to children's lower baseline cortisol and lower cortisol reactivity in response to cognitive testing (Fernald, Burke, & Gunnar, Reference Fernald, Burke and Gunnar2008). Lower HPAA activity has also been observed in adolescents living in a relocation camp following exposure to Hurricane Katrina when compared to a demographically matched control sample (Vigil, Geary, Granger, & Flinn, Reference Vigil, Geary, Granger and Flinn2010).

Furthermore, extreme levels of neglect have also been associated with dampened HPAA activity. Children who were raised in Romanian and Russian orphanages show blunted early morning cortisol and flattened diurnal rhythms (Carlson & Earls, Reference Carlson and Earls1997; Gunnar, Morison, Chisholm, & Schuder, Reference Gunnar, Morison, Chisholm and Schuder2001; Gunnar & Vazquez, Reference Gunnar and Vazquez2001). Hyporesponsivity has also been reported in children living in foster homes in the United States (Bruce, Fisher, Pears, & Levine, Reference Bruce, Fisher, Pears and Levine2009; Dozier et al., Reference Dozier, Manni, Gordon, Peloso, Gunnar and Stovall-McClough2006). The link between experiences of adversity and downregulated stress physiology is further strengthened by studies showing that preventive interventions aimed at improving children's caregiving environment have an effect on normalizing dampened HPAA diurnal rhythms (Cicchetti, Rogosch, Toth, & Sturge-Apple, Reference Cicchetti, Rogosch, Toth and Sturge-Apple2011; Dozier, Peloso, Lewis, Laurenceau, & Levine, Reference Dozier, Peloso, Lewis, Laurenceau and Levine2008; Fisher & Stoolmiller, Reference Fisher and Stoolmiller2008; Fisher, Van Ryzin, & Gunnar, Reference Fisher, Van Ryzin and Gunnar2011). Given the current state of the literature, it is more accurate to state that high adversity exposure may lead to dysregulated physiological phenotypes than to a particular pattern of hyper- or hyporesponsivity.

To account for the presence of the physiologically underresponsive phenotype in the context of high adversity, Del Giudice, Ellis, and Shirtcliff (Reference Del Giudice, Ellis and Shirtcliff2011) have recently proposed the adaptive calibration model, which builds on biological sensitivity to context theory. According to this model, two different phenotypes emerge in the context of high adversity. Exposure to frequent stressors associated with dangerous or unpredictable environments may give rise to a vigilant profile characterized by high SNS and HPAA basal levels and reactivity and low vagal tone and PNS reactivity. It is hypothesized that these vigilant individuals will experience high levels of anxiety and heightened sensitivity to contextual threats, which is expressed in females through withdrawn behaviors and in males though agonistic behaviors. In contrast, exposure to severe, chronic and traumatic stress may give rise to an unemotional profile characterized by generally low basal activity and reactivity across both HPAA and ANS systems in response to social-evaluative and performance-related stressors. Del Guidice and colleagues (Reference Del Giudice, Ellis and Shirtcliff2011) suggest that unemotional males are more likely to engage in proactive, instrumental aggression and show high levels of impulsivity, callousness, and risk taking, whereas unemotional females may show relational aggression, detached social relationships, and low levels of parental investment. The adaptive calibration model provides an important set of testable hypotheses that not only accounts for some divergent findings in literature, but also takes into account the specificity of different stress response systems. The fact that experiences of high adversity have been linked to both hypo- and hyperresponsive phenotypes opens an important research area for resilient researchers. If both of these phenotypes predispose individuals to higher risk of developing internalizing and externalizing behaviors, it is important to determine whether and how physiological reactivity differentiates resilient and maladaptive functioning.

Epigenetics

Physiological reactivity is subserved by a set of neurobiological processes that can be studied at deeper, subcellular levels of analysis (Marsland, Bachen, Cohen, Rabin, & Manuck, Reference Marsland, Bachen, Cohen, Rabin and Manuck2002). Alterations in epigenetic modifications of chromatin structure, DNA methylation, and histone acetylation in response to acute or chronic stressors can calibrate the transcription of genes involved in the development and regulation of physiological reactivity. Recent studies show significant associations between early adverse experiences and variability in the expression and epigenetic modification of stress-responsive genes. For example, animal models reveal that variation in early maternal care is related to DNA methylation of the regulatory region of the gene encoding the glucocorticoid receptor (Meaney, Reference Meaney2010; Szyf, Weaver, Champagne, Diorio, & Meaney, Reference Szyf, Weaver, Champagne, Diorio and Meaney2005; Weaver et al., Reference Weaver, D'Alessio, Brown, Hellstrom, Dymov and Sharma2007). In addition, increased expression of genes related to sympathetic innervation of lymphoid tissues has been observed in rhesus macaques exposed to unstable social hierarchy structures (Sloan et al., Reference Sloan, Capitanio, Tarara, Mendoza, Mason and Cole2007).

In humans, a study of suicide victims has linked history of childhood abuse to epigenetic regulation of the glucocorticoid receptor in the hippocampus (McGowan et al., Reference McGowan, Sasaki, D'Alessio, Dymov, Labonté and Szyf2009). Oberlander et al. (Reference Oberlander, Weinberg, Papsdorf, Grunau, Misri and Devlin2008) showed that fetal exposure to third-trimester maternal depression is related to increased methylation of the glucocorticoid receptor gene and exaggerated salivary cortisol responses to a stressor at 3 months of age. Miller and colleagues (Reference Miller, Chen, Fok, Walker, Lim and Nicholls2009) found that low SES in early childhood was associated with differential expression of stress-related genes in adulthood, controlling for concurrent social class, perceived stress, and lifestyle practices. More relevant to resilience researchers, retrospective report of maternal warmth was found to moderate the link between low SES and epigenetically modified proinflammatory response (Chen, Miller, Kobor, & Cole, Reference Chen, Miller, Kobor and Cole2011). Recently, Essex and colleagues (in press) linked early parental stress, assessed during infancy and preschool, to differential methylation in adolescents. Together, these studies represent a new scientific frontier that has the potential to discover the role that early experiences play in “programming” individuals' stress responsivity on a molecular level across time. Given that both prenatal and neonatal experiences have been linked to differences in physiological responsivity (Davis, Glynn, Waffarn, & Sandman, Reference Davis, Glynn, Waffarn and Sandman2011; Grunau et al., Reference Grunau, Holsti, Haley, Oberlander, Weinberg and Solimano2005), epigenetic analyses may help us better understand when and at what level experience-based alterations to physiological response occur.

Adaptation and allostatic load

Although this section has focused primarily on demonstrating links between various levels and types of adversity and measures of physiological reactivity, implicit in this review is the notion that dysregulated physiological reactivity is associated with maladaptive functioning, as indexed by the presence of psychopathology symptoms and absence of competence. It is important to note that from an evolutionary perspective, both hyperresponsive and hyporesponsive physiological profiles in the context of adversity may be considered adaptive, as they secure better survival odds and a chance to pass on genes even if they reflect behavioral strategies that include higher levels of aggression, risk taking, and vigilant behavior (cf. Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011; Ellis, Boyce, Belsky, Bakermans-Kranenburg, & van IJzendoorn, Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and van Ijzendoorn2011). For example, Meaney and colleagues reported that although rodent pups exposed to low maternal care become more anxious, physiologically reactive adults with poorer learning capacity, they also show better adjustment and learning in stressful contexts (Bagot, Tse, Nguyen, Wong, & Meaney, Reference Bagot, Tse, Nguyen, Wong and Meaney2011). In humans, high physiological reactivity has been linked to higher levels of anxious and vigilant behaviors among maltreated children (Cicchetti & Rogosch, Reference Cicchetti and Rogosch2001b; Tarullo & Gunnar, Reference Tarullo and Gunnar2006), but for children facing life-threatening adversity, heightened arousal may be protective as it enables them to respond quicker to potentially dangerous events. Nevertheless, developmental psychopathology and resilience researchers consider high levels of internalizing and externalizing behaviors maladaptive, as they represent a risk for low achievement of salient developmental tasks such as formation of positive social relationships and successful attainment of academic or work skills (Masten & Obradović, Reference Masten and Obradović2006). Further, if circumstances change in a way that makes the environment less threatening, extreme vigilance may lead to maladaptive behaviors in response to ambiguous or neutral events.

Notwithstanding the potential short-term benefits of high reactivity, a prolonged state of physiological and psychological arousal can have serious long-term consequences for both physical and mental health. Chronic activation of stress response systems may lead to cumulative “wear and tear” on the body, a phenomenon known as an allostatic load (McEwen, Reference McEwen1998, Reference McEwen2000). Allostatic load has been linked to maladaptive family processes (Repetti, Robles, & Reynolds, Reference Repetti, Robles and Reynolds2011) as well as higher levels of health problems in adulthood (Seeman et al., Reference Seeman, Gruenewald, Karlamangla, Sidney, Liu and Mcewen2010). Thus, resilience researchers need to rethink their definition of adaptation and also weigh the benefits and costs of short-term and long-term adaptation when considering implications of children's physiological reactivity for resilience processes. As we move forward, it will be crucial that these conceptual differences in the definition of adaptive functioning be addressed.

Summary

Researchers have made significant progress in understanding how experiences of adversity relate to children's physiological responsivity. However, most studies do not take into account that both experience and physiology can change over time. We need to go beyond relating childhood adversity to a single measure of concurrent or subsequent physiological reactivity and more systematically examine how early experience shapes longitudinal trajectories of physiological reactivity. Two recent studies provide great exemplars for future work. Blair, Raver, Granger, Mills-Koonce, and Hibel (Reference Blair, Raver, Granger, Mills-Koonce and Hibel2011) have showed how change in infants' resting HPAA activity at 7, 15, 24, and 48 months of age relates to early home instability and economic insufficiency. Using latent growth curve analyses, Cicchetti and colleagues (Reference Cicchetti, Rogosch, Toth and Sturge-Apple2011) have demonstrated that maltreated infants who were randomly assigned to interventions aimed at reducing effects of maltreatment showed a similar cortisol pattern to a nonmaltreated comparison group at 13, 19, 26, and 38 months of age, whereas maltreated children who were assigned to standard community services displayed increasingly lower levels of morning cortisol. In addition, it will be important to investigate transactional effects between longitudinal changes in both context and reactivity across different developmental periods and determine whether there are periods during which context exerts stronger influence on physiological reactivity. For example, a 2-year longitudinal study revealed that initial levels in vagal tone predicted changes in maternal restrictive behaviors, whereas the opposite was not the case (Kennedy, Rubin, Hastings, & Maisel, Reference Kennedy, Rubin, Hastings and Maisel2004). Finally, given that resilience develops over time and is shaped by the confluence of various negative and positive processes, it is important to examine how physiological reactivity interacts with both risk and protective factors and how the timing and length of different experiences influences this interplay.

Biological sensitivity to context

At the heart of resilience research is the notion that children vary in the extent to which they are affected by adverse experiences. Early researchers in the field described resilient children as “invulnerable,” implying that there is some intrinsic trait that protects them from deleterious environmental effects. Although researchers have since abandoned this term, recognizing that resilience requires a complex set of processes that enable at-risk children to overcome the odds (Masten & Obradović, Reference Masten and Obradović2006), the notion that some children are more vulnerable to stress has prevailed. Children who are behaviorally or biologically reactive have been found to be more vulnerable to environmental adversity than their less reactive peers (Obradović & Boyce, Reference Obradović and Boyce2009). In the last decade, numerous studies have shown that reactivity moderates adversity effects such that children who have more difficult temperaments, a more reactive physiology, or particular gene alleles are at higher risk for maladaptive outcomes in the context of high adversity (Caspi et al., Reference Caspi, McClay, Moffitt, Mill, Martin and Craig2002; Caspi et al., Reference Caspi, Sugden, Moffitt, Taylor, Craig and Harrington2003; El-Sheikh, Keller, & Erath, Reference El-Sheikh, Keller and Erath2007; Fowles et al., Reference Fowles, Kochanska and Murray2000; Morris et al., Reference Morris, Silk, Steinberg, Sessa, Avenevoli and Essex2002). Such findings reflect what researchers long ago described as the diathesis–stress model (Gottesman & Shields, Reference Gottesman and Shields1967; Monroe & Simons, Reference Monroe and Simons1991), also known as the dual-risk model, which posits that dual or increased risk arises from an interaction between the child's internal risk or diathesis and exposure to environmental risk. Accordingly, high physiological reactivity has been considered a marker of diathesis or vulnerability for all children, whereas low reactivity is often seen as characteristic of resilient children.

However, two recent evolutionary-based theories have partially challenged this view of reactivity, converging on the idea that heightened biological and behavioral reactivity make some children more susceptible than others to both high and low quality environments (Ellis et al., Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and van Ijzendoorn2011). Boyce and Ellis (Reference Boyce and Ellis2005) have proposed that physiological reactivity should be reconceptualized as biological sensitivity to context, arguing that high physiological reactivity may be maladaptive in contexts of adversity, but healthy and promotive in contexts of nurturance and protection (Boyce & Ellis, Reference Boyce and Ellis2005; Ellis et al., Reference Ellis, Essex and Boyce2005). Conversely, children with low physiological reactivity may be less affected by the quality of their surroundings. Similarly, Belsky and colleagues have advanced the differential susceptibility hypothesis, which posits that children who have more difficult or reactive temperaments as indexed by high negative emotionality are also more susceptible, “for better and for worse,” to both positive and negative environmental influences (Belsky, Reference Belsky, Ellis and Bjorklund2005; Belsky, Bakermans-Kranenburg, & van IJzendoorn, Reference Belsky, Bakermans-Kranenburg and van IJzendoorn2007).

A growing body of empirical work supports these theories, using various behavioral, physiological, and genetic constructs as indices of susceptibility (Belsky & Pluess, Reference Belsky and Pluess2009; Obradović & Boyce, Reference Obradović and Boyce2009). Most relevant to this review, individual differences in physiological ANS and HPAA reactivity to laboratory challenges have been found to moderate the effects of family environment on children's adaptive functioning (Boyce et al., Reference Boyce, Essex, Alkon, Goldsmith, Kraemer and Kupfer2006; Cummings, El-Sheikh, Kouros, & Keller, Reference Cummings, El-Sheikh, Kouros and Keller2007; El-Sheikh et al., Reference El-Sheikh, Keller and Erath2007; Obradović, Bush, Stamperdahl, Adler, & Boyce, Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010). For example, in a community sample of kindergarten children, my colleagues and I found that high PNS reactivity exacerbated risk for maladaptation in the context of high family adversity, whereas in the context of low adversity it promoted higher levels of academic achievement, school competence, and prosocial behaviors (Obradović et al., Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010). Recent studies have shown that physiological reactivity also moderates the effects of nonfamilial contexts, as indexed by peer victimization and teacher–child relationship quality (Essex, Armstrong, Burk, Goldsmith, & Boyce, Reference Essex, Armstrong, Burk, Goldsmith and Boyce2011; Rudolph, Troop-Gordon, & Granger, Reference Rudolph, Troop-Gordon and Granger2010).

It is important to emphasize that this new conceptualization of high stress reactivity as high biological sensitivity or susceptibility does not contradict the traditional dual-risk and diathesis–stress models, but rather it extends them. These theories converge on the idea that in contexts of adversity high reactivity indeed represents a risk factor for maladaptive developmental outcomes. Consequently, low reactivity offers protection against negative life experiences and may be linked to resilience adaptation. However, although many studies have demonstrated a link between low reactivity and better adjustment in the context of adversity, we should not assume that only low reactivity is associated with resilience. For resilience researchers, the most important tenet of the biological sensitivity to context theory is that high reactivity is not a unitary construct that always implies vulnerability, but rather it represents an index of plasticity (Boyce & Ellis, Reference Boyce and Ellis2005; Ellis et al., Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and van Ijzendoorn2011). Moreover, the effect of high physiological sensitivity on developmental outcomes depends on the context to which the child is reacting to. Because resilience often emerges over time from the combination of both positive and negative influences, it is important to identify when it is adaptive for at-risk children to show high physiological reactivity. Finally, resilience researchers need to keep in mind that physiological reactivity is not a fixed trait, but a complex, dynamic construct that emerges as a product of early genetic and environmental influences and can also be shaped by later contextual experiences.

The nature of stressors

In examining the role of physiological reactivity for resilience processes, researchers need to investigate whether the role of reactivity changes across different types of adversity. In contrast to studies demonstrating that high PNS reactivity can serve as an index of biological sensitivity to context (Obradović et al., Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010; Shannon, Beauchaine, Brenner, Neuhaus, & Gatzke-Kopp, Reference Shannon, Beauchaine, Brenner, Neuhaus and Gatzke-Kopp2007), researchers studying the effects of marital conflict and domestic violence on children's adaptation have found that high PNS reactivity acts as a protective factor that buffers children against the effects of marital adversity. Adaptation of children who displayed high PNS reactivity to laboratory challenges was less influenced by the presence of marital conflict, whereas children with low PNS reactivity were reported to be more vulnerable, exhibiting more behavioral problems, poorer health, and lower academic achievement, in the context of high marital conflict or domestic violence (El-Sheikh, Harger, & Whitson, Reference El-Sheikh, Harger and Whitson2001; El-Sheikh & Whitson, Reference El-Sheikh and Whitson2006; Katz, Reference Katz2007; Katz & Gottman, Reference Katz and Gottman1997). High PNS reactivity to interpersonal stress also emerged as a protective factor against the effects of hostile-withdrawn parenting on peer conflict (Leary & Katz, Reference Leary and Katz2004). These studies suggest that children who come from conflictual families are less likely to show behavioral and health problems if they show higher physiologically reactivity to interpersonal laboratory challenges. In these situations, high reactivity may be indicative of better emotion regulation and active coping when faced with interpersonal challenges in their daily lives.

Intrigued by the apparent discrepancies between these findings and studies showing high PNS reactivity as a plasticity factor (Obradović et al., Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010; Shannon et al., Reference Shannon, Beauchaine, Brenner, Neuhaus and Gatzke-Kopp2007), my colleagues and I have recently examined whether the link between ANS reactivity and children's adaptation in the context of marital conflict differs with the nature of the laboratory challenge task employed to elicit and measure children's stress response. We have hypothesized that laboratory challenges that embody the types of adversity salient in children's lives may have different predictive value than laboratory challenges that are less personal. We have found that children who displayed high RSA reactivity to a cognitive challenge (i.e., number recall) and children who displayed low RSA reactivity to an interpersonal challenge (i.e., watching bullying movie) were most affected by the degree of marital conflict exposure (Obradović, Bush, & Boyce, Reference Obradović, Bush and Boyce2011a). In other words, high PNS reactivity to cognitive and interpersonal challenges may indicate different markers of sensitivity to marital conflict and thus relate in a contrasting manner to resilient adaptation. Although these findings need to be replicated, examining the predictive validity of reactivity in response to different laboratory stimuli offers a way to reconcile two seemingly contradictory lines of research, both examining PNS physiology as a moderator of environmental effects on adaptation.

Other research has also demonstrated the importance of considering the nature of physiological stressors. Hinnant and El-Sheikh (Reference Hinnant and El-Sheikh2009) reported that physiological reactivity to different types of laboratory challenges uniquely predicted different behavioral outcomes in children. The interactive effect of baseline RSA and reactivity to an interpersonal stressor emerged as a significant predictor of later internalizing symptoms, whereas the interactive effect of baseline RSA and reactivity to a cognitive stressor emerged as a significant predictor of later externalizing symptoms. Moreover, physiological reactivity has been shown to be highly context-specific, as children do not exhibit consistently high or low reactivity across different laboratory challenges (Alkon et al., Reference Alkon, Goldstein, Smider, Essex, Kupfer and Boyce2003; Chen, Matthews, Salomon, & Ewart, Reference Chen, Matthews, Salomon and Ewart2002; Quigley & Stifter, Reference Quigley and Stifter2006; Salomon, Matthews, & Allen, Reference Salomon, Matthews and Allen2000). The salience of laboratory challenges may also change with time as children grow older and encounter different types of stressors in their environment. For example, Stroud and colleagues (Reference Stroud, Foster, Papandonatos, Handwerger, Granger and Kivlighan2009b) have found different patterns of physiological responsivity to performance and peer rejection stressors between children and adolescents. Together these studies point to the need to attend not only to the context in which children are growing up but also to the nature of the stimuli to which children are responding when reactivity is evoked and measured. One way to identify when and how high reactivity contributes to resilient adaptation is through gaining a better understanding of the role that laboratory context plays in approximating how children's bodies react to different types of real-life stressors.

The nature of physiological systems

Although different physiological systems respond to environmental threats, demands, and challenges in a highly integrated manner, emerging evidence suggests a possible degree of specificity across different indices of ANS and HPA activity. According to Porges' polyvagal theory, PNS activity reflects one's ability to engage with and adapt to environmental challenges without activating the metabolically costly flight or fight response (Porges, Reference Porges2003, Reference Porges2007). Because PNS also inhibits the activation of SNS stress response via vagal brake, RSA has been frequently regarded as a measure of children's capacity for self-regulation (Beauchaine, Reference Beauchaine2001; Porges, Reference Porges2007). Studies have linked both high baseline RSA and RSA reactivity to processes that support positive adaptation and coping such as self-soothing behaviors, attention focusing, positive emotion regulation strategies, and higher levels of executive functioning (Blair & Peters, Reference Blair and Peters2003; Calkins, Reference Calkins1997; Calkins & Keane, Reference Calkins and Keane2004; Fabes, Eisenberg, & Eisenbud, Reference Fabes, Eisenberg and Eisenbud1993; Feldman & Eidelman, Reference Feldman and Eidelman2009; Fox & Field, Reference Fox and Field1989; Marcovitch et al., Reference Marcovitch, Leigh, Calkins, Leerks, O'Brien and Blankson2010; Suess, Porges, & Plude, Reference Suess, Porges and Plude1994). Further, some studies suggest that RSA baseline levels may be better markers of biological sensitivity, as they reflect an organism's preparedness or anticipation to respond to stressors and challenges. For example, high basal RSA has been found to protect children from the effects of marital conflict on behavior problems, academic achievement, and health, whereas low baseline RSA places children at risk for various behavioral problems (El-Sheikh et al., Reference El-Sheikh, Harger and Whitson2001; El-Sheikh & Whitson, Reference El-Sheikh and Whitson2006; Katz & Gottman, Reference Katz and Gottman1995, Reference Katz and Gottman1997). Focusing on PNS activity may be especially relevant for understanding coping with everyday mild to moderate stressors that may be particularly challenging to children who are also exposed to chronic adversity.

In contrast, SNS is activated by a threatening situation that requires active avoidance of a stressor by mobilization of the fight or flight response (Porges, Reference Porges2003). Beauchaine and colleagues have further proposed that two indices of SNS response may be linked to behavioral approach and behavioral inhibition systems (Beauchaine et al., Reference Beauchaine, Gatzke-Kopp and Mead2007; Brenner, Beauchaine, & Sylvers, Reference Brenner, Beauchaine and Sylvers2005). Studies show that SNS reactivity as indexed by changes in skin conductance level can be elicited by the threat of punishment, and thus may reflect individual differences in punishment avoidance, inhibition, anxiety, and fearfulness (Beauchaine et al., Reference Beauchaine, Gatzke-Kopp and Mead2007; Fowles et al., Reference Fowles, Kochanska and Murray2000; Shannon et al., Reference Shannon, Beauchaine, Brenner, Neuhaus and Gatzke-Kopp2007). Conversely, SNS reactivity as indexed by changes in preejection period seems to be particularly sensitive to rewards or anticipation of rewards, and as such may reflect individual differences in approach motivation, reward sensitivity, and aggressive tendencies (Beauchaine et al., Reference Beauchaine, Gatzke-Kopp and Mead2007; Brenner et al., Reference Brenner, Beauchaine and Sylvers2005; Crowell et al., Reference Crowell, Beauchaine, Gatzke-Kopp, Sylvers, Mead and Chipman-Chacon2006). Analogous to SNS, HPAA has evolved to respond to direct threats to one's well-being. Because researchers cannot subject participants to significant physical threats in a laboratory context, Dickerson and Kemeny (Reference Dickerson and Kemeny2004) conducted a meta-analysis of adult studies to show that HPAA reactivity is most reliably activated by social threats to self that are novel, unpredictable, and uncontrollable. Various laboratory protocols have been developed to study children's and adolescent's HPAA reactivity in response to threats elicited by social evaluation and peer rejection (Stroud et al., Reference Stroud, Foster, Papandonatos, Handwerger, Granger and Kivlighan2009a; Yim, Quas, Cahill, & Hayakawa, Reference Yim, Quas, Cahill and Hayakawa2010). HPAA activity has also been studied outside the laboratory context by examining how various environmental contexts relate to changes in HPAA diurnal rhythm (Cicchetti, Rogosch, Gunnar, & Toth, Reference Cicchetti, Rogosch, Gunnar and Toth2010; Dozier et al., Reference Dozier, Manni, Gordon, Peloso, Gunnar and Stovall-McClough2006), or how emotional states predict momentary and daily changes in HPAA activity (Doane & Adam, Reference Doane and Adam2010).

Researchers need to bear in mind the potentially unique roles that different physiological systems have when incorporating measures of physiological reactivity in studies of resilience. First, it is important to choose laboratory protocols that are carefully designed to elicit and capture variability in physiological reactivity of a particular system. For example, SNS reactivity to fear may be best captured by exposing young children to a mild threat, because severe threat may appropriately elicit high levels of arousal in all children (Buss, Davidson, Kalin, & Goldsmith, Reference Buss, Davidson, Kalin and Goldsmith2004). Furthermore, stressors without explicit threat to self may not elicit HPAA reactivity, potentially confounding poor methodology with bona fide nonsignificant findings. Second, what constitutes a marker of high susceptibility to context may vary across different systems. Although high SNS and HPAA reactivity have been more consistently shown to represent high susceptibility to context (El-Sheikh et al., Reference El-Sheikh, Keller and Erath2007; Rudolph et al., Reference Rudolph, Troop-Gordon and Granger2010), some studies suggest that low PNS reactivity better captures children's susceptibility to context, as children high in PNS reactivity appear to less affected by environmental influences (El-Sheikh & Whitson, Reference El-Sheikh and Whitson2006; Hastings et al., Reference Hastings, Sullivan, McShane, Coplan, Utendale and Vyncke2008). Recent studies also show that different measures of the HPAA diurnal pattern, such as cortisol awakening response, morning and evening levels, or day to day variability, rather than cortisol reactivity to laboratory challenges, may be better indicators of environmental sensitivity (Adam et al., Reference Adam, Doane, Zinbarg, Mineka, Craske and Griffith2010; Corbett, Schupp, Levine, & Mendoza, Reference Corbett, Schupp, Levine and Mendoza2009; Kudielka & Wüst, Reference Kudielka and Wüst2009). A more comprehensive study of how different markers of stress response systems vary in their associations with adversity and adaptation may help explain some of the current inconsistencies in literature and advance our understanding of how physiological reactivity contributes to resilient adaptation.

Severity and comorbidity of symptoms

In addition to considering the nature of stressors and systems that are employed to represent children's physiological reactivity, researchers need to be cognizant of the nature of their sample when studying physiological reactivity. For example, high RSA reactivity has been linked to better emotion regulation and attention in community samples of children (Blair, Reference Blair2003; Calkins, Reference Calkins1997; Suess et al., Reference Suess, Porges and Plude1994) as well as to high levels of behavior problems in clinical samples (Beauchaine et al., Reference Beauchaine, Gatzke-Kopp and Mead2007; Crowell et al., Reference Crowell, Beauchaine, McCauley, Smith, Stevens and Sylvers2005). This inconsistency, however, may be partially accounted for by the fact that most studies employ standardized residual scores to represent children's physiological reactivity. This means that reactivity scores tend to reflect how an individual child compares to other children in the sample rather than the child's own absolute level of arousal. High reactivity in a community sample may be equivalent to more moderate reactivity in a clinical sample, whereas high reactivity in a clinical sample may translate to excessive reactivity in a community sample. Therefore, it is crucial that researchers start providing thresholds for low, moderate, high, and excessive reactivity in absolute terms, rather than defining reactivity only in relative terms that may vary across different samples. Such benchmarks may also shed important light on whether high rather than excessive physiological reactivity can promote resilient adaptation.

Resilience researchers have long recognized that adaptation is a multidimensional construct and that children's functioning should be concurrently assessed and considered across various domains (Buckner, Mezzacappa, & Beardslee, Reference Buckner, Mezzacappa and Beardslee2003; Masten et al., Reference Masten, Burt, Roisman, Obradovic, Long and Tellegen2004; Obradović, Reference Obradović2010). This may be especially important to take into account when investigating how stress physiology relates to adaptation in the context of adversity. Some inconsistencies in the literature can be explained by the comorbidity of symptoms that show opposite relations with physiological reactivity. For example, although pure externalizing problems tend to be associated with low reactivity, co-occurring internalizing and externalizing behavior problems have been associated with heightened reactivity in both community and clinical samples (Boyce et al., Reference Boyce, Quas, Alkon, Smider, Essex and Kupfer2001; Calkins et al., Reference Calkins, Graziano and Keane2007; Kibler, Prosser, & Ma, Reference Kibler, Prosser and Ma2004). Likewise, the established negative relations between HPAA activity and externalizing symptoms may not be reliably found in maltreated children, because they tend to show elevated HPAA levels, especially if they display internalizing symptoms (Tarullo & Gunnar, Reference Tarullo and Gunnar2006). In a recent study of toddlers exposed to interparental aggression, Davies, Sturge-Apple, and Cicchetti (Reference Davies, Sturge-Apple and Cicchetti2011) further demonstrated the complexity of links between reactivity and adaptation. They reported that 1-year increase in cortisol reactivity was related to both increase in internalizing symptoms and decrease in attention and hyperactivity problems. Resilience researchers should measure adaptation across multiple domains of functioning as well as employ person-focused analytic approaches to further our understanding of how physiological reactivity relates to resilient and maladaptive adaptation.

Summary

A compelling body of work shows that physiological reactivity can act as a powerful moderator of adversity effects and, as such, is fundamental to our understanding of resilience processes. Biological sensitivity to context theory (Boyce & Ellis, Reference Boyce and Ellis2005) offers a new way of conceptualizing physiological reactivity as an index of plasticity rather than vulnerability, such that highly reactive children are more susceptible to both positive and negative environmental influences. However, at least one tenet of the biological sensitivity to context theory, which is that high physiological reactivity represents a risk factor in the context of high adversity, needs to be revisited, as numerous studies have linked higher levels of physiological reactivity with resilient adaptation, suggesting that the role of physiological reactivity may change across different types of stressors, biological systems, behavioral outcomes, and samples. Highly reactive children, aside from being more affected by adverse experiences, may also be more susceptible to processes known to promote resilience, such as prosocial peer groups, positive parenting, alternate mentors, or afterschool programs, as well as preventive intervention programs. Unfortunately, there is a lack of studies examining how individual differences in physiological reactivity interact with positive experiences promoting resilience. Future studies thus need to provide a more nuanced conceptualization of different physiological responses as measures of susceptibility to various positive and negative environmental influences concurrently present in the child's life.

Physiological response trajectories

Traditional approaches for calculating physiological reactivity involve measuring the child's physiological response to a nonchallenging situation such as sitting still (i.e., baseline or pretask measure) and a laboratory challenge (i.e., challenge or posttask measure). The two values are then used to derive a single difference or a residual score that represents the change or reactivity from baseline in response to the challenge. If study protocol includes multiple challenges, reactivity scores are averaged or examine separately in analyses. Researchers studying HPAA reactivity sometimes obtain more than two measures of salivary cortisol in order to better capture the response pattern of this slower acting system. Multiple assessment points are frequently used to calculate the “area under the curve,” which represents the total physiological arousal during laboratory protocol (Pruessner, Kirschbaum, Meinlschmid, & Hellhammer, Reference Pruessner, Kirschbaum, Meinlschmid and Hellhammer2003). Alternatively, children are classified into “responders” or “nonresponders” based on an empirically or theoretically determined threshold that signifies a meaningful increase in physiological arousal.

All these approaches yield a single index of reactivity, which tends to oversimplify the dynamic nature of physiological response to a challenge. It is crucial that future studies incorporate multiple physiological measures that capture variability in onset, offset, and duration of the physiological arousal. For example, Brooker and Buss (Reference Brooker and Buss2010) showed that both dynamic and static measures of RSA change during a fear-inducing challenge were important in understanding toddlers' affect and emotion behaviors. In addition, Santucci et al. (Reference Santucci, Silk, Shaw, Gentzler, Fox and Kovacs2008) found that lower RSA recovery, but not reactivity, following an emotional challenge was associated with poor emotion regulation strategies in response to frustration. Growth curve modeling can be used to examine whether trajectories of stress response show linear or nonlinear change during a task, across a day, or over a longer period of time. These techniques also allow researchers to identify various static and dynamic covariates of stress response trajectories. For example, Shirtcliff and Essex (Reference Shirtcliff and Essex2008) used growth curve analysis to show how interindividual variability in basal cortisol levels and diurnal trajectories relates to adolescent mental health during fifth and seventh grade. Examining the entire trajectory of a physiological response to a stressor may help determine whether all children in high-risk contexts show similar levels of arousal in response to an acute stressor, but resilient children also demonstrate faster physiological recovery indicative of better self-regulation strategies.

Van Ryzin, Chatham, Kryzer, Kertes, and Gunnar (Reference Van Ryzin, Chatham, Kryzer, Kertes and Gunnar2009) have recently demonstrated advantages and disadvantages of two analytic approaches for identifying different trajectories of the HPAA diurnal rhythm. They show how growth curve modeling can be a useful technique to identify predictors of the variability around the sample's average trajectory response in situations when one can assume a certain level of homogeneity in response patterns. On the other hand, group-based trajectory modeling can be employed with heterogeneous populations to identify distinct groups of children who show dramatically different response patterns. Researchers can also apply these techniques to understanding how the time course of ANS response to an acute stressor relates to variability in adaptation. A group-based trajectory modeling may be particularly effective in identifying atypical response patterns associated with resilient adaptation. In contrast, growth curve modeling would enable researchers to investigate time-varying covariates of physiological reactivity.

Dynamic interplay among multiple physiological systems

Most studies linking physiological reactivity and children's behaviors focus only on markers of one biological system, despite the fact that stressful or challenging experiences trigger a set of highly integrated physiological responses that implicate multiple systems. Although decades ago Berntson and colleagues recommended studying so-called autonomic space by taking into account the simultaneous activation of both the SNS and PNS (Berntson, Cacioppo, Quigley, & Fabro, Reference Berntson, Cacioppo, Quigley and Fabro1994), only a few studies have examined concurrent SNS and PNS activity in children. Some researchers use a person-focused analytic approach by first identifying four distinct ANS reactivity profiles: (a) coactivation, as indexed by SNS and PNS activation; (b) coinhibition, as indexed by SNS and PNS withdrawal; (c) reciprocal parasympathetic activation, as indexed by SNS withdrawal and PNS activation; and (d) reciprocal sympathetic activation, as indexed by SNS activation and PNS withdrawal. Once the groups have been identified, researchers can compare them on various behavioral outcomes. Other researchers employ a variable-focused approach by examining how the two-way interaction between continuous measures of SNS and PNS reactivity relates to children's behavior.

Studies that employ these analytic strategies reveal that the concurrent activity and interaction of the two ANS branches differentiates groups of children who show distinct patterns of behavioral problems or susceptibilities to contextual influences (Boyce et al., Reference Boyce, Quas, Alkon, Smider, Essex and Kupfer2001; El-Sheikh et al., Reference El-Sheikh, Kouros, Erath, Cummings, Keller and Staton2009; Keller & El-Sheikh, Reference Keller and El-Sheikh2009). In a recent monograph, El-Sheikh and colleagues (Reference El-Sheikh, Kouros, Erath, Cummings, Keller and Staton2009) showed that, in the context of high marital conflict, coinhibition and coactivation of the SNS and PNS systems served as a risk factor for behavioral problems, whereas reciprocal activation offered some protection. However, the prevalence of the four ANS profiles seems to vary across different developmental periods and types of challenges (Alkon et al., Reference Alkon, Goldstein, Smider, Essex, Kupfer and Boyce2003; Salomon et al., Reference Salomon, Matthews and Allen2000), emphasizing the need to investigate whether the interplay between the two systems is influenced by exposure or adaptation to various contextual experiences.

Examining the interaction effects of ANS and HPAA reactivity may also further understanding of processes that contribute to variability in adaptation. According to Bauer, Quas, and Boyce (Reference Bauer, Quas and Boyce2002), higher risk for behavioral problems may be found in children who show either (a) high or low activation of both the ANS and HPAA systems (the additive model); or (b) unbalanced, asymmetrical activation of both systems (the interactive model). A few recent studies provide initial evidence for the additive model. For example, adolescents who displayed both low HPAA and SNS reactivity showed highest levels of externalizing symptoms (Gordis, Granger, Susman, & Trickett, Reference Gordis, Granger, Susman and Trickett2006), whereas children who displayed high levels of both SNS and HPA reactivity were at highest risk for internalizing and externalizing symptoms (El-Sheikh, Erath, Buckhalt, Granger, & Mize, Reference El-Sheikh, Erath, Buckhalt, Granger and Mize2008). However, a recent study supports the interactive model by showing that lower HPAA activity was related to lower self-esteem only in adolescents who also demonstrated high SNS activity (Vigil et al., Reference Vigil, Geary, Granger and Flinn2010).

In addition, researchers have started to examine how children's baseline or anticipatory levels of arousal interact with physiological reactivity to challenges to predict behavior. For example, Hinnant and El-Sheikh (Reference Hinnant and El-Sheikh2009) reported that the interplay between children's baseline RSA value and RSA regulation in response to an interpersonal stressor predicted behavior problems 2 years later. Low basal RSA and high RSA reactivity presented the highest risk for internalizing, whereas low basal RSA and low RSA reactivity presented the highest risk for externalizing symptoms. It is thus important that future research considers how both the individual's physiological readiness and the actual stress response work together to promote adaptation. In a recent paper, Del Giudice, Hinnant, Ellis, and El-Sheikh (in press) have taken a first step to test the hypotheses put forth in the adaptation calibration model by using latent profile analysis to identify four groups of children who show similar basal and reactivity SNS and HPAA responses. Future studies need to explicitly investigate how physiological responsivity profiles and patterns are associated with positive adaptation in the context of adversity. Studies that reveal that reciprocal PNS and SNS activation or asymmetrical SNS and HPA levels provide some protection against symptoms of psychopathology provide an initial direction for identifying physiological profiles associated with resilient adaptation.

The role of appraisal and regulation

Although physiological reactivity has been shown to moderate the effects of context on adaptation, it is generally considered to be an unconscious and automatic response to environmental challenges and stressors. This characterization of physiological reactivity can conjure up the image of a passive child who is more or less affected by environment, depending on whether he possesses high or low reactivity phenotype. We know this to be far from the truth. First, children's interpretations of context can determine the degree to which it appears threatening. For example, hostile interpretations of ambiguous stimuli have been found to partially mediate concurrent and longitudinal associations between socioeconomic disadvantage and increased ANS reactivity in both children and adolescents (Chen, Langer, Raphaelson, & Matthews, Reference Chen, Langer, Raphaelson and Matthews2004). In a recent study, Parker and colleagues found that stress-inoculated monkeys may differentially appraise stressors in adulthood, showing physiological reactivity similar to controls in response to severe stressors, but dampened reactivity in response to moderate stressors (Parker, Buckmaster, Lindley, Schatzberg, & Lyons, Reference Parker, Buckmaster, Lindley, Schatzberg and Lyons2011). Second, children's appraisal of physiological arousal may also moderate the relation between physiological reactivity and adaptive behavior. In a recent study, Jamieson and colleagues experimentally manipulated college students' appraisal of physiological arousal (Jamieson, Mendes, Blackstock, & Schmader, Reference Jamieson, Mendes, Blackstock and Schmader2010). Students who were taught to reappraise their arousal as a performance-enhancing experience had higher levels of SNS reactivity and better scores on both practice and actual GRE math tests than students who were not assigned to the reappraisal condition.

Finally, children vary in their ability to effortfully regulate arousal and control physiological recovery. Studies link moderate levels of physiological reactivity to better executive functioning skills in children (Blair, Granger, & Razza, Reference Blair, Granger and Razza2005; Marcovitch et al., Reference Marcovitch, Leigh, Calkins, Leerks, O'Brien and Blankson2010), and it is feasible that some children who show high physiological responsivity to challenges also display better cognitive control over that arousal. Understanding the role of executive functioning skills in self-regulating arousal will be an important next step for resilience researchers. Although exposure to high levels of environmental adversity may predispose most children to develop highly sensitive physiological phenotypes, resilient children may develop self-regulatory skills that produce fast and efficient recovery. In other words, high reactivity and good self-regulation skills may differentiate resilient children from children whose high reactivity and poor self-regulation skills puts them at risk for developing behavioral problems. At the same time, the link between chronic exposure to stressors (e.g., poverty, harsh parenting, neglect) and poor self-regulation skills, as indexed by executive functions, may be partially explained by the vulnerability of the prefrontal cortex to stress hormones. In adult and animal samples, researchers have shown that the activation of both HPAA and SNS systems compromises the activity of the brain regions subserving executive functions (Arnsten, Reference Arnsten2009; Arnsten & Li, Reference Arnsten and Li2005; Lupien, McEwen, Gunnar, & Heim, Reference Lupien, McEwen, Gunnar and Heim2009). Heightened stress reactivity has been linked to poor performance on EF tasks (Blair, Reference Blair2010; Lisonbee, Pendry, Mize, & Gwynn, Reference Lisonbee, Pendry, Mize and Gwynn2010; Qin, Hermans, van Marle, Luo, & Fernández, Reference Qin, Hermans, van Marle, Luo and Fernández2009; Williams, Suchy, & Rau, Reference Williams, Suchy and Rau2009). Given that a recent study also shows that higher HPA activity mediates the effect of family adversity on children's cognitive abilities (Blair, Granger, et al., Reference Blair, Granger, Willoughby, Mills-Koonce, Cox and Greenberg2011), it will be important to further investigate the complex relation that may emerge between physiological arousal and executive functioning in the context of adversity.

Longitudinal continuity and change

Researchers studying physiological responsivity to normative stressful events, such as starting a new school year and creating new friends, have shown that the role of physiological reactivity can vary across time. Gunnar and colleagues have reported that elevated cortisol during the first weeks of school, when children forge new peer groups and connections, may be an adaptive response, as it has been associated with higher social competence (Gunnar, Tout, de Haan, Pierce, & Stansbury, Reference Gunnar, Tout, de Haan, Pierce and Stansbury1997). Conversely, elevated cortisol later in the school year, when peer groups have been fully formed, has been linked to solitary and negative social behaviors. A recent study by Tarullo, Mliner, and Gunnar (Reference Tarullo, Mliner and Gunnar2011) extended these findings by examining how social dynamics change over a school year and how children's temperamental differences affect social interactions. Children who showed high and average levels of exuberance and dominance showed a decrease in cortisol levels across a preschool year. In contrast, highly inhibited children showed two different longitudinal patterns. Those who were popular and formed friendships showed an increase in cortisol levels across the school year, whereas those who had no friends showed consistently high cortisol levels. It appears that for dominant, outgoing kids it is adaptive to show elevated HPAA response only at the beginning of the year while negotiating a position in a new social hierarchy. In contrast, it may be adaptive for shy and risk-aversive children to show low reactivity in fall when dominant peers “fight” for their hierarchy positions, followed by an increase in spring as they work to overcome their avoidance tendencies in order to form friendships. My colleagues and I have also found that links between high physiological reactivity to laboratory challenges and teacher-reported competence in kindergarten varied across the school year and as a function of children's dominance status and gender (Obradović, Bush, & Boyce, Reference Obradović, Bush and Boyce2011b). Together these studies highlight the need to examine how relations between physiological reactivity and adaptation change as both the contextual landscape and one's role in it evolve, even over short periods of time.

Unfortunately, many studies examine only concurrent links between children's physiological reactivity, adaptive functioning, and adversity exposure. Even studies linking measures of physiology and adversity to subsequent indices of adaptation rarely employ repeated measures that can reveal how these constructs interact over time. Future studies should employ developmental models to identify transactional effects between physiological reactivity, positive and negative experiences, and adaptation across time and different developmental periods. It will be important to determine not only how main and interactive effects of adversity and reactivity on adaptation change over time, but also whether changes in adaptation and context in turn can shape stability of physiological reactivity.

Summary

Researchers need to adopt a more dynamic approach to measuring, analyzing and interpreting various processes representing physiological reactivity to environmental stimuli. Future studies need to address the dynamic nature of physiological reactivity both on a micro, momentary level as well as on a macro, longitudinal level. A first set of studies will need to take into account variability in (a) response trajectories by capturing difference in children's anticipation, reactivity and recovery, (b) synchronization of multiple biological systems that can either augment or diminish each other's effects, and (c) children's appraisal of and control over their physiological arousal. A second set of studies will need to investigate how associations between physiological response, adaptation, and adversity change over time by employing advanced analytic strategies such as growth curve modeling, latent transition analysis, and structural equation modeling. Although it is beyond the scope of this paper, researchers will also need to consider how the interplay between physiological reactivity, adversity, and adaptation also vary as a function of children's genetic variability and temperamental tendencies (Obradović & Boyce, Reference Obradović and Boyce2009). Epigenetic modifications may represent a key mechanism in linking different processes. For example, a recent study shows that the interplay between genetic polymorphisms and the degree of promoter region DNA methylation significantly predicted unresolved emotional state of mind regarding early experiences of loss and trauma (van IJzendoorn, Caspers, Bakermans-Kranenburg, Beach, & Philibert, Reference van IJzendoorn, Caspers, Bakermans-Kranenburg, Beach and Philibert2010).