Decades of research demonstrate that exposure to environmental adversity places children and adults at elevated risk for developing cognitive, social, emotional, and health problems (Boyce, Reference Boyce and Masten2007; Luthar, Reference Luthar1999; McLoyd, Reference McLoyd1998; Shonkoff, Boyce, & McEwen, Reference Shonkoff, Boyce and McEwen2009). Although it is well established that disorders of development and health are more prevalent among individuals from high-risk families, there is striking variation in the psychological adjustment and physical health of children and adults exposed to both low and high levels of adversity (Luthar, Reference Luthar, Cicchetti and Cohen2006; Masten & Obradović, Reference Masten and Obradović2006). In recent years, researchers have made significant progress in understanding how environmental exposures interact with genotypes and phenotypes to differentially shape human development. It has become increasingly clear that individuals with different characteristics vary not only in whether and how much they are negatively affected (in terms of conventionally defined mental health outcomes) by environmental stressors and adversity (e.g., Caspi et al., Reference Caspi, McClay, Moffitt, Mill, Martin and Craig2002, Reference Caspi, Sugden, Moffitt, Taylor, Craig and Harrington2003) but also in the extent to which they are positively influenced by environmental resources and supports (Bakermans-Kranenburg & van IJzendoorn, Reference Bakermans-Kranenburg and van IJzendoorn2011 [this issue]; Blair, Reference Blair2002; Kochanska, Kim, Barry, & Philibert, in press; Quas, Bauer, & Boyce, Reference Quas, Bauer and Boyce2004). Most notable, however, is the recurrent finding, further highlighted by the articles in this Special Section, that the very characteristics of individuals that make them disproportionately vulnerable to adversity sometimes also make them disproportionately likely to benefit from contextual support (e.g., Bakermans-Kranenburg & van IJzendoorn, Reference Bakermans-Kranenburg and van IJzendoorn2007; Belsky, Reference Belsky1997a, Reference Belsky, Ellis and Bjorklund2005; Boyce & Ellis, Reference Boyce and Ellis2005; Boyce et al., Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995). In this paper, we propose that such individual differences in susceptibility to the environment have several defining characteristics:

1. 1. Individuals characterized by heightened environmental susceptibility display enhanced sensitivity to both negative and positive environments, that is, to both risk-promoting and development-enhancing environmental conditions.

2. 2. This enhanced sensitivity increases developmental receptivity to the environment. That is, more susceptible individuals are more likely to experience sustained developmental change, not just transient fluctuations in functioning, in response to environmental exposures

3. 3. Susceptibility to the environment is instantiated in the biology of the nervous system; it is neurobiological susceptibility. Genetic susceptibility factors operate through neurobiological processes and behavioral indicators of susceptibility are grounded in neurobiology.

4. 4. Developmental experience plays a role, along with heritable polygenic variation, in determining individual differences in neurobiological susceptibility.

5. 5. Individuals of all ages (children and adults) vary in neurobiological susceptibility to environmental influences and, within individuals, susceptibility may vary across the life span.

6. 6. Individual differences in neurobiological susceptibility are adaptive in the evolutionary sense and have been conserved by fluctuating selective pressures that generate different fitness payoffs across different social, physical, and historical contexts (or at least did so during the course of human evolution).

7. 7. Variation in neurobiological susceptibility to the environment, therefore, constitutes a central mechanism in the regulation of alternative patterns of human development; specifically, differential susceptibility moderates the effects of environmental exposures on developmental and life outcomes. Ultimately, this means that the development of some individuals, more than others, will be influenced by their experiences and environments (even if these were exactly the same).

We begin this paper by outlining the defining features of the currently dominant perspective on environmental risk and human frailty: diathesis–stress/dual risk. We then review challenges to this perspective based on emerging theory and data from the evolutionary, developmental, and health sciences. These challenges highlight the need for a paradigm shift in conceptualizing Person × Environment interactions in development. In this context we frame an evolutionary–neurodevelopmental theory of how individual differences in neurobiological susceptibility to the environment may account for the impressive variation in the potency of environmental effects on adaptation, development, and health.

The Dominant Paradigm: Diathesis–Stress/Dual Risk

Students of human development widely appreciate that individuals vary in whether and how much they are negatively affected by environmental stressors, ranging from exposures to poverty to hostile or insensitive parenting to low quality child care, to name just a few well-studied phenomena. Perhaps the most striking evidence that personal characteristics condition or moderate such environmental effects is found in developmental research on Temperament × Parenting interactions (Rothbart & Bates, Reference Rothbart, Bates, Eisenberg, Damon and Lerner2006) and psychiatric research on Gene × Environment (G × E) interactions (Burmeister, McInnis, & Zollner, Reference Burmeister, McInnis and Zollner2008). Work in both areas is guided primarily, even if not exclusively, by what developmentalists regard as the transactional/dual-risk model (Sameroff, Reference Sameroff and Mussen1983) and what psychiatrists and others studying psychopathology regard as the diathesis–stress model (Gottesman & Shields, Reference Gottesman and Shields1967; Monroe & Simons, Reference Monroe and Simons1991; Zuckerman, Reference Zuckerman1999). Central to both frameworks is the view that some individuals, because of a specific “vulnerability” that may be behavioral in character (e.g., difficult temperament), physiological or endophenotypic in nature (e.g., heightened biological reactivity to stress), or genetic in origin (e.g., serotonin linked polymorphic region [5-HTTLPR] short alleles), are disproportionately or even exclusively likely to be affected adversely by an environmental stressor. According to prevailing views, it is the child with a “difficult” (or negatively emotional) temperament, for instance, or individuals carrying certain “vulnerability genes” or “risk alleles” who are most likely to develop or function poorly, including manifesting psychopathological conditions such as depression, when they are exposed to adversity.

The “dual-risk” designation derives from the synergistic effect of a risk (or diathesis) inherent in the individual interacting with one operative in the environment. Where some people are regarded as especially susceptible to adversity because of their personal vulnerabilities, other people lacking such vulnerabilities who do not succumb to the adversity in question are considered to be resilient (Cicchetti, Reference Cicchetti1993; Cicchetti & Garmezy, Reference Cicchetti and Garmezy1993; Luthar, Reference Luthar, Cicchetti and Cohen2006; Masten & Obradović, Reference Masten and Obradović2006), often as a result of personal protective factors (e.g., low stress reactivity, nonrisk genotypes). Implicit in the diathesis–stress framework is the view that the children and adults who are vulnerable or resilient because of their personal characteristics respond more or less similarly to nonadverse and supportive or enriched environmental conditions. Thus, central to the diathesis–stress view is the assumption that vulnerable and resilient individuals develop differently principally under conditions of environmental stress. One consequence of this assumption, as illustrated in Figure 1, is that many studies do not measure either the full range of environments (just adversity and its absence, e.g., maltreatment vs. no maltreatment) or a full range of psychological/behavioral functioning (just dysfunction and its absence, e.g., depressed vs. not depressed). Until recently, little attention has been paid to whether, or the extent to which, personal characteristics condition or moderate the effects of supportive or enriched environmental contexts on highly competent functioning and positive well-being.

Figure 1.

The diathesis–stress/dual risk model (solid black and solid gray lines) and the differential susceptibility model (solid gray and dotted gray lines). The two models are partly overlapping, and selection of a restricted range of environments (i.e., just adversity and the absence of adversity, left-side rectangle) renders the two models indiscernible. In this context, negative outcomes are experienced only by individuals displaying high susceptibility/developmental vulnerability. A focus on positive environments as well as inclusion of the full range of environments reveals the difference between the diathesis–stress and differential susceptibility models. As shown in the right-side rectangle, the differential susceptibility model contends that more susceptible individuals in positive environments will show more favorable outcomes (i.e., developmental enhancement); more susceptible individuals are thus disproportionately influenced by both negative and positive environments. Adapted from “Genetic Vulnerability or Differential Susceptibility in Child Development: The Case of Attachment [Research Review],” by M. Bakermans-Kranenburg and M. H. van IJzendoorn, Reference Van Zeijl, Mesman, Stolk, Alink, van IJzendoorn and Bakermans-Kranenburg2007, Journal of Child Psychology and Psychiatry, 48, 1160–1173. Copyright 2007 by Wiley–Blackwell. Adapted with permission.

Beyond Diathesis–Stress/Dual Risk: Evolutionary Models of Adaptive Development

An evolutionary perspective challenges the prevailing developmental psychopathological analysis of dysfunctional or maladaptive outcomes within settings of adversity. In particular, it contends that both stressful and supportive environments have been part of human experience throughout our evolutionary history, and that developmental systems shaped by natural selection respond adaptively to both kinds of contexts. Thus, when people encounter stressful environments, this does not so much disturb their development as direct or regulate it toward strategies that are adaptive under stressful conditions, even if those strategies are currently harmful in terms of the long-term welfare of the individual or society as a whole (see Hinde & Stevenson-Hinde, Reference Hinde and Stevenson-Hinde1990; Main, Reference Main1990).

Consider the extensive experimental work conducted by Michael Meaney and colleagues showing that putatively low quality maternal care in the rat (i.e., low levels of maternal licking and grooming) alters pups' stress physiology and brain morphology. Although such changes seem disadvantageous (i.e., higher corticosterone levels, shorter dendritic branch lengths, and lower spine density in hippocampal neurons), they actually enhance learning and memory processes under stressful conditions (Champagne et al., Reference Champagne, Bagot, van Hasselt, Ramakers, Meaney and de Kloet2008). Moreover, such physiological and morphological changes mediate the effects of maternal behavior on central features of defensive and reproductive strategies: behavior under threat, open-field exploration, pubertal development, sexual behavior, and parenting (Cameron et al., Reference Cameron, Champagne, Parent, Fish, Osaki-Kuroda and Meaney2005, Reference Cameron, Del Corpo, Diorio, Mackallister, Sharma and Meaney2008); and they do so in ways consistent with evolutionary models of adaptive reproductive strategies (Belsky, Steinberg, & Draper, Reference Belsky, Steinberg and Draper1991; Chisholm, Reference Chisholm1999).

In the rodent model, then, enhanced learning under stressful conditions, increased fearful and defensive behaviors, accelerated sexual maturation, increased sexual behavior, and reduced parental investment in offspring apparently represent strategic (i.e., functional) ways of developing when neglected. In this context, neglect itself can be regarded as a mechanism through which rat parents guide their offspring's development toward optimal survival and reproductive strategies under conditions of adversity. It would seem mistaken therefore to view diminished licking and grooming as “poor maternal care” or the development induced by such care as “disturbed.” From an evolutionary perspective, the care provided by the putatively neglectful parents may be appropriate preparation of their offspring for the ecological conditions into which they are likely to mature.

It is important to note that optimal adaptation (in the evolutionary sense) to challenging environments is not without real consequences and costs. Harsh environments often harm or kill children, and the fact that children developmentally adapt to such rearing conditions (reviewed in Ellis, Figueredo, Brumbach, & Schlomer, Reference Ellis, Figueredo, Brumbach and Schlomer2009; Pollak, Reference Pollak2008) does not imply that such conditions either promote child well-being or should be accepted as unmodifiable facts of life (i.e., David Hume's “naturalistic fallacy”). There can be no doubt that high-stress environments that are dangerous and lack essential resources, compared with low-stress environments that are safe and well resourced, undermine fitness. Developmental adaptations to high-stress environments enable individuals to make the best of a bad situation (i.e., to mitigate the inevitable fitness costs), even though “the best” may still constitute a high-risk strategy that jeopardizes the person's health and survival (e.g., Mulvihill, Reference Mulvihill2005; Shonkoff et al., Reference Shonkoff, Boyce and McEwen2009). Further, there are genuinely novel environments, such as Romanian or Ukrainian orphanages (Dobrova-Krol, van IJzendoorn, Bakermans-Kranenburg, & Juffer, Reference Dobrova-Krol, van IJzendoorn, Bakermans-Kranenburg and Juffer2010; Nelson et al., Reference Nelson, Zeanah, Fox, Marshall, Smyke and Guthrie2007), that are beyond the normative range of conditions encountered over human evolution. Children's brains and bodies simply could not have been selected to respond adaptively to collective rearing by paid, custodial, nonkin caregivers (Hrdy, Reference Hrdy1999). Exposures to such challenging yet (evolutionarily) unprecedented conditions can be expected to induce pathological development, not evolutionarily adaptive strategies.

In sum, an evolutionary–developmental perspective emphasizes conditional adaptation: “evolved mechanisms that detect and respond to specific features of childhood environments, features that have proven reliable over evolutionary time in predicting the nature of the social and physical world into which children will mature, and entrain developmental pathways that reliably matched those features during a species' natural selective history” (Boyce & Ellis, Reference Boyce and Ellis2005, p. 290; for a comprehensive treatment of conditional adaptation, see West-Eberhard, Reference West-Eberhard2003; for applications to human development, see Belsky et al., Reference Belsky, Steinberg and Draper1991; Chisholm, Reference Chisholm1999; Ellis, Reference Ellis2004). From within such a perspective, the highly susceptible child who responds to a dangerous environment by developing insecure attachments, adopting an opportunistic interpersonal orientation, and sustaining an early sexual debut is no less functional than the context-sensitive child who responds to a well-resourced and supportive social environment by developing the opposing characteristics and orientations. A further implication is that efforts to reduce the pain and suffering of children growing up under stressful conditions need to take into consideration the local sense in which risky and seemingly self-destructive behaviors may be adaptive. Children have evolved to function competently, that is, to survive and ultimately reproduce, in a variety of contexts. The default assumption should be that alternative patterns of development in response to both stressful and supportive environmental conditions (within the range encountered over human evolution) constitute adaptive variation.

Evolutionary–Developmental Theories of Differential Susceptibility

In addition to shaping species-typical developmental responses to diverse environmental conditions, natural selection has also maintained variation (adaptive individual differences) in neurobiological susceptibility to the environment. Two different evolutionary accounts of such variation have emerged in recent years: biological sensitivity to context theory (BSCT; Boyce et al., Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995; Boyce & Ellis, Reference Boyce and Ellis2005; Ellis, Essex, & Boyce, Reference Ellis, Essex and Boyce2005) and differential susceptibility theory (DST; Belsky, Reference Belsky1997a, Reference Belsky1997b, Reference Belsky, Ellis and Bjorklund2005; Belsky, Bakermans-Kranenburg, & van IJzendoorn, Reference Belsky, Bakermans-Kranenburg and van IJzendoorn2007).Footnote ¹ With a central focus on Person × Environment interactions, both models advance the role of organismic characteristics in moderating the effects of both stressful and supportive environmental conditions on human development (see Figure 1). Further, both presume that there are individual differences in sensitivity to environmental influence but not in the manner of diathesis–stress. Both theories incorporate the traditional diathesis–stress view while extending it, by making the critical observation that those individuals most likely to be adversely affected (according to conventional mental health standards) by negative environmental conditions are also most likely to benefit from supportive ones.

Focusing on the role of the person in moderating environmental effects on development, DST and BSCT follow in the tradition of Bronfenbrenner's (1993) person–process–context model, which posits that parenting and other environmental factors may vary in their developmental influence as a function of the characteristics of the child. The theories also converge with Wachs and Gandour's (1983) organismic–specificity hypothesis, which posits differential reactivity of children to similar rearing experiences as a function of attributive differences in children's cognitive, behavioral, or emotional characteristics.

Before proceeding with our comparison of BSCT and DST, it is important to note that these perspectives share much in common with Aron and Aron's (1997; Aron, Aron, & Davies, Reference Aron, Aron and Davies2005) theory of sensory-processing sensitivity (SPS). The main difference is that DST and BSCT began with a focus on child-developmental processes, whereas SPS started with a focus on cognitive processes in adults (in terms of variation in depth of processing of sensory information, with more sensitive individuals processing information more thoroughly before acting). Although some SPS studies have included retrospective assessments of childhood experiences, SPS did not originate from a developmental perspective. As such, there is not a validated measure of SPS in children, and longitudinal work on the role of SPS in regulating child (or adult) developmental outcomes has not been conducted. Accordingly, it is premature to attempt to integrate SPS theory with BSCT and DST at this time. Nonetheless, as SPS moves into the child-developmental domain and BSCT and DST more systematically engage adult cognitive processes, such an integration is likely.

BSCT

The concept of biological sensitivity to context has its early roots in a report by Boyce and colleagues (1995), presenting two studies of naturally occurring environmental adversities and biological reactivity as predictors of respiratory illnesses in 3- to 5-year-old children. First, the results revealed that children showing low cardiovascular or immune reactivity to stressors had approximately equal rates of respiratory illnesses in both low and high adversity settings.Footnote ² Second, and consistent with the prevailing diathesis–stress model, highly biologically reactive children exposed to high adversity child care settings or home environments had substantially higher illness incidences than all other groups of children. Third, the unexpected finding was that highly reactive children living in lower adversity conditions (i.e., more supportive child care or family settings) had the lowest illness rates, which were significantly lower than even low reactivity children in comparable settings. Boyce et al. (Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995, p. 419) concluded that

. . . a subgroup of children may exist that sustains hyperdynamic biological responses to psychologically stressful events and experiences. Based on the results presented here, such children might be expected to encounter poorer health in high-stress contexts and unusually positive health outcomes in low-stress contexts. One plausible explanation for such a pattern of findings is the possibility that reactive children are more sensitive or more susceptible to the characteristics of the social environment. . . . Children with a heightened sensitivity to psychosocial processes emanating in the environment might then be expected to experience unusually poor outcomes in high-stress, unsupportive social conditions. The same children might flourish, on the other hand, under low-stress, nurturing, and predictable conditions . . .

Boyce et al. (Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995) thus advanced what would later become known as the differential susceptibility hypothesis (which Belsky, Reference Belsky1997a, Reference Belsky1997b, promulgated on a purely theoretical basis, without awareness of the work of Boyce and colleagues): that children differ in their susceptibility to environmental influence in a “for better and for worse” manner (with worse defined in terms of psychopathology and physical health problems, not necessarily worse fitness outcomes, an issue to be addressed in more detail below). Further, the initial Boyce et al. (Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995) research, together with subsequent work (Boyce & Ellis, Reference Boyce and Ellis2005), identified a physiological mechanism of environmental susceptibility—autonomic, adrenocortical, or immune reactivity to psychosocial stressors—and proposed that psychobiologic reactivity moderated the effects of early environmental exposures on physical and mental health outcomes in a bivalent manner. More reactive children displayed heightened sensitivity to both positive and negative environmental influences and thus were given the shorthand designation of orchid children, signifying their special susceptibility to both highly stressful and highly nurturing environments. In contrast, children low in reactivity were designated as dandelion children, reflecting their relative ability to function adequately in species-typical circumstances of all varieties (Boyce & Ellis, Reference Boyce and Ellis2005).

Although the findings of Boyce et al. (Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995) stimulated a provisional interpretation of how environmental exposures and psychobiologic reactivity worked together in regulating children's mental and physical health, conspicuously missing was a broader, more heuristic theoretical framework in which psychobiologic reactivity could be interpreted and explained. Boyce and Ellis's (2005; see also Ellis & Boyce, Reference Ellis and Boyce2008; Ellis et al., Reference Ellis, Essex and Boyce2005; Ellis, Jackson, & Boyce, Reference Ellis, Jackson and Boyce2006) BSCT was an effort to provide such an evolutionary functional analysis, advancing as it did two key propositions. The first involved a new hypothesis about the function of the stress response systems and the second a novel evolutionary hypothesis about the developmental calibration of these systems. Each is considered here.

With respect to the function of the stress response systems, it was clear that biological reactivity to stressors comprised an integrated system of central neural and peripheral neuroendocrine responses designed to prepare the organism for challenge or threat. In contrast, according to BSCT, these “stress response” systems also function to increase susceptibility to resources and support in the ambient environment (e.g., positive social opportunities, cooperative information). This dual function signified the need to conceptualize stress reactivity more broadly as biological sensitivity to context, which Boyce and Ellis (Reference Boyce and Ellis2005) defined as neurobiological susceptibility to both cost-inflicting and benefit-conferring features of the environment and operationalized as an endophenotypic property indexed by heightened reactivity in one or more of the stress response systems. Depending on levels of nurturance and support versus harshness and unpredictability in their developmental environments, highly reactive children experience either the best or the worst of psychiatric and biomedical outcomes within the populations from which they are drawn (Boyce, Reference Boyce, Bornstein and Genevro1996; Boyce et al., Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995, Reference Boyce, Essex, Alkon, Goldsmith, Kraemer and Kupfer2006; Bubier, Drabick, & Breiner, Reference Bubier, Drabick and Breiner2009; Ellis, Shirtcliff, Boyce, Deardorff, & Essex, Reference Ellis, Shirtcliff, Boyce, Deardorff and Essex2011 [this issue]; Essex, Armstrong, Burk, Goldsmith, & Boyce, Reference Essex, Armstrong, Burk, Goldsmith and Boyce2011 [this issue]; Obradović, Bush, & Boyce, Reference Obradović, Bush and Boyce2011 [this issue]; Obradović, Bush, Stamperdahl, Adler, & Boyce, Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010; Quas et al., Reference Quas, Bauer and Boyce2004). BSCT therefore posits that individual differences in the magnitude of biological stress responses function to regulate openness or susceptibility to environmental influences, ranging from harmful to protective.

Given past evidence that early trauma can increase stress reactivity and newer evidence that high reactivity can enhance developmental functioning in highly supportive settings, Boyce and Ellis (Reference Boyce and Ellis2005) postulated a curvilinear, U-shaped relation between levels of early support–adversity and the magnitude of biological response dispositions (see Figure 2). Specifically, Boyce and Ellis hypothesized that (a) exposure to acutely stressful childhood environments upregulates biological sensitivity to context, increasing the capacity and tendency of individuals to detect and respond to environmental dangers and threats; (b) exposure to especially supportive childhood environments also upregulates biological sensitivity to context, increasing susceptibility to social resources and support; and (c) by contrast and typical of the majority of children, exposure to childhood environments that are not extreme in either direction downregulates biological sensitivity to context, buffering individuals against the chronic stressors encountered in a world that is neither highly threatening nor consistently safe. Exploratory analyses in two studies offered confirmatory evidence that the lowest prevalences of high reactivity phenotypes were found in conditions of moderate stress and that both tails of the support–adversity distribution were associated with higher proportions of reactive children (Ellis et al., Reference Ellis, Essex and Boyce2005; see also Gunnar, Frenn, Wewerka, & Van Ryzin, Reference Gunnar, Frenn, Wewerka and Van Ryzin2009).

Figure 2.

The hypothesized curvilinear relation between experiences of support and protection versus stress and adversity in early environments and biological sensitivity to context. Comparisons of subjects at points A and B lead to the conclusion that low support/high stress results in heightened biological sensitivity to context. In contrast, comparisons at points C and D generates the inference that low support/high stress produces diminished biological sensitivity to context. Adapted from “Biological Sensitivity to Context: I. An Evolutionary–Developmental Theory of the Origins and Functions of Stress Reactivity,” by W. T. Boyce and B. J. Ellis, 2005, Development and Psychopathology, 17, 271–301. Copyright 2005 by Cambridge University Press. Adapted with permission.

In sum, BSCT advanced the claim that developmental variation in biological sensitivity to context has been maintained by natural selection, because differences in biological sensitivity to context reliably produced different fitness outcomes in different childhood environments encountered over evolutionary history. It is important that in a parallel but independent trajectory of theoretical work, Belsky's (Reference Belsky1997a, Reference Belsky1997b, Reference Belsky, Ellis and Bjorklund2005) DST was concurrently emerging from a very different set of evolutionary observations and ideas.

Toward an Integrated Differential Susceptibility Paradigm

Even though BSCT and DST emerged independently and differ in important respects, they share much in common. First, both theories, based as they are on an evolutionary analysis of human development, advance the claim that individuals differ systematically in their susceptibility to environmental influences and seek to explain the nature of such individual differences. As summarized below, this begins with the recognition that susceptibility to the environment is instantiated in the multiple genetic polymorphisms, endophenotypic mechanisms, and behavioral phenotypes that operate as susceptibility factors, moderating the influence of environmental exposures on developmental and life outcomes. Endophenotypes constitute a necessary link between genes and behavior, whereby (single or multiple) genetic markers of differential susceptibility operate through neurobiological processes and behavioral indicators of differential susceptibility are grounded in neurobiology. Consequently, whatever the level of analysis employed in a given study, neurobiological susceptibility to the environment is the fundamental construct of interest. This raises the possibility that the genotypically, endophenotypically, and behaviorally susceptible individuals identified in various studies may actually be the same people (Belsky et al., Reference Belsky, Bakermans-Kranenburg and van IJzendoorn2007; Obradovic & Boyce, 2009).

Second, DST and BSCT both presume that individuals should differ in their neurobiological susceptibility to environmental influence, and that such differential susceptibility underlies many reliable interactions between features of persons and features of environments in guiding human development and functioning. Both theories embrace, implicitly if not explicitly, Bronfenbrenner's (1979) dictum that, when it comes to human development, the main effects are in the interactions, whether they are G × E, Temperament × Parenting, Stress Reactivity × Family Stress, or other Person × Environment processes. More specifically, central to both BSCT and DST is the assumption that environmental influences on developmental and life outcomes are moderated by neurobiological susceptibility to the environment. Both theories thus conceptualize variation in neurobiological susceptibility as a central mechanism in the regulation of alternative patterns of human development.

Third, neither theory presumes that differential susceptibility is restricted to any one developmental period or even just to childhood; both appreciate that at all developmental stages during the life span, individuals may differ in the extent to which they are affected by both supportive and challenging environments. Along these lines, research documenting differential susceptibility in children, adolescents, and adults is reviewed below. Nonetheless, BSCT and DST were both originally developed to explain neurobiological susceptibility in childhood, as clearly reflected in the preceding summaries of these perspectives. Consequently, even though the theories logically extend to adulthood, relatively little is known about differential susceptibility across the life span (such as stability vs. change).

Fourth, both DST and BSCT define susceptible individuals as experiencing sustained change in response to environmental exposures. This explicit focus on at least somewhat enduring developmental change distinguishes BSCT and DST from models of variation in sensitivity to environmental stimuli, such as Strelau's (1983) theory of reactivity or Stelmack and Geen's (1992) model of introversion. The central construct in these personality theories—individual differences in the intensity of response to sensory stimulation—is a necessary but not sufficient condition for differential susceptibility to the environment. By definition, neurobiological susceptibility involves the additional step of moderating the effects of environmental exposures on developmental and life outcomes (with more susceptible individuals experiencing more developmental change in response to life experiences). Although this does not imply that such developmental changes are set in stone, it does not preclude the possibility that these environmentally induced effects are long lasting, if not permanent. Nevertheless, it remains an open question as to whether neurobiological susceptibility is stable over time and thus whether some individuals may become more or less susceptible to environmental influences as they develop, or even whether some individuals may be particularly susceptible at one point in time but not another.

Fifth, a core assertion of both DST and BSCT is that individuals differ in neurobiological susceptibility to environmental contexts that are both positive in character (i.e., afford resources and support that potentially enhance fitness) and negative in character (i.e., embody stressors and adversities that potentially undermine fitness). In other words, and in contrast to widely embraced diathesis–stress models, it is not simply that some individuals are more susceptible to the negative effects of adversity, making them vulnerable. Rather, BSCT and DST regard those disproportionately vulnerable to adversity as also disproportionately likely to benefit from supportive and enriching environments. Such individuals are thus affected by the environment in a manner that can be characterized as “for better and for worse.” Another way of describing or even conceptualizing the individual differences in question is in terms of reaction norms (Manuck, Reference Manuck2009). Whereas some individuals have a wide range of reaction in terms of their developmental outcomes and functioning, depending on the environments they encounter, others have a much narrower range of reaction, responding less markedly, if at all, to positive and negative life experiences.

It is critical to appreciate that differential susceptibility to positive and negative environments has different implications when viewed from developmental-psychopathology and evolutionary perspectives. In the developmental-psychopathology framework, heightened neurobiological susceptibility increases the ability and tendency of individuals to experience “good” outcomes in positive environments (i.e., “for better,” as defined by dominant Western values; e.g., secure attachment, happiness, high self-esteem, emotion regulation, educational and professional success, stable marriage) and “bad” outcomes in negative environments (i.e., “for worse,” as defined by that same value system; e.g., insecure attachment, substance abuse, conduct problems, depression, school failure, teenage pregnancy). By contrast, according to the evolutionary perspectives central to both DST and BSCT, heightened neurobiological susceptibility to the environment functions to direct or regulate development in ways that, over human evolution, recurrently matched individuals to both positive and negative environments, thereby promoting reproductive fitness. In positive environments, this translates into adjusting development to optimize reproductively relevant processes and behaviors such as growth, status, fertility, and offspring quality. This form of conditional adaptation would typically be considered “for better” in a developmental psychopathology framework. In negative environments, however, this translates into “making the best of a bad situation,” often resulting in developmental outcomes that are typically regarded as “nonoptimal” in Western culture. A key difference, then, between the evolutionary and developmental psychopathology perspectives is that evolutionary models conceptualize conditional adaptation to negative environments as an output of evolved developmental systems shaped by natural selection in the service of fitness goals. Consequently, even though susceptible individuals in negative environments may be especially vulnerable to poor mental health outcomes (as defined by dominant Western values), they may still be acting in ways that promote or once promoted status and reproductive success in dangerous environments (e.g., gang membership in bad neighborhoods: see Palmer & Tilley, Reference Palmer and Tilley1995; advantage taking, sexual promiscuity, limited parental investment).

Beyond these points of agreement, it is worth highlighting that BSCT and DST both derive from evolutionary analyses of human development that, even if not identical, are grounded in the view that natural selection maintains alternative patterns of development (phenotypic variation) in the context of multiniche environments (e.g., Hinde & Stevenson-Hinde, Reference Hinde and Stevenson-Hinde1990; Penke, Jaap, Denissen, & Miller, Reference Penke, Jaap, Dennissen and Miller2007). Both theories posit that variation in neurobiological susceptibility to the environment has been adaptively structured and functions as a central mechanism in regulating alternative developmental pathways to match, as well possible, different environmental niches. Fundamental to this view is the assumption that optimal developmental strategies vary as a function of the physical, economic, and social parameters of an individual's specific environment. Levels of neurobiological susceptibility that promote success in some environments may therefore lead to failure in others. It is this kind of environmental heterogeneity (multiniche environments in which a trait's effect on fitness varies across time or space) that provides the ecological basis for the maintenance of adaptive phenotypic variation (whether through balancing selection, conditional adaptation, bet-hedging, or a combination thereof; see Ellis et al., Reference Ellis, Jackson and Boyce2006, Reference Ellis, Figueredo, Brumbach and Schlomer2009; Penke et al., Reference Penke, Jaap, Dennissen and Miller2007).

In sum, DST and BSCT largely converge on an integrated theory of neurobiological susceptibility to the environment. Taken together, these perspectives shed new light on the potency and impotency of a broad range of environmental contexts, from highly positive and enriching to dangerous and corrosive, to shape the gamut of developmental outcomes.

Biobehavioral and Neurogenomic Bases of Differential Susceptibility

A substantial but scattered body of empirical findings has emerged that is consistent with the BSCT and DST proposals. Only recently has attention been explicitly drawn to the common findings of this diverse set of studies (Belsky, Reference Belsky, Ellis and Bjorklund2005; Belsky et al., Reference Belsky, Bakermans-Kranenburg and van IJzendoorn2007; Belsky & Pluess, Reference Belsky and Pluess2009a; Boyce & Ellis, Reference Boyce and Ellis2005; Obradovic & Boyce, 2009), and this Special Section of Development and Psychopathology is the first publication in which papers addressing differential susceptibility will share a common source. As background, we provide a brief summary of thought and conjecture on the neurogenomic and endophenotypic mechanisms that may underlie individual differences in context sensitivity, as well as an illustrative synopsis of evidence gathered prior to the Special Section in support of the differential susceptibility paradigm.

BSCT and DST converge on a common, unifying claim that differences in openness to environmental influence are grounded in and subserved by neurobiological variation in sensitivity to contextual signals and cues. BSCT originated in empirical observations of differences in children's autonomic and adrenocortical reactivity to challenge, whereas DST initially advanced no mechanistic hypotheses about how individual differences in susceptibility operated, but focused instead on the temperamental, phenotypic descriptors of context sensitive children. Theory and data suggest that a common, context-sensitive endophenotype may plausibly underlie differential susceptibility at multiple levels of analysis, ranging from behavioral indicators to peripheral neuroendocrine pathways, brain circuitry, and both genetic and epigenetic variation. Common to each level of analysis is a biobehavioral process involving heightened susceptibility to both risk-promoting and development-enhancing environmental contexts.

In considering such a claim, note that bivalent effects of environmental and phenotypic factors are increasingly known and appreciated within the biological and social sciences. A recent paper, for example, shows that dietary folate conveys protective effects against colorectal cancer, unless an underlying neoplastic process triggers a reversal of such effects, rendering folate a cofactor in the process of tumorigenesis (Mason, Reference Mason2009). In another example, Gluckman, Hanson, Cooper, and Thornburg (Reference Gluckman, Hanson, Cooper and Thornburg2008) contend that a conditional adaptation to fetal undernutrition leads to the development of a “thrifty phenotype” involving insulin resistance, which is protective and survival-enhancing in resource-poor environments, but increases risk for endothelial dysfunction, obesity, and risk for cardiovascular diseases in resource-rich ones. Such a phenomenon is thought to have occurred among individuals who were prenatally exposed to famine during the Dutch Hunger Winter of 1944–1945, who experienced a transition to food abundance in later life, and who thus acquired an enhanced risk for cardiovascular disease and early mortality (Painter et al., Reference Painter, de Rooij, Bossuyt, Simmers, Osmond and Baker2006; Ravelli et al., Reference Ravelli, van der Meulen, Michels, Osmond, Barker and Hales1998; Ravelli, van der Meulen, Osmond, Barker, & Bleker, Reference Ravelli, van der Meulen, Osmond, Barker and Bleker1999). These examples are compelling illustrations of how a single biological agent or process can have bivalent effects that are context contingent.

The observation that bivalent effects of the risk-altering factor in question may drive environmentally contingent, U-shaped distributions of the factor within human populations is also important. Ellis at al. (2005) presented provisional evidence that stress reactivity has a U-shaped distribution along a gradient of favorable to unfavorable social and family settings, Macrì and Würbel (Reference Macrì and Würbel2006) demonstrated a similar U-shaped distribution for adrenocortical reactivity in rodents, and Gluckman et al. (Reference Gluckman, Hanson, Cooper and Thornburg2008) showed that childhood obesity and insulin resistance are more prevalent at both ends of the birth weight spectrum. In the case of stress reactivity, conditional adaptation may bias early development in low- and high-stress contexts toward high sensitivity endophenotypes. Similarly, prenatal undernutrition may predispose fetuses to allocate nutrients disproportionately to adipose tissues in order to augment survival chances in underresourced settings or epochs, whereas fetal overnutrition may also lead to hyperinsulinemia and fat deposition, resulting in more accelerated infant weight gain and higher relative risks of obesity and later cardiovascular disease.

A biologically sensitive endophenotype producing such bivalent effects might be subserved by systematic differences in function or even structure at multiple, hierarchically organized levels of complexity. For illustrative purposes, we summarize recent evidence of such effects at (a) genetic, (b) epigenetic, (c) neural, (d) neuroendocrine, and (e) behavioral levels of analysis. Evidence for genetic moderation of environmental effects in a for better and for worse manner can be found in research carried out across the life span (for reviews, see Bakermans-Kranenburg & van IJzendoorn, Reference Bakermans-Kranenburg and van IJzendoorn2007; Belsky et al., Reference Belsky, Jonassaint, Pluess, Stanton, Brummett and Williams2009; Belsky & Pluess, Reference Belsky and Pluess2009a). One Dutch study showed that maternal sensitivity observed when children were 10 months of age predicted externalizing problems reported by mothers more than 2 years later, but only for children carrying the 7-repeat dopamine receptor D4 (DRD4) allele (Bakermans-Kranenburg & van IJzendoorn, Reference Bakermans-Kranenburg and van IJzendoorn2006); such children displayed the most externalizing behavior observed when mothers were judged insensitive but the least when mothers were judged highly sensitive (for similar results, see Sheese, Voelker, Rothbart, & Posner, Reference Sheese, Voelker, Rothbart and Posner2007). Mills-Koonce and associates (2007) report analogous findings in the case of children carrying the A1+ allele of the DRD2 polymorphism in their work linking sensitive mothering at 6 and 12 months of age with children's affective problems at age 3 years. In adolescence, Eley and colleagues (2004) observed that girls growing up in more and less risky family environments manifested higher and lower levels of depression, respectively, although this proved true principally in the case of those homozygous for short alleles on the serotonin-transporter gene (5-HTTLPR). These results proved strikingly similar to those depression-related findings subsequently reported by Taylor and associates (2006) studying effects of life events in young adulthood. In another study of adults, van IJzendoorn, Bakermans-Kranenburg, and Mesman (2008) found evidence for dopamine-related genes (catechol-o-methyltransferase [COMT] and DRD4) moderating—in G × G × E fashion—the effect of daily hassles on parenting. Adults with DRD4 7-repeat and COMTval proved to be less responsive to their toddlers when confronted with more than average daily hassles. In the case of fewer than average daily hassles, however, they showed the highest levels of responsive parenting. In a large Finnish G × E investigation focused on the serotonin 2A receptor, adults carrying one or more T alleles of the HTR2A T102C polymorphism scored highest on harm avoidance if they grew up in low socioeconomic status households but lowest if they grew up in high socioeconomic status families (Jokela, Lehtimaki, & Keltikangas-Jarvinen, Reference Jokela, Lehtimaki and Keltikangas-Jarvinen2007). Finally, studies of groups of captive and free-ranging rhesus macaques have also shown how variation in the promoter region of the SLC6A4, serotonin transporter gene, which influences the tone and responsivity of serotonergic circuitry, influences the infant monkey's sensitivity to perturbations in the rearing environment (Barr et al., Reference Barr, Newman, Becker, Parker, Champoux and Lesch2003, Reference Barr, Newman, Shannon, Parker, Dvoskin and Becker2004).

Bakermans-Kranenburg and van IJzendoorn (Reference Bakermans-Kranenburg and van IJzendoorn2007) and Belsky and Pluess (Reference Belsky and Pluess2009a) have suggested that endophenotypic variation in environmental susceptibility might be underpinned by allelic variation in the dopaminergic and serotonergic circuitry of the brain that govern thresholds of responsiveness to reward and punishment. The research reviewed above demonstrating that polymorphisms in the DRD4 gene or the serotonin transporter gene are associated with either highly adaptive or suboptimal child and adult developmental endpoints, depending upon rearing experiences and life events, would offer empirical support for such a mechanism. Allelic variation in DRD4 polymorphisms, for example, might plausibly influence susceptibility to context through differences in attention, state regulation, orienting responses, or thresholds for rewards, all behaviors empirically linked to dopaminergic neural circuits (Bakermans-Kranenburg, van IJzendoorn, Mesman, Alink, & Juffer, Reference Bakermans-Kranenburg, van IJzendoorn, Mesman, Alink and Juffer2008); for an extensive discussion of the mesolimbic dopamine system as a potential susceptibility mechanism, see Gatzke-Kopp (2010).

At the epigenetic level of analysis, cutting edge animal work reveals how differences in social or physical environmental exposures can reprogram phenotypic differences in biobehavioral reactivity to adversity. Meaney, Szyf, and colleagues' (e.g., Weaver et al., Reference Weaver, Cervoni, Champagne, D'Alessio, Sharma and Seckl2004) research program demonstrates that natural variation in maternal behavior in the rat induces changes in glucocorticoid receptor gene expression through altered histone acetylation, DNA methylation, and NGFI-A transcription factor binding. Such changes in glucocorticoid receptor expression alter the reactivity of the hypothalamic–pituitary–adrenocortical axis, calibrating the individual's level of hypothalamic–pituitary–adrenocortical axis responsivity to stress for the remainder of the life span. McGowan, Meaney, and Szyf (Reference McGowan, Meaney and Szyf2008) demonstrated in rodents how such effects may be reversed with drastic environmental changes (e.g., in diet, such as l-methionine supplementation) that influence epigenetic changes in critical brain loci.

In a series of studies on humans, Philibert and colleagues showed that methylation levels of the CpG island upstream from SCL6A4 were associated with abuse during childhood (Beach, Brody, Todorov, Gunter, & Philibert, Reference Beach, Brody, Todorov, Gunter and Philibert2010) and that product levels of the serotonergic system differed according to degree of methylation (Philibert et al., Reference Philibert, Madan, Anderson, Cadoret, Packer and Sandhu2007). In addition, van IJzendoorn and colleagues found that higher levels of methylation of 5-HTT were associated with increased risk of unresolved responses to trauma in carriers of the usually protective 5-HTTLPR ll variant, thus affecting setpoints for reactivity to traumatic stress (van IJzendoorn, Caspers, Bakermans-Kranenburg, Beach, & Philibert, Reference Van IJzendoorn, Caspers, Bakermans-Kranenburg, Beach and Philibert2010). They argued that methylation may serve as the interface between the neurobiological basis of human development and the environment.

Further genetic and epigenetic processes in humans, which involve the developmental shaping of neurotransmitter and molecular signaling pathways, appear to regulate brain structures that mediate reward, fear, and emotional reactivity (Feder, Nestler, & Charney, Reference Feder, Nestler and Charney2009). Thus, at the level of neural function, differential susceptibility to environmental exposures may be determined by systematic differences in the functioning of specific brain circuitry, neuronal activity, and neurotransmitter production, processing, and metabolism. A variety of brain regions have been implicated in the filtering of incoming sensory information, including the temporal cortex (Boutros et al., Reference Boutros, Torello, Barker, Tueting, Wu and Nasrallah1995), prefrontal cortex (Shimamura, Reference Shimamura2000), amygdala (Hariri et al., Reference Hariri, Drabant, Munoz, Kolachana, Mattay and Egan2005), and thalamus (McCormick & Bal, Reference McCormick and Bal1994), and such filtering might arguably act as a neural substrate of context sensitivity. Certain clinical conditions, such as chronic pain (Miller, Reference Miller2000) and autism (Kern et al., Reference Kern, Trivedi, Grannemann, Garver, Johnson and Andrews2007; Leekam, Nieto, Libby, Wing, & Gould, Reference Leekam, Nieto, Libby, Wing and Gould2007), also involve systematic differences in sensory sensitivity in the auditory, visual, tactile, or olfactory modalities. Jagiellowicz et al. (in press) observed that individuals with exceptional sensitivity to internal and external stimuli, including social and emotional cues, showed greater functional magnetic resonance imaging activation of brain regions involved in higher order visual processing and the cerebellum when detecting minor changes in visual stimuli. A variety of molecular signaling pathways utilizing peptides and neurotransmitters (e.g., norepinephrine [NE], serotonin, dopamine, neuropeptide Y, and brain-derived neurotrophic factor) are involved in the activation and regulation of these circuits, along with polymorphisms and haplotypes within the genes involved in the expression of signaling molecules and their receptors (Feder et al., Reference Feder, Nestler and Charney2009).

As elucidated in the research of several investigators, including Boyce (Alkon et al., Reference Alkon, Lippert, Vujan, Rodriquez, Boyce and Eskenazi2006; Boyce et al., Reference Boyce, Chesney, Alkon, Tschann, Adams and Chesterman1995, Reference Boyce, Quas, Alkon, Smider, Essex and Kupfer2001; Boyce & Ellis, Reference Boyce and Ellis2005; Obradovic et al., 2010); McEwen (Reference McEwen2007); Kagan, Reznick, and Snidman (Reference Kagan, Reznick and Snidman1988); Gunnar (Gunnar & Quevedo, Reference Gunnar and Quevedo2007; Lupien, McEwen, Gunnar, & Heim, Reference Lupien, McEwen, Gunnar and Heim2009); and their colleagues, context sensitivity also appears to be embodied in the differential reactivity of the two peripheral neuroendocrine stress response systems: corticotropin-releasing hormone and locus coeruleus–NE (LC-NE) systems. There is extensive individual variability in the reactivity of these systems to standardized laboratory challenges, and individuals with heightened responsivity in either or both systems appear to sustain the worst or the best of the observed health and developmental outcomes, contingent upon the level of adversity or support prevalent in the immediate social environment. As discussed earlier, Boyce and associates (1995) reported that 3- to 5-year-old children with higher blood pressure reactivity or higher immune reactivity, which are both strongly influenced by activation of the corticotropin-releasing hormone and LC-NE systems, exhibited higher rates of respiratory illness than other children when growing up in stressful rearing contexts, yet under low-stress conditions such highly reactive children had a significantly lower incidence of respiratory illnesses than other children. More recently, Obradovic and associates (2010) found that 5- to 6-year-old children with high reactivity of the parasympathetic nervous system (modulated by LC-NE system responses) were rated as less prosocial when growing up under conditions of high contextual adversity and more prosocial under more favorable contextual conditions, compared to children with low parasympathetic reactivity. In addition, children with high cortisol reactivity proved more prosocial under conditions of low adversity and less prosocial under high adversity relative to children with low cortisol reactivity. In a study of adults, Gannon, Banks, Shelton, and Luchetta (Reference Gannon, Banks, Shelton and Luchetta1989) observed that, in comparison to undergraduates showing low autonomic reactivity, more highly reactive students experienced lower levels of physical symptoms and depression when experiencing few daily hassles but higher levels when experiencing many hassles.

Most behavioral phenotypic markers of susceptibility that have been identified involve negative emotionality in one form or another (Belsky, Reference Belsky, Ellis and Bjorklund2005; Belsky & Pluess, Reference Belsky and Pluess2009a; Kochanska, Aksan, & Joy, Reference Kochanska, Aksan and Joy2007; Van Zeijl et al., Reference Van Zeijl, Mesman, Stolk, Alink, van IJzendoorn and Bakermans-Kranenburg2007), likely because of the predominance of vulnerability oriented research based on a diathesis–stress model and thus focused on this putative “risk factor.” In a study of temperament and maternal discipline in relation to externalizing problems in early childhood, Van Zeijl et al. (Reference Van Zeijl, Mesman, Stolk, Alink, van IJzendoorn and Bakermans-Kranenburg2007) found that children with difficult temperaments were more susceptible to both negative and positive discipline, compared with children of relatively easy temperament. Using data from the large-scale National Institute of Child Health and Human Development Study of Early Child Care and Youth Development, Pluess and Belsky (Reference Pluess and Belsky2009) reported that, compared to children with easy temperaments, difficult children had more behavior problems early in their school careers when exposed to low-quality child care during infancy or early childhood but fewer problems when quality was high. Other inquiries drawing on National Institute of Child Health and Human Development data, but focusing upon effects of parenting, as well as additional developmental outcomes and children's functioning at older ages, have generated similar results (Bradley & Corwyn, Reference Bradley and Corwyn2008; Dopkins-Stright, Cranley-Gallagher & Kelley, 2008; Pluess & Belsky, Reference Pluess and Belsky2010). Further, Lengua's (Reference Lengua2008) investigation of a Temperament × Parenting interaction during the middle-childhood years showed that children who were highly prone to negative emotion in the form of frustration increased in externalizing problems over time when mothers were rejecting, but decreased when mothers manifested little rejection; no such parenting effects were evident for same-age peers scoring low on frustration. In young adulthood, Aron et al. (Reference Aron, Aron and Davies2005, Studies 2 and 3) observed that a problematic childrearing history predicted high levels of self-reported shyness and negative affectivity among undergraduate students, whereas its absence predicted low levels of these same dependent constructs; this relation obtained principally in the case of students scoring high on SPS (see also Aron & Aron, Reference Aron and Aron1997).

A difficult ontogenetic issue is the question of where, within this spectrum of genomic, epigenomic, neural, neuroendocrine, and behavioral mechanisms, differential susceptibility actually resides in humans and other species. Is neuroendocrine reactivity the actual mechanism by which environmental susceptibility generates greater or lesser risks to and benefits for health and development? Or is reactivity simply a physiological marker of heightened susceptibility? Relative to variations in neural circuitry activation, are genetic and epigenetic variations closer to or more distant from the true mediating events of sensitivity to context? These are philosophical or even rhetorical questions to some degree, because developmental sensitivity to context may well be instantiated at all known levels of biological abstraction. Each level is, after all, hierarchically and mechanistically related to that just above and that just below on the scales of size and complexity. In contrast, serious consideration of the causal processes that generate differential susceptibility cannot ignore the difficulty of pursuing and defining the causal events more and less proximal to observed differences in context sensitivity. The next generation of research on differential susceptibility to the environment will need to observationally and experimentally contend with the issue of its causal origins.

Comparison and Evaluation of Evolutionary Models and Their Implications

Although understanding neurobiological susceptibility to the environment will require detailed mechanistic knowledge, we know from biology that the study of proximate mechanisms is insufficient by itself and must be complemented by the study of ultimate causation in terms of what the mechanistic machinery is designed to do. Despite major commonalities, BSCT and DST posit different evolutionary models of neurobiological susceptibility that generate different predictions about the development and distribution of susceptible phenotypes across ecological and social contexts. Articulating the bases of these different predictions, including hypothesized fitness consequences in varying physical and social environments, should enable scholars to design studies to test the models and potentially distinguish between them.

Basic and Applied Issues in Differential Susceptibility Research

The preceding discussions of proximate mechanisms and adaptive function, to say nothing of evidence assembled to date highlighting differential susceptibility (for reviews, see Bakermans-Kranenburg & van IJzendoorn, 2010; Belsky & Pluess, Reference Belsky and Pluess2009a; Obradovic & Boyce, 2009), raise a host of issues and possibilities for both basic and applied research in the future. These include, as will be discussed below, methodological considerations in designing differential susceptibility research; issues of ecological, cultural and racial–ethnic variation; the need for experimental validation; statistical concerns in evaluating differential susceptibility findings; and ethical dilemmas involved in using differential susceptibility to shape intervention strategies and social programs.

Agenda for Future Research

Differential susceptibility is a burgeoning new field of study with an exciting agenda for future research. Topping this agenda is uncovering a high-resolution map of the mechanisms of neurobiological susceptibility and testing their evolutionary functions. A careful analysis of mechanisms is needed to determine whether different research teams working at different levels of analysis (genotypic, endophenotypic, and behavioral) have been identifying the same more or less susceptible individuals using different but related susceptibility markers. Ultimately, a detailed understanding of how neurobiological mechanisms regulate differential susceptibility should provide a solid foundation for shaping programs and interventions to maximally benefit children (and adults) of all kinds. The path forward is integration across levels: elucidating the hierarchical and bidirectional relations between genotypes, endophenotypes, and behavior that together constitute neurobiological susceptibility to the environment and explain its variation across individuals. At the same time, an evolutionary perspective is needed to understand what all of the neurobiology has been selected to do. As ongoing research firmly establishes the evolved functions of variation in neurobiological susceptibility (i.e., why differential susceptibility has been maintained by natural selection), strong hypotheses regarding the development and distribution of susceptible phenotypes across time, space, and social context will follow. Finally, progress toward understanding proximate mechanisms and evolutionary function will depend on careful consideration of the highly interdependent and context-dependent components of differential susceptibility—organismic characteristics, environmental factors, and developmental outcomes—and their ecological, cultural, and racial–ethnic dimensions.

In addition to these metatheoretical issues, a number of more specific issues await future research. A first question is how domain specific versus domain general is neurobiological susceptibility to the environment (see Belsky, Reference Belsky, Ellis and Bjorklund2005). Are some individuals more susceptible than others to a wide and diverse array of developmental inputs vis-à-vis a wide and diverse array of developmental outcomes? Are some individuals highly susceptible to parental influences on academic achievement, whereas others prove highly susceptible to peer effects on risk-taking behavior? It is easy to imagine that there could be individual differences in the degree to which susceptibility is broad versus narrow, such that some are highly susceptible to many environmental effects on multiple outcomes, others only to some effects on some outcomes, and still others not susceptible to many or even any. Domain specificity concurs with the widely embraced notion that different neurobiological systems are sensitive to different environmental inputs and influence different developmental outputs. Essex et al. (Reference Essex, Armstrong, Burk, Goldsmith and Boyce2011 [this issue]) and Obradovic et al. (2011 [this issue]) each present evidence supporting domain specificity.

Although we have sometimes used typological language to describe more susceptible children (e.g., designating them as orchid children), it would be mistaken to infer that susceptibility is more discrete than continuous, that is, that some individuals are highly sensitive to environmental contexts while others simply are not. Perhaps the better way to conceptualize the issue (highlighted in footnote 2) is in terms of individual differences in the neurobiological susceptibility (continuous variation), regardless of whether such susceptibility to the environment is domain general or domain specific. Thus, it may make more sense to conceptualize differential susceptibility as a continuous dimension rather than in categorical terms. Evidence from Caspi et al.'s (2003) groundbreaking G × E study is consistent with this argument. It showed that stressful life events exerted it greatest impact on those homozygous for the short serotonin-transporter allele and least for those homozygous for the long allele, with those heterozygotes carrying one of each allele falling in between. Other evidence such as this led Belsky and Pluess (Reference Belsky and Pluess2009a, Reference Belsky and Pluess2009b) to propose a “plasticity gradient” along which individuals vary.

How developmental time periods relate to differential susceptibility, including whether susceptibility varies within and/or across individuals over time, also merits attention in future work. Thus, research needs to address whether those who are most and least malleable in response to the environment (or particular features of the environment) early in life remain so later in life, no matter how early and later are operationalized. One can imagine, for example, a highly susceptible child growing up in what might be regarded as a neutral environment that is neither particularly supportive nor especially stressful. What happens when this individual finds himself in an extremely supportive or stressful context in middle childhood or adulthood? Is this individual still highly susceptible to the anticipated positive and negative effects of such contexts? Has his heightened context sensitivity been reduced by time and development? What happens when a person who has grown up in a very supportive (or stressful) early environment finds herself later in life in a dramatically contrasting context? Does she remain as susceptible to its benefits (or costs) as she might have been earlier in life?

Regardless of whether stability or change characterizes differential susceptibility, research will also need to determine why that is the case. That is, through what neurobiological processes do stability or change in environmental susceptibility operate? For example, do those individuals who are highly physiologically reactive and thus most environmentally sensitive according to BSCT remain highly physiologically reactive as they develop? Whether or not this is so, does future susceptibility to the environment remain mediated by heightened physiological reactivity, or do other processes take over this function? An important related issue is whether differential susceptibility functions similarly in children and adults. It may be that differential susceptibility in children primarily moderates the sensitivity of developmental trajectories to environmental conditions, whereas differential susceptibility in adults primarily moderates the ability to adapt to new niches.

Another important issue for future research will be determining how for better and for worse processes unfold. Neurogenomic and endophenotypic mechanisms of differential susceptibility must be manifest through behavioral interactions with the environment. For example, do neurobiologically susceptible children in supportive environments experience “better” outcomes because their enhanced sensitivity makes them better at detecting positive opportunities and learning to capitalize on them (e.g., seeing a teacher as a prospective mentor, taking advice from a parent, displaying heightened awareness of and responsivity to external stimuli)? Such processes would be consistent with the profile of orchid children as more reflective and perhaps more conscious of self and the environment (Ellis & Boyce, Reference Ellis and Boyce2008). Conversely, do neurobiologically susceptible children in adverse contexts experience “worse” outcomes because their enhanced reactivity promotes a hostile attribution bias, facilitates development of an exploitive interpersonal style (e.g., increased abilities for deception; Mealey, Reference Mealey1995), or lowers thresholds for detecting threat in ambiguous or unfamiliar situations (e.g., elevated sensitivity to threat cues such as angry faces; Pollak, Reference Pollak2008), prompting them to respond more quickly and aggressively to perceived threat? A full understanding of these for better and for worse outcomes awaits mapping of the downstream behavioral processes through which neurogenomic and endophenotypic mechanisms of differential susceptibility are integrated and expressed.

The last issue to be considered here concerns the ontogenetic origins of differential susceptibility and, more specifically, identifying specific genetic and environmental influences on heightened susceptibility to environmental influence. Boyce and Ellis' (2005) BSCT advanced the claim that physiological reactivity arises in response to both very positive and very negative environments (i.e., the U-shaped curve). Belsky's (1997a, 1997b; Belsky et al., Reference Belsky, Jonassaint, Pluess, Stanton, Brummett and Williams2009) original DST was based on the presumption that differences in susceptibility were more or less genotypic in origin. But neither perspective precluded the alternative. Pluess and Belsky (Reference Pluess and Belsky2011 [this issue]) advance the hypothesis that, as a result of maternal stress, there may be prenatal programming of postnatal plasticity and that individuals may be differentially susceptible to prenatal programming for genetic reasons. BSCT also embraces such G × E possibilities.