Beyond Risk and Protective Factors: An Adaptation-Based Approach to Resilience

The specialization and sensitization hypotheses

Drawing on an evolutionary-developmental framework, we propose the specialization hypothesis: Harsh, unpredictable environments do not exclusively impair cognitive abilities; instead, individuals become developmentally adapted (“specialized” and potentially enhanced) for solving problems that are ecologically relevant in such environments (Frankenhuis & de Weerth, 2013). For example, in rapidly changing environments, heightened attention-shifting may enable individuals to take advantage of fleeting opportunities, even if frequent shifting interferes with sustained attention (see Mittal et al., 2015). To improve intervention outcomes in stress-adapted children and youth, we need to uncover a high-resolution map of specific cognitive abilities that are enhanced as a result of growing up under high-adversity conditions. That would enable design of interventions that work with, instead of against, these abilities.

A corollary of the specialization hypothesis is the sensitization hypothesis: The hypothesized advantages in cognitive function among people who grow up under stressful conditions (as per the specialization hypothesis) are manifested primarily under currently stressful conditions (i.e., earlier-life experiences sensitize later responses to stress). For example, stress-adapted youth may be advantaged at attention shifting under conditions of current stress and uncertainty, but not in benign, nonthreatening circumstances (Mittal et al., 2015).

The sensitization hypothesis assumes that the day-to-day experiences and circumstances of stress-adapted individuals are qualitatively different from those of individuals from low-risk backgrounds and, therefore, that testing stress-adapted children and youth under standard laboratory conditions may disadvantage them by not allowing them to show their abilities in context (i.e., their abilities to solve problems and achieve goals within their local ecology). Stress-adapted children and youth may instead perform certain tasks better in settings that do not attempt to minimize the reality of daily stressors and uncertainties. This could include contexts that expressly highlight the prevalence of daily stressors (e.g., reminders that we live in a world where resources are uncertain) or environments in which people in a room are allowed to move and talk, which may simulate the contexts in which stress-adapted individuals developed their skills. In total, the sensitization hypothesis necessitates studying test performance-by-environment interactions. It involves testing for skills and abilities under different conditions (i.e., experimental manipulations of psychological or environmental states). ¹ Gaining knowledge about environmental conditions that are promotive versus harmful to the performance of stress-adapted children and youth is critical to designing interventions that work with this population.

Focusing a theoretical and empirical lens on how early exposures to harsh, unpredictable environments adaptively influence cognitive abilities is critical for understanding how individuals developing in such contexts learn, remember, solve problems, and make decisions. An extraordinarily large body of research has documented the maladaptive consequences of early life stress. This raises the question: Why should knowledge about the cognitive strengths of children and youth who are adapted to harsh, unpredictable environments (of which we know so little) be any less useful than knowledge about their impairments (of which we know so much)? The better we understand cognitive adaptations to harsh, unpredictable environments, including specialization and sensitization effects, the more effectively we can tailor education, policy, and interventions to fit the needs and potentials of stress-adapted children and youth. This adaptation-based approach to resilience exemplifies using psychology to improve people’s lives because it illuminates the unique strengths and abilities that develop in response to high-stress environments—and how to use those attributes to enhance learning and developmental outcomes in stress-adapted individuals.

Overview

We begin by reviewing the well-established negative effects of psychosocial adversity on cognitive development, and then summarize how this issue has been conceptualized and addressed in the resilience literature. We then present the theoretical background for an alternative adaptation-based approach to resilience that is based in life history theory. To explicate this new perspective, we discuss and evaluate the specialization and sensitization hypotheses through a review of relevant theoretical and empirical literatures, focusing on cognitive adaptations to chronic or repeated childhood adversity in birds, rodents, and humans. We conclude by discussing the implications of the adaptation-based approach to resilience for intervention and propose future directions aimed at increasing research efforts and knowledge in this area of multidisciplinary inquiry.

The Effects of Adversity on Cognitive and Academic Outcomes

The physical, material, and social hardships of poverty encompass a wide range of contexts that may negatively affect child cognitive development and achievement. These contexts include neighborhood danger; exposure to environmental chemicals; bad housing conditions characterized by noise, crowding, and violence; neglectful and abusive parenting; parental mental and physical health problems; family instability resulting in disrupted relationships with caregivers; residential instability; low-quality child care; and peer and school violence (e.g., Blair & Raver, 2012a; Bradley & Corwyn, 2002; Duncan & Brooks-Gunn, 2000). Herein we refer to children and youth experiencing the diverse hardships of poverty as growing up in “high-risk” or “harsh, unpredictable” environments (and thus being stress-adapted). ²

One of the most robust findings in the field of human development is the pervasive negative effect of poverty on cognitive, learning, and achievement outcomes (e.g., Bradley & Corwyn, 2002; Conger & Donnellan, 2007; Duncan, Magnuson, Kalil, & Ziol-Guest, 2012; McLoyd, 1998). Lower levels of language skill and vocabulary are apparent as early as 18 months in poor children, and lower levels of reading and math skills are evident in kindergarten, with these discrepancies tending to get worse over time (Fernald, Marchman, & Weisleder, 2013; Mulligan, Hastedt, & McCarroll, 2012). Low family income is a strong and consistent predictor of reduced executive-function abilities (Blair et al., 2011; Evans & Schamberg, 2009), learning disabilities involving reading, writing and mathematics, and overall lower scores on standard intelligence tests and scholastic tests (Heberle & Carter, 2015). Children from impoverished backgrounds are also at an elevated risk for grade repetition, expulsion and suspension from school, and school dropout (Ross et al., 2012).

Extant theory and research has focused on understanding how the diverse hardships of poverty lead to such poor developmental outcomes. A common approach has been to count risk factors in a child’s life. Consistent with cumulative risk models of development (e.g., Evans et al., 2013; Sameroff et al., 1987; Seifer et al., 1996), this approach involves compositing multiple sources of stress in family environments or examining the additive effects of multiple stressors. The underlying hypothesis in this approach is that the more stressors children are exposed to, the more their developmental competencies will be compromised.

A more fine-grained version of the cumulative risk approach involves testing mediation models, which focus on intervening mechanisms in the relations between childhood stress and subsequent neurobiological and cognitive outcomes. Much of this work focuses on the mediating roles of parental condition and functioning (e.g., depression, low marital quality) and the resulting quality of caregiving (e.g., harsh discipline, low parental sensitivity) in explaining the effects of poverty on child development (e.g., Belsky, Steinberg, & Draper, 1991; Conger & Conger, 2002). Some models seek to identify intervening biological mechanisms, such as patterns of child stress physiology (e.g., Blair & Raver, 2012b; Del Giudice, Ellis, & Shirtcliff, 2011) or the structure and function of key brain areas such as the prefrontal cortex, amygdala, and hippocampus (e.g., Karatoreos & McEwen, 2013). Cumulative risk models have commonly been framed in terms of diathesis stress, where exposures to childhood adversities interact with personal vulnerabilities (e.g., difficult temperament, heightened biological reactivity to stress, low-activity MAOA allele) to predict child developmental outcomes. In the diathesis stress framework, certain children or youth are vulnerable or resilient because of personal characteristics that moderate environmental risk.

However the theoretical pie is sliced, these approaches attempt to account for deficits in learning and behavior. Children from high-adversity backgrounds are considered to be at risk for impaired development, and these models attempt to explain the causes, mediators, and moderators of that impairment.

The Traditional Strength-Based Approach to Resilience

Despite the overall low achievement levels of children and youth from socioeconomically disadvantaged backgrounds, there is striking variation in the outcomes of individuals exposed to high adversity. Some of these individuals thrive, or at least “beat the odds,” despite their high-risk background. This observation has led to a resilience literature that has pushed back against cumulative risk approaches and instead has focused on developmental assets, emphasizing the importance of positive resources and promotive factors that enable children and youth to overcome their challenging life circumstances.

Much is now known about the factors associated with resilient outcomes in young people. As reviewed by A. S. Masten (2001, 2014), major predictors of resilient outcomes in youth from high-risk environments include such individual factors as intelligence and problem-solving skills, hope and optimism, self-control, planfulness, and motivation to succeed; such relationship factors as effective caregiving and parenting quality, close relationships with capable adults, and close friends and romantic partners; and larger social system factors that provide resources and protection such as effective schools and communities. These established resilience factors have provided the foundation for interventions that focus on promoting resilience. In total, the resilience literature asks “What does it takes to succeed?” and “How can we build these strengths and qualities in children and youth?”

Fighting an uphill battle that is difficult to win

Whether the goal is to increase reserve capacity, repair and reverse stress-adapted systems, get the cat to retract its claws, or build a better cognitive toolbox, extant intervention strategies share the common goal of trying to get children and youth from high-risk backgrounds to act, think, and feel more like children and youth from low-risk backgrounds. Such interventions involve helping stress-adapted children and youth to compensate for their weaknesses or otherwise overcome the negative effects of growing up under harsh, unpredictable conditions. Although each of the kinds of interventions described by the different resilience metaphors has achieved some empirical success (e.g., Ager, 2013; Durlak et al., 2011; A. S. Masten & Cicchetti, 2016), these approaches are ultimately limited because they do not attempt to leverage—and thus cannot capitalize on—the unique strengths and abilities that develop in response to harsh, unpredictable environments. Broadly speaking, intervention efforts may be stuck in a pattern of fighting against (rather than working with) functional adaptations to stress (Ellis & Del Giudice, 2014; Ellis et al., 2012).

As a case in point, consider a well-established (and much heralded) finding from the social development literature: The social and cognitive skills that children exhibit in kindergarten, such as prosocial skills and self-control, predict their health, education, and employment outcomes in young adulthood (Duckworth, Quinn, & Tsukayama, 2012; Jones, Greenberg, & Crowley, 2015; Moffitt et al., 2011). While impressive, these results do not mean that children displaying low levels of these skills are impaired (as per the deficit model) or that we should necessarily intervene to improve these skills (as per standard resilience models; Frankenhuis, Panchanathan, & Nettle, 2016). Two caveats apply that inform the current adaptation-based approach to resilience.

First, among low-socioeconomic status (SES) children and youth, cognitive skills such as self-control may act as a “double-edged sword,” facilitating academic success and psychosocial adjustment while undermining cardiometabolic health (as reflected in obesity, blood pressure, and stress hormones) and inducing faster epigenetic aging (Brody et al., 2013; Chen, Miller, Brody, & Lei, 2014; Miller, Yu, Chen, & Brody, 2015). These results challenge the notion of universally “good” or “bad” skills and instead suggest that different skills are likely to be adaptive in different contexts (see especially the discussion of “successful intelligence” later in this article). For example, although high levels of cognitive control aid performance on goal-based tasks that rely on a narrow focus of attention, low levels of cognitive control (more typical of stress-adapted children and youth) may enhance performance on open-ended tasks that depend on acquiring and using environmental information from diverse sources (Amer, Campbell, & Hasher, 2016).

Second, although children who are fortunate enough to have strong social and cognitive skills in kindergarten have many positive outcomes, many programs that attempt to build such skills in stress-adapted children have had limited success. For example, through an intensive program that targeted stress-adapted children over their first 10 years in grade school, the Fast Track intervention set out to build self-control skills, anger coping strategies, and interpersonal problem-solving skills (Conduct Problems Prevention Research Group, 1992). Although the program cost about $60,000 per child, it had little impact on social-cognitive processes (see Table 1 of Dodge, Godwin, & Conduct Problems Prevention Research Group, 2013) and, if anything, revealed how difficult it is to get stress-adapted children to think and act more like children from more low-risk backgrounds.

OPEN IN VIEWER

Table 1.

Developmental Stress: Evidence for Adaptive Phenotypic Effects in Birds

Hypothesis type	Developmental Stress Exposure	Phenotypic Effects	Proposed Function
Specialization	Embryonic exposure to yolk corticosterone	Production of smaller nestlings that achieve faster take-off speeds (barn swallows, European starlings, great tits: Crino & Breuner, 2015). Enhanced flight as shown by more mature flight muscles, lower wing loading, and better in-flight performance by fledglings (European starlings: Chin et al., 2009).	Enhanced flight capacity promotes survival in predator dense environments (as signaled by maternal and embryonic glucocorticoids).
Specialization	Low food accessibility, early-life social isolation (in nestlings/fledglings)	In adulthood, enhanced associative learning (zebra finches: Brust, Krüger, Naguib, & Krause, 2014; Kriengwatana, Farrell, Aitken, Garcia, & MacDougall-Shackleton, 2015; chickens: Goerlich, Nätt, Elfwing, Macdonald, & Jensen, 2012) but impaired hippocampal-dependent contextual learning on spatial tasks (Western scrub-jays: Pravosudov, Lavenex, & Omanska, 2005; zebra finches: Kriengwatana et al., 2015).	Enhanced associative learning (over slower, cognitively expensive contextual learning) promotes immediate responding to environmental challenges (prevalent in resource-poor environments).
Specialization	Unpredictable stress regimen (e.g., catching, handling, noise) as fledglings	Within 6 days following stress exposure: enhanced prefrontal cortex-dependent reversal learning (i.e., flexibility relearning a new rewarded location) in a spatial task (Japanese quail: Calandreau et al., 2011).	Enhanced reversal learning promotes survival in an unpredictable environment.
Specialization a	Limited and unpredictable food supply a	Enhanced food caching intensity and memory for cache location (i.e., spatial memory; black-capped chickadees, tufted titmice, marsh tits: Hurly, 1992; Pravosudov & Clayton, 2001, 2002; Pravosudov & Grubb, 1997). b	Enhanced food caching and memory for stored food locations promotes survival and reproduction in environments where food is scarce or unpredictable.
Specialization	Posthatching exposure to corticosterone	In adulthood, more frequent switching between social learning strategies (i.e., discounting of parental information in favor of learning from nonparental adults); weaker affiliation with parents in favor of flock mates; attainment of more central social network positions; improved performance on a novel foraging task; higher quality and quantity of offspring in males (zebra finches: Boogert, Farine, & Spencer, 2014; Crino, Driscoll, Ton, & Breuner, 2014; Crino, Prather, Driscoll, Good, & Breuner, 2014; Farine, Spencer, & Boogert, 2015).	In unpredictable environments, exposure to elevated developmental stress (as signaled by glucocorticoids) may indicate one’s parents are struggling to cope with current conditions. If other adults know better, copying them should yield better foraging strategies. Chicks may also be selected to switch to individual trial-and-error learning in this context. Elevated corticosterone triggers life history trade-offs (e.g., shorter life span) favoring current over future reproduction.
Specialization	Prehatching exposure to corticosteroids; postnatal food unpredictability (nestlings and fledglings)	In adulthood: Prehatching stress led to increased activity level, exploration in a novel environment, and mimicking of foraging strategies demonstrated by conspecifics; posthatching stress led to more risk taking to find food and avoidance of foraging strategies demonstrated by conspecifics. Combined pre- and posthatching stress resulted in the highest levels of exploratory and risk-taking behavior (Japanese quail: Boogert, Zimmer, & Spencer, 2013; Zimmer, Boogert, & Spencer, 2013).	Elevated exploratory and risk-taking behavior under developmentally stressful conditions promotes greater food acquisition. Food unpredictability promotes novel foraging strategies relative to those demonstrated by conspecifics.
Sensitization	Low-quality diet as nestlings	During a brief period of food restriction, adult zebra finches that experienced poor early nutrition were faster to engage in exploratory and foraging behavior (Krause, Honarmand, Wetzel, & Naguib, 2009).	Rapid exploration and foraging promotes greater food acquisition in a nutritionally poor environment.

a

Timing of developmental stress exposure not specified. ^bAge at testing not specified.

The assumption underlying Fast Track, and other interventions like it, is that children growing up under conditions of poverty and violence are damaged by their experiences, and that we (scientists, policy makers, educators) can repair that damage through interventions that train stress-adapted children to be more like children from safe, stable environments. Although exposures to high-stress environments certainly jeopardize health and survival (e.g., Mulvihill, 2005; Shonkoff et al., 2009), and traditional interventions approaches are part of the solution to that problem, the challenge is that extant interventions work against, rather than with, social and cognitive adaptations to high-stress environments (Ellis & Del Giudice, 2014; Ellis et al., 2012); thus, they are fighting an uphill battle that is difficult to win. Furthermore, interventions have potential costs (including iatrogenic effects) as well as benefits, which vary across developmental contexts. As stated by Ellis et al. (2012),

From a conditional adaptation perspective, the first question to ask is whether intervention is appropriate. Seemingly harmful risk-taking behaviors may be adaptive in the context of competitive or dangerous environments; therefore, preventing or changing these behaviors could be equivalent to declawing the cat—removing the psychological and behavioral weaponry necessary to survive and control resources in one’s local ecology. (p. 610)

This logic may shed light on the bivalent effects of self-control discussed above. Although self-control may be adaptive for children living in safe, stable environments, high levels of self-control may create a mismatch for children living in harsh, unpredictable environments where opportunities are fleeting and it is important to obtain more immediate rewards (see Frankenhuis et al., 2016; Mittal et al., 2015).

At an applied level, a bigger problem with the deficit approach—including its value judgments regarding putatively desirable versus undesirable capacities and behaviors—is that it can be belittling and disrespectful to the members of marginalized and low-income communities who we are trying to engage through policy and interventions. As one community stakeholder noted, “there is a tendency to look at people from underserved communities as somehow inferior” (Acosta et al., 2016, p. 40). In contrast, the adaptation-based approach to resilience recognizes, utilizes, and values the skills and abilities that develop in response to high-risk environments. It emphasizes being appreciated and respected for the skills you do have—and using these skills as building blocks for success—rather than being stigmatized for what you lack relative to others.

The Adaptation-Based Approach to Resilience

In contrast to traditional strength-based approaches to resilience, the current adaptation-based approach focuses on generating—and putting to use—a high-resolution map of specific cognitive abilities that are enhanced in children and youth growing up under harsh, unpredictable conditions. The scientific goal is to chart the enhanced social-cognitive skills of individuals who grow up in high-stress environments (see literature reviews below, next two sections), and the applied goal is to leverage these abilities to enhance intervention outcomes in stress-adapted individuals (as discussed in detail in the discussion section). Rather than declawing the cat, for example, the adaptation-based approach to resilience considers ways to take advantage of the cat’s claws to navigate life’s challenges. This adaptation-based approach complements, rather than competes with, traditional strength-based approaches to resilience.

Central to the adaptation-based approach is the concept of trade-offs in development. All organisms live in a world of limited resources; for example, the energy that can be extracted from the environment in a given amount of time is intrinsically limited. Time itself is a limited good (e.g., the time spent by an organism looking for food cannot be used to care for offspring). Such constraints dictate that different life domains—bodily maintenance, physical growth, brain development, reproduction—cannot all be maximized at once. Instead, organisms are selected to make trade-offs that prioritize resource expenditures, so that greater investment of time or resources in one domain occurs at the expense of investment in competing domains. For example, resources spent on an inflammatory host response to fight infection cannot be spent on reproductive effort; thus, the benefits of an inflammatory host response are may trade off against the costs of lower ovarian function in women and reduced musculoskeletal function in men (Clancy et al., 2013; Muehlenbein & Bribiescas, 2005).

According to life history theory (Charnov, 1993; Roff, 1992; Stearns, 1992), these kinds of trade-offs over development are not random; they have been shaped by natural selection to maximize fitness—survival and reproduction—within the specific environment that an organism develops and in relation to its somatic condition. That means, for example, that organisms growing up in food-rich versus food-poor environments, or in safe versus dangerous environments, or in good versus bad health, face systematically different resource-allocation trade-offs and constraints. According to life history theory, each trade-off constitutes a decision node in allocation of resources, and each decision node influences the next (opening up some options, foreclosing others) in an unending chain over the life course (Ellis, Figueredo, Brumbach, & Schlomer, 2009). These trade-offs progressively favor one developmental trajectory over another, resulting in coherent, integrated suites of physiological and behavioral traits that form the individual’s life history strategy.

Human life history strategies appear to vary along a dimension of fast versus slow, reflecting the different trade-offs that individuals face in different environmental contexts (Del Giudice, Gangestad, & Kaplan, 2015; Ellis et al., 2009; Figueredo, de Baca, & Woodley, 2013; Figueredo et al., 2006). Although there is ongoing debate concerning the best way to characterize human life history variation (e.g., Copping, Campbell, & Muncer, 2014; Figueredo et al., 2015), a large body of research suggests that fast life histories are more risky and present oriented (e.g., taking benefits opportunistically with little regard for long-term consequences), prioritize mating effort (e.g., competitive risk-taking, aggression), include earlier sexual development and reproduction, and involve lower levels of parental investment per offspring. By contrast, slower life histories are less risky and more long-term oriented (e.g., greater self-regulation, more investment in long-term relationships, a reciprocally rewarding interpersonal orientation), include later sexual development and reproduction, and involve higher levels of parental investment per offspring (e.g., Belsky et al., 1991; Chisholm, 1999; Del Giudice et al., 2015; Ellis et al., 2009; Figueredo et al., 2013; Figueredo et al., 2006; Gibbons et al., 2012).

Variation in the development of life history strategies is sensitive to environmental factors, such as energy availability, extrinsic morbidity–mortality, and predictability of environmental conditions (Ellis et al., 2009; Kuzawa & Bragg, 2012). ³ For example, faster life histories result (in part) from trade-offs imposed by high levels of extrinsic morbidity-mortality (i.e., external sources of disability and death that are largely insensitive to the adaptive decisions of the organism). In a world of fleeting opportunities and threats without warning, the benefits of investing in morbidity and mortality reduction are low relative to the costs; consequently, future reproduction and other long-term investments are devalued. A fast strategy in this context that maximizes short-term gains (such as through high-risk behaviors that leverage positions in status hierarchies and access to mates) may enhance fitness despite the long-term costs (Ellis et al., 2012; Frankenhuis et al., 2016; Yao, Långström, Temrin, & Walum, 2014). In total, for both fast and slow strategies, trade-offs over development function to match the individual to local environmental conditions; ⁴ depending on those conditions, individuals can benefit from pursuing either faster or slower strategies.

We hypothesize that different skill sets will be associated with fast versus slow strategies, reflecting different cost-benefit trade-offs. As per the specialization and sensitization hypotheses, fast strategists should possess an adaptive suite of social/cognitive skills and abilities that are specialized for thriving in harsh, unpredictable environments (see Ellis & Del Giudice, 2014; Frankenhuis & de Weerth, 2013; Mittal et al., 2015), even though those very same skills and abilities may be costly (or less beneficial) in safe, stable environments. For example, an individual growing up in a chaotic/unpredictable environment may prioritize development of attention-shifting skills (to take advantage of fleeting opportunities and avoid unpredictable threats) at the cost of deprioritizing inhibitory control, whereas an individual growing up in a safe environment may make the opposite trade-offs (see Blackwell, Chatham, Wiseheart, & Munakata, 2014, for empirical evidence of such trade-offs in children).

Although suites of skills should be associated with different life history strategies, we do not expect that all individuals pursuing a particular strategy will have the same skill sets. People should invest in skills and abilities that are relevant in their developmental context. For example, a psychosocially neglected child and a physically abused child can be expected to develop overlapping skill sets that reflect the development of faster life history strategies in both family contexts (e.g., skill sets related to successfully attaining immediate rewards), but nonoverlapping skill sets that reflect differential exposures to violence. Thus, there should be divergence in skills and abilities across individuals who differ in levels and types of stress exposures.

In summary, we hypothesize that stress-adapted skills and abilities result from resource-allocation trade-offs that prioritize expenditures in ways that “make the best of a bad job” (by specializing skills and abilities to match high-adversity contexts), even though “the best” may constitute a high-risk strategy with substantial costs (see The Effects of Adversity on Cognitive and Academic Outcomes). These costs reflect the very nature of developmental trade-offs under harsh conditions (when “reserve capacity” is low): One system is diminished so that another system can be enhanced or preserved (e.g., Muehlenbein & Bribiescas, 2005; Pike, 2005). In the scientific literature on stress, however, these countervailing costs and benefits have not been equally studied. We know vastly more about the detrimental effects of childhood stress than its benefits in context.

The current adaptation-based approach to resilience attempts to address this lacuna. Its larger goal is to uncover the psychological adaptations that allow one to successfully navigate the challenges faced in high-stress developmental contexts. This approach converges on a central question: “What are the attention, learning, memory, problem-solving, and decision-making strategies that are promoted by exposures to childhood adversity?” At an applied level, this approach emphasizes: “What do youth from high-risk environments do well?” and “How can we work with, rather than against, these strengths to promote better intervention outcomes?” We now turn to a selective review of empirical studies of the specialization and sensitization hypotheses, focusing on both human and nonhuman animal literatures.

Literature Review on the Specialization and Sensitization Hypotheses: Animal Research

Consistent with the specialization and sensitization hypotheses, a large body of animal research suggests that early-life stress can enhance cognition and behavior. The majority of these studies have employed either avian or rodent models. Avian species provide a powerful model because of their widely varying life history strategies and because the same species often inhabits many different environments. Rodents provide a strong model for experimental studies of the effects of stress on the brain, behavior, and cognition because of the degree of control that researchers can exert while studying them in the laboratory, and because physiological stress response systems have been highly conserved in the evolutionary history of mammalian species.

Rodents

Much of what is known about the effects of early life stress on development comes from a vast rodent literature reported in over 6,000 peer-reviewed research articles extending over a century (reviewed in Howell, Neigh, & Sanchez, 2016). The large majority of this work has been conducted with rats and mice and has extensively examined the effects of both prenatal and postnatal stress (Howell et al., 2016). Much of the postnatal work focuses on observing or manipulating the powerful dam-pup relationship (e.g., maternal separation). Other common developmental stress manipulations include social isolation, restraint stress, social instability stress, social defeat stress, predatory stress, and the combination of multiple stressors. As in the bird literature, much of the rodent literature has shown that developmental exposures to stress have various negative effects such as disrupted HPA-axis and amygdala function, sustained anxiety-like and depressive-like behaviors, lower levels of adult hippocampal neurogenesis, impaired spatial learning and memory, impaired reversal learning, and increased vulnerability to drug abuse and addiction (e.g., Howell et al., 2016; Oitzl, Workel, Fluttert, Frösch, & De Kloet, 2000; Oomen et al., 2010). This body of work is important and has had many translational implications for understanding the developmental processes and biological mechanisms through which early stress “gets under the skin” to alter the phenotype.

However, these well-documented deleterious effects of stress are not the whole story. The other part of the story is that rodents show coherent developmental responses to ecologically relevant stressors by altering their phenotypes to match high-adversity contexts. This research, summarized in Table 2, has documented a variety of adaptive responses to early-life stress, ranging from reproductive adaptations (e.g., early puberty, greater skill at attracting mates) to cognitive adaptations (e.g., faster fear conditioning, enhanced striatal-dependent response learning, enhanced memory retention for early life events) to social adaptations (e.g., earlier and more frequent play behavior, increased dominance-related behavior and higher social rank).

OPEN IN VIEWER

Table 2.

Developmental Stress: Evidence for Adaptive Phenotypic Effects in Rodents

Hypothesis Type	Developmental Stress Exposure	Phenotypic Effects (Age at Testing)	Proposed Function
Specialization	Prenatal stress (repeated restraint of dams during pregnancy); predator stress, restraint, and tail shock (in adults)	Impaired hippocampus-dependent place learning (juveniles and adults; Kim, Lee, Han, & Packard, 2001; Lemaire, Koehl, Le Moal, & Abrous, 2000; Park et al., 2008; Wu et al., 2007), but enhanced striatal-dependent response learning and biased use of response learning on tasks that can be solved using either place learning or response learning (i.e., dual solution tasks; adults; Kim et al., 2001; Leong & Packard, 2014).	Enhanced stimulus-response/associative learning (over slower, cognitively expensive contextual learning) promotes immediate responding to environmental challenges in a dangerous world (as signaled by various developmental stress exposures).
Specialization	Juvenile exposure to elevated platform stress followed by adult exposure to acute swim stress	Rats that experienced both juvenile and adult stress (compared with rats that had experienced juvenile stress only, adult stress only, or no stress) displayed greater anxiety and enhanced spatial learning and memory as evidenced by performance in the Morris water maze (adults; Avital & Richter-Levin, 2005).	Match between juvenile programming environment and later adult environment promotes effective coping with danger (e.g., anxious/defensive behaviors and enhanced spatial learning/memory improve predator-avoidance).
Specialization	Low maternal licking and grooming (infancy); earlier age at weaning (postnatal day 21–28)	Earlier and more frequent play behavior (preweaning period; Franks, Champagne, & Curley, 2015); females: increased dominance-related behaviors and higher social rank (adults; Parent et al., 2013); males: higher levels of play fighting (juveniles; Parent & Meaney, 2008) and stronger defensive responses to an intruder (adults; Menard & Hakvoort, 2007).	Heightened pace of social development, more socially competitive behavior, and stronger defensive responding promote survival and earlier reproduction in harsh, competitive environments (as signaled by low maternal investment).
Specialization	Low maternal licking and grooming (infancy)	In females, earlier puberty and greater skill at attracting mates; in a competitive mating context, greater success at getting pregnant (adults; Cameron et al., 2008; Parent et al., 2013; Sakhai, Kriegsfeld, & Francis, 2011).	Harsh, unpredictable environments (as signaled by low maternal investment) promote trade-offs favoring current over future reproduction.
Specialization/sensitization	High maternal stress (produced by poor nesting materials during postnatal days 1–6); maternal deprivation (daily separations during first 2 weeks of life)	Accelerated maturation of (a) brain regions important for emotion expression, associative learning, and memory and (b) behaviors supported by these regions (e.g., faster fear conditioning, slower fear extinction, enhanced memory retention for early life events; postnatal days 7–47: infancy to adolescence; Callaghan & Tottenham, 2016a, 2016b; R. Richardson, Cowan, Callaghan, & Kan, 2016).	Stress acceleration hypothesis: Accelerated maturation reflects the allocation of developmental resources toward emotional systems and associative learning and memory in ways that confer a survival advantage through earlier self-regulation in harsh environments (as signaled by absent or inconsistent parental care; Callaghan & Tottenham, 2016b).
Specialization/sensitization	Predator exposure, social and physical stress (e.g., crowding, damp bedding) in adolescence	Faster decision making (shorter time to correct a choice and locate a food reward after an error); faster to explore novel environments and objects; enhanced reversal learning; increased monitoring of environment for threats; under currently threatening conditions (predation cues, bright lights), rats that experienced stress during adolescence performed better in a timed-foraging task (adults; Chaby, Cavigelli, et al., 2015; Chaby, Cavigelli, White, Wang, & Braithwaite, 2013; Chaby et al., 2016; Chaby, Sheriff, Hirrlinger, & Braithwaite, 2015).	Faster decision making and exploratory behavior, heightened vigilance, and reversal learning promote survival in harsh, unpredictable environments. Match between adolescent programming environment and current conditions promotes more successful foraging under threat.
Sensitization	Low maternal licking and grooming (infancy); maternal deprivation (24-hour separation at Postpartum Day 3)	Under low-stress conditions, reduced performance on tests of spatial learning and memory; but under high-stress conditions, enhanced hippocampal long-term potentiation (a cellular model of learning/memory) and enhanced memory on a hippocampal-dependent contextual fear-conditioning task (adults; Bagot et al., 2009; Champagne et al., 2008; Oomen et al., 2010).	Harsh, unpredictable early programming environments (as signaled by low maternal investment) prepare the animal to function under conditions of adversity later in life.
Sensitization	Maternal deprivation (daily separations over first 3 weeks of life); unpredictable stress in early adulthood	Following the unpredictable stress regimen, rats that had experienced maternal deprivation (compared with rats that had not) showed better hippocampal performance in a contextual fear conditioning test and enhanced contextual memory (adults; Zalosnik, Pollano, Trujillo, Suárez, & Durando, 2014).	Match between early programming environment and later adult environment promotes survival in dangerous environments through enhanced contextual fear conditioning and memory.

Although Table 2 reviews documented adaptive phenotypic responses to developmental stress, our ability to draw strong conclusions from the rodent literature has been limited by highly variable findings across laboratories (Macrì, 2013). The first column in Table 2, denoting whether a study tested for specialization or sensitization, may be relevant to explaining this replication problem. If advantages in social and cognitive skills in individuals who grew up under stressful conditions manifest primarily under currently stressful conditions (i.e., sensitization), then simply exposing an animal to developmental stress and then later testing it under generic conditions may not produce interpretable or consistent results. For example, under low-stress conditions in which rats are extensively habituated to testing conditions, rats that received high levels of parental investment from their mothers (i.e., high levels of licking and grooming) show enhanced performance on tests of spatial learning and memory, especially object recognition tests and the Morris water maze (reviewed in Bagot et al., 2009). However, when these highly nurtured rats are tested under high-stress conditions, they show reduced hippocampal long-term potentiation (LTP; a cellular model of learning and memory) and reduced memory on a hippocampal-dependent contextual fear-conditioning task. Consistent with the sensitization hypothesis, it is instead rats that experienced low levels of licking and grooming that excel in these stressful contexts (Bagot et al., 2009; Champagne et al., 2008).

Such sensitization effects may explain why rats who grow up under safe, stable laboratory conditions but then face acute stressors in adulthood (e.g., predation stress, restraint and tail shock, food restriction) show widely variable outcomes: they are not developmentally prepared to cope with these challenges (Macrì, 2013). A core assumption of the current adaptation-based approach to resilience is that individuals growing up in harsh, unpredictable environments specialize their cognitive abilities to match high-adversity contexts, as did the rats that experienced low levels of licking and grooming and then coped more successfully with acute challenges in adulthood. In the rodent literature, many of the studies documenting beneficial phenotypic effects of developmental stress exposures have employed this kind of a sensitization design (Table 2), in which the adult stress task is reminiscent of earlier life conditions.

Despite some ongoing issues regarding contradictory findings in the rodent literature, researchers have made steps toward explaining the beneficial phenotypic effects documented in Table 2 as adaptations in context. As in the avian literature, the rodent literature provides many apparent examples of functional adaptations to stress. As shown in Table 2, several of these cases involve specialization. For example, rodents exposed to various early-life stressors tend to prioritize development of the striatal-dependent associative learning and memory system (also commonly referred to as a “stimulus-response” system), which supports immediate responding to environmental challenges (Kim, Lee, Han, & Packard, 2001; Lemaire, Koehl, Le Moal, & Abrous, 2000; Leong & Packard, 2014; Park, Zoladz, Conrad, Fleshner, & Diamond, 2008). Likewise, maternal deprivation early in life promotes developmental trade-offs favoring current over future reproduction (i.e., rodents shift toward faster life histories), including accelerated pace of development, more socially competitive behavior, greater skill at attracting mates and achieving social dominance, and more competitive success at getting pregnant (Cameron, Fish, & Meaney, 2008; Franks, Champagne, & Curley, 2015; Parent, Del Corpo, Cameron, & Meaney, 2013; Parent & Meaney, 2008; Sakhai, Kriegsfeld, & Francis, 2011).

In other cases, as summarized in Table 2, adaptation in context involves sensitization. For example, rats that were exposed to a regimen of stressors in adolescence (i.e., encounters with predators, unpredictable social and physical stressors), compared with rats in an unstressed control condition, showed enhanced performance on a timed foraging task (i.e., they were more efficient at transitioning between foraging patches and consumed more food) when tested under high-threat conditions (i.e., predation cues, bright lights; Chaby, Sheriff, Hirrlinger, & Braithwaite, 2015). Likewise, developmental matching between the early programming environment (maternal deprivation) and the later adult environment (e.g., alarm bells, food deprivation, restraint) resulted in better hippocampal-dependent performance in a contextual fear conditioning test and enhanced contextual memory (Zalosnik, Pollano, Trujillo, Suárez, & Durando, 2014).

Taken together, a growing body of comparative avian and rodent research supports both the specialization and sensitization hypotheses: Harsh and unpredictable early-life environments do not solely impair behavior and cognition but specialize it for solving recurring adaptive problems faced in high-adversity contexts. The human literature, to which we now turn, also provides reasonable support for these hypotheses.

Literature Review on the Specialization and Sensitization Hypotheses: Human Research

As discussed earlier, the deficit model is the prevailing approach to studying the effects of stress on human development. This approach has been productive in mapping the pathways and mechanisms through which early-life stress disrupts neurobiological systems and cognitive development (see The Effects of Adversity on Cognitive and Academic Outcomes). The success of this approach, however, and the resulting dominant position of the deficit model in the field, has largely foreclosed consideration of the functional role of adaptations to stress in regulating normal variation in development across diverse contexts. Consequently, we know relatively little about beneficial phenotypic effects of developmental exposures to stress. However, with the recent emergence of evolutionary-developmental models, and especially life history theory, in the psychological sciences (e.g., Belsky et al., 1991; Del Giudice et al., 2011; Ellis & Del Giudice, 2014; Frankenhuis & de Weerth, 2013; Frankenhuis et al., 2016; Griskevicius, Tybur, Delton, & Robertson, 2011; Wenner, Bianchi, Figueredo, Rushton, & Jacobs, 2013), such beneficial effects have started to receive increasing attention.

There are two basic challenges in studying the effects of stress in regulating human development. First, the human literature is inherently limited relative to the avian and rodent literatures because early-life stress can only be measured and not experimentally manipulated. Thus, our state of knowledge is largely correlational. Second, although current stress exposures can be experimentally manipulated (e.g., through priming), ethical constraints dictate that those manipulations are relatively mild; many ecologically valid forms of threat and danger cannot be studied experimentally. Despite these limitations, there is wide variation in childhood exposures to harshness and unpredictability, and that variation can be validly measured (e.g., Belsky, Schlomer, & Ellis, 2012; Raver, Blair, & Garrett-Peters, 2015; Simpson, Griskevicius, Kuo, Sung, & Collins, 2012). Furthermore, a large corpus of research has shown that even the limited stress manipulations employed by psychologists can have substantial effects on human cognition and behavior (e.g., Griskevicius et al., 2011; Hill, Rodeheffer, DelPriore, & Butterfield, 2013; Mittal et al., 2015). Taken together, these factors enable meaningful (though less than definitive) tests of the specialization and sensitization hypotheses in humans.

An emerging body of human research has begun to document how early stress exposures regulate the development of skills and abilities to match high-adversity contexts. This research, summarized in Table 3, has documented a variety of adaptations to early-life stress, ranging from enhanced social-emotional skills (e.g., emotion recognition, empathic accuracy) to enhanced memory in specific domains (e.g., early-life memory retention; memory for negative, emotionally laden, or stressful events; working memory in the form of rapid tracking and memory updating), enhanced learning in specific domains (e.g., learning about animal danger, procedural learning), enhanced cognitive speed and accuracy in specific domains (e.g., recognition of angry or fearful faces), increased attention-shifting ability, and enhanced reward-oriented problem solving.

OPEN IN VIEWER

Table 3.

Developmental Stress: Evidence for Adaptive Phenotypic Effects in Humans

Hypothesis Type	Developmental Stress Exposure	Phenotypic Effects (Age at Testing)	Proposed Function
Specialization	Environmental harshness (maternal disengagement/low household income) at 2 years of age	Enhanced reward-oriented problem-solving but reduced abstract visual problem-solving (at age 4), but only in children scoring higher on hawk temperament traits (Suor et al., 2017).	Reward-oriented problem-solving leverages access to immediate rewards under harsh conditions but trades off against abstract problem-solving skills in this context.
Specialization	Interparental aggression (6, 15, 24, 35, and 58 months)	More exposure to interparental verbal aggression predicted greater accuracy at recognizing emotions (joy, sadness, anger, and fear); but more exposure to interparental physical aggression predicted lower emotion recognition accuracy (58 months; Raver, Blair, & Garrett-Peters, 2015).	Enhanced emotion recognition improves behavior prediction, which is vital to survival in hostile environments. Negative effect of physical aggression constitutes impairment (e.g., neurobiological disruption).
Specialization	Parental divorce/separation (< 7 years); exposure to stepparents and other extra adults while growing up	Early exposure to divorce, greater separation stress, and more nonbiological-parent adults in the household were associated with earlier memory retention (i.e., recall of earlier childhood events; Artioli, Cicogna, Occhionero, & Reese, 2012; Artioli & Reese, 2014; Artioli, Reese, & Hayne, 2015). a	Stress acceleration hypothesis (Callaghan & Tottenham, 2016b; described in Table 2).
Specialization	Low maternal age, maternal smoking during pregnancy, low family SES (early childhood)	Enhanced performance on response shifting but lower performance on verbal memory in kindergarteners (Vandenbroucke et al., 2016).	Enhanced response shifting promotes anticipation of threats and fleeting opportunities in harsh/unpredictable environments (common at low SES). This skill trades off against other executive functions.
Specialization	Insecure attachment as determined by the Strange Situation (12 months)	Compared with securely attached children, better recall of negative events (e.g., spilling juice) but worse recall of positive events (e.g., receiving a birthday present) seen in a puppet show (3 years of age; Belsky, Spritz, & Crnic, 1996).	Enhanced recall of negative events promotes detection/avoidance of negative events in the future, which are more likely to occur in harsh family environments (as developmentally embedded through insecure attachment). This skill trades off against recall of positive events.
Specialization	Low perceived or objective SES (at time of testing)	Enhanced response inhibition; better empathic accuracy; physiological responses and emotional contagion patterns more empathically linked to a social interaction partner (college and community adult samples; Kraus, Piff, Mendoza-Denton, Rheinschmidt, & Keltner, 2012; Na & Chan, 2016).	Enhanced empathic accuracy promotes behavioral prediction/management of external social forces/individuals that influence one’s life outcomes (which occurs chronically for low SES individuals).
Specialization	High risk from dangerous animals (birth to age at testing)	Enhanced learning about animal danger but not about dangers posed by material artifacts (ages 4–9 years: Barrett, Peterson, & Frankenhuis, 2016).	Enhanced learning about dangerous animals promotes survival in environments where animal hazards are prevalent.
Specialization	Institutional care; unpredictable/unstable family environments (early childhood)	Enhanced performance on a risk-taking task that rewards exploitation (i.e., taking advantage of an immediate and small, but probable, reward; ages 6–15 years; Humphreys et al., 2015); reduced effortful control (delay control) on a “hot” task involving fleeting rewards, but not on a “cool” task lacking an affective-motivational component (age 4 years, 6 years: Sturge-Apple et al., 2016).	Exploiting immediate rewards optimizes expected returns in unpredictable environments, where individuals have little control over future reward outcomes.
Specialization	Harsh parenting (birth to 16 years, assessed retrospectively)	Enhanced deception detection in college sample; effect did not replicate in community sample (Frankenhuis, Roelofs, & de Vries, in press).	Increased deception detection reduces probability of morbidity, mortality, and exploitation in hostile family environment.
Specialization/sensitization	Maternal anxiety/stress/depressive symptoms (prenatal; 3–24 months postnatal)	Enhanced psychomotor and mental development (3–24 months postnatal) when exposed to concordant prenatal and postnatal maternal depression (DiPietro, Novak, Costigan, Atella, & Reusing, 2006; Sandman, Davis, & Glynn, 2012)	Matching hypothesis: Prenatal stress prepares the organism for stressful conditions later in life.
Specialization/sensitization	Severe neglect and/or abuse as determined by protective services (birth to age at testing)	Faster orientation to angry faces and voices; greater accuracy in identifying angry facial expressions from degraded stimuli; greater speed (but not accuracy) in identifying fearful faces (8–15 years; C. L. Masten et al., 2008; Pollak, 2008; Pollak, Messner, Kistler, & Cohn, 2009).	Faster and more accurate detection of threat promotes survival in hostile environments.
Specialization/sensitization	Child neglect; sexual and/or physical abuse as determined by protective services (birth to age at testing)	Heightened attention to and memory for negative, emotionally laden or stressful information (Goodman, Quas, & Ogle, 2009); e.g., enhanced memory for a doctor who performed an invasive examination (3–16 years; Eisen, Goodman, Qin, Davis, & Crayton, 2007) or distracting aggressive stimuli (e.g., guns, swords; 4–9 years; Rieder & Cicchetti, 1989).	Enhanced detection of and memory for threat promotes survival in hostile environments.
Sensitization	High childhood unpredictability/chaos (birth to 10 years) b	Under primed conditions of economic decline/uncertainty, enhanced attention-shifting ability but reduced inhibitory control (deliberate overriding of dominant responses; college student samples; community sample: age 37; Mittal et al., 2015).	Enhanced shifting ability promotes detection of threats and taking advantage of fleeting opportunities in chaotic/unpredictable environments. Attention shifting trades off against inhibitory control.
Specialization/sensitization	High childhood unpredictability/chaos (birth to 10 years, assessed retrospectively)	Under primed conditions of economic decline/uncertainty, enhanced working memory in the form of rapid tracking and memory updating but reduced working memory in the form of long-term storage and information retention in the face of distraction (community adult samples; Young, Griskevicius, Simpson, Waters, & Mittal, 2016).	Enhanced rapid tracking and updating ability in working memory promotes the availability of current information, which is essential in chaotic/unpredictable environments, but trades off against longer term working memory functions.
Sensitization	Low SES (at time at testing)	Under primed conditions of high financial demand, enhanced procedural learning (acquiring stimulus-response associations) but reduced performance on cognitive functions that rely heavily on working memory (college and community adult samples; Dang et al., 2016; Mani, Mullainathan, Shafir, & Zhao, 2013).	Enhanced procedural learning promotes acquisition of stimulus-response associations, which promote fast responding to environmental challenges (common at low SES), but trade off against working memory (DeCaro, Thomas, & Beilock, 2008).

a

Results based on a middle-childhood sample (age at testing: 7–11 years) and multiple college student samples. ^bChildhood unpredictability was assessed retrospectively in the college student samples and prospectively in the community sample.

The findings summarized in Table 3 primarily focus on one part of the story: beneficial phenotypic effects. The other part of the story—detrimental phenotypic effects—has already been extraordinarily well documented (as reviewed in The Effects of Adversity on Cognitive and Academic Outcomes). A well-rounded analysis thus suggests that developmental exposures to stress have a mix of adaptive and harmful effects, as highlighted by several of the entries in Table 3 (e.g., exposure to interparental verbal aggression improves, but exposure to interparental physical aggression reduces, accuracy at recognizing emotions; Raver et al., 2015). Such mixed findings may reflect, in part, inadequate attention to sensitization effects (in terms of matching between earlier and later stress exposures). For example, human infants show reduced psychomotor and mental development during the first year of life when they have been exposed to discordant, as opposed to concordant, levels of prenatal and postnatal maternal depression, even though the concordant condition involves greater cumulative exposure to stress (Sandman, Davis, & Glynn, 2012). Studies that have examined the effects of maternal depression, anxiety, and stress during pregnancy on child cognitive outcomes, without taking into account matching with postnatal environments, have produced mixed results (e.g., DiPietro, Novak, Costigan, Atella, & Reusing, 2006; Glover, 2014).

Many of the findings reported in Table 3 can be interpreted as adaptations in context. Several cases involve specialization. For example, children who have experienced severe neglect or abuse tend to exhibit improved detection, learning, and memory on tasks involving stimuli that are ecologically relevant to them (reviewed in Frankenhuis & de Weerth, 2013), such as enhanced memory for a doctor who performed an invasive examination (Eisen, Goodman, Qin, Davis, & Crayton, 2007), or enhanced recall of distracting aggressive stimuli (e.g., guns, swords; Rieder & Cicchetti, 1989), or faster orientation to angry faces and voices (Pollak, 2008; Pollak, Messner, Kistler, & Cohn, 2009). These cognitive skills may promote survival in hostile environments. Other research conducted with more normative samples also supports the specialization hypothesis. Much of this work employs SES as an indicator of developmental stress. One significant finding to emerge from this literature is that lower-SES individuals have an advantage in social-cognitive tasks involving contextual information, such as the ability to accurately read others’ affective states (reviewed in Kraus, Piff, Mendoza-Denton, Rheinschmidt, & Keltner, 2012). For example, in one study, high-school-educated university employees outperformed college-educated university employees on a standard test of empathic accuracy (which involved labeling, with emotion terms, different posed facial expressions; Mayer, Salovey, & Caruso, 2002). Enhanced empathic accuracy may promote behavioral prediction and management of external social forces and individuals that exert substantial control over one’s life (which occurs chronically for people at low SES; see Kraus et al., 2012).

In other cases, as summarized in Table 3, adaptation in context involves sensitization (whereby skills and abilities promoted by exposures to early-life stress are enhanced under currently stressful conditions that are reminiscent of earlier childhood experiences). For example, individuals who grew up under conditions of low SES (compared with others who grew up under high SES conditions) show enhanced procedural learning (i.e., stimulus–response mapping in categorization tasks), but reduced performance on cognitive functions that rely heavily on working memory, when tested under primed conditions of high financial demand (Dang et al., 2016; Mani et al., 2013). Although procedural learning may be enhanced under the kinds of high-stress conditions that chronically occur at low SES, procedural learning processes apparently trade off against working memory (DeCaro, Thomas, & Beilock, 2008). ⁵ Likewise, in a longitudinal study of a children born into poverty, adults who experienced high family unpredictability during their first 10 years of life (i.e., parental transitions, residential changes, parental job changes), compared with others who experienced low family unpredictability, showed enhanced ability in attention shifting (a component of executive function that involves efficiently switching between different tasks) when tested under conditions of primed economic decline/uncertainty (Mittal et al., 2015). Unpredictable early-life environments also enhance aspects of working memory central to tracking novel information in the environment. Youth who grew up under more unpredictable environmental conditions, compared with others who were reared in more predictable environments, were able to track a larger amount of information in their working memory when tested under conditions of primed economic decline/uncertainty (Young, Griskevicius, Simpson, Waters, & Mittal, 2016). These kinds of enhanced abilities in attention shifting and working memory are likely to promote the detection of threats and taking advantage of fleeting opportunities in chaotic/unpredictable environments. At the same time, these enhanced skills may trade off against poorer inhibitory control (Mittal et al., 2015) and worse performance on aspects of working memory that involved long-term storage and information retention in the face of distraction (Young et al., 2016).

In total, research in birds, rodents, and humans suggests that developmental exposures to stress can improve forms of attention, perception, learning, memory, and problem solving that are ecologically relevant in harsh or unpredictable environments. These findings suggest that early-life stress not only impairs cognitive development (as in the well-documented negative phenotypic effects) but also directs or regulates cognitive development toward prioritizing skills and abilities that are adaptive in context (as per the specialization hypothesis). Many of these skills and abilities, moreover, are primarily manifest in currently stressful contexts where they would provide their most powerful fitness-relevant advantages (as per the sensitization hypothesis).

Discussion

How does repeated or chronic childhood adversity shape biobehavioral development and, through it, social and cognitive abilities? In the developmental sciences, there is a widely accepted answer to this question. Instantiated in various deficit models, such as cumulative risk (e.g., Evans et al., 2013; Sameroff et al., 1987; Seifer et al., 1996), toxic stress (Shonkoff et al., 2012), and allostatic load (Lupien et al., 2006; McEwen & Stellar, 1993), that answer posits a striking duality: Biobehavioral responses to stress may be adaptive in the short term, but protracted activation of stress responsive systems is maladaptive and toxic in the long term. Repeated or chronic childhood adversity causes disruptions of brain structure and function, resulting in dysregulation of neurobiological mediators “that are the precursors of later impairments in learning and behavior as well as the roots of chronic, stress-related physical and mental illness” (Shonkoff et al., 2012, p. e236). In these deficit models, children from high-stress backgrounds are considered to be at risk for impaired development, and the intervention goal is to prevent, reduce, or repair the damage.

Missing from this traditional intervention approach is an attempt to leverage the unique strengths and abilities that develop in response to high-stress environments. As instantiated in the specialization and sensitization hypotheses, a core assumption of evolutionary-developmental models is that exposures to stress do not so much impair development as direct or regulate it toward strategies that are adaptive under stressful conditions (see The Adaptation-Based Approach to Resilience). From this perspective, deficit models miss something fundamental about development: They miss the coherent, functional biobehavioral changes that occur in response to stress over time, including regulation of alternative life history strategies (Ellis & Del Giudice, 2014). These changes not only promote adaptation to harsh, unpredictable childhood environments (as reflected in such traits as heightened vigilance, attention shifting, and empathic accuracy), but also shape longer-term developmental trajectories to match expected future conditions. Fast life histories involve prioritizing investment in specific skills and abilities that enable one to survive and reproduce under harsh, unpredictable conditions, even though those very same skills and abilities may be costly (or less beneficial) in safe, stable environments.

The implications of this theory are far-reaching. If individual differences in skills and abilities largely reflect adaptation in context, then stress-adapted individuals growing up in harsh, unpredictable environments should develop heightened skills and abilities relevant to solving adaptive problems faced in those environments (specialization) and might outperform individuals growing up in safe, stable environments when tested under conditions resembling the high-risk contexts to which they are ostensibly adapted (sensitization). As reviewed in Tables 1 to 3 and discussed in the previous two sections, research in birds, rodents, and humans suggests that developmental exposures to stress can improve forms of attention, perception, learning, memory, and problem solving that are ecologically relevant to survival and reproduction in harsh, unpredictable environments. Although the effects of early life stress are clearly mixed with a combination of positive and negative effects, the current literature suggests that exposures to adversity regulate cognitive development toward prioritizing skills and abilities that are adaptive in context (as per the specialization hypothesis). Many of these skills and abilities, moreover, are primarily manifest in currently stressful contexts where they would provide the greatest fitness-relevant advantages (as per the sensitization hypothesis). Such sensitization effects may reflect (in part) the lower reserve capacity of stress-adapted individuals, who may need to conserve resources under basal conditions in order to more fully activate relevant neurocognitive processes in context.

Despite our overall emphasis on the cognitive strengths of stress-adapted children and youth, the specialization and sensitization hypotheses have both hard and soft versions. The hard version focuses on enhanced abilities; it proposes that stress-adapted individuals have “hidden talents” that exceed the abilities of individuals growing up under more safe, stable conditions. By contrast, the soft version focuses on equalization; it also proposes that stress-adapted individuals have hidden talents, but that these talents enable them to perform as well as individuals from low-risk environments (i.e., to “level the playing field”) in certain domains. Both the hard and soft versions of specialization/sensitization highlight the need to uncover a high-resolution map of specific skills and abilities that either are (a) enhanced by or (b) maintained in spite of growing up under harsh, unpredictable conditions, as both (a) and (b) are highly relevant to the design of interventions that work with these skills and abilities. Moreover, even a weaker version of (b), in which disparities in relevant skills and abilities are reduced but not eliminated, would still be informative for intervention.

Conclusion

Comparative research on birds, rodents, and humans highlights the role of stress in regulating development of adaptively relevant skills and abilities. Leveraging these strengths could help stress-adapted individuals achieve their full potential and lead more satisfying and productive lives. This adaptation-based approach to resilience thus converges on a pressing research agenda: to uncover a high-resolution map of the skills and abilities of children and youth from high-risk backgrounds (specialization), and to determine how developmental exposures to stress interact with current psychological states and conditions to regulate learning and performance (sensitization). In the future, we envision an assessment battery (the first of its kind) that captures the “hidden talents” of stress-adapted young people, which can be used to inform efforts and programs that potentiate their success in education, employment, and civic life. This approach affords a more positive view of youth from diverse backgrounds that both avoids stigma and communicates a distinctive strength-based message to the public, the individuals affected by adversity, and policy makers.

In pursuing this adaptation-based approach to resilience, the field will need to pay careful attention to (a) dimensions of childhood stress (e.g., acute vs. chronic vs. unpredictable), (b) types of developmental stress (e.g., resource deprivation, child abuse vs. child neglect), (c) the timing of stress exposures (e.g., sensitive periods, delays between developmental stress exposures and the manifestation of beneficial/detrimental effects), (d) whether a given stress exposure is still occurring at the time of testing, (e) interactions between developmental stress exposures and current states (i.e., under what conditions, and for which abilities, do we observe sensitization effects?), (f) age at testing (i.e., at what age do adaptively relevant skills emerge?), and (g) intervening neurobiological and cognitive mechanisms that mediate the development of stress-adapted skills. Moreover, any tests of the specialization and sensitization hypotheses should explicitly consider the extent to which the content of testing materials (e.g., use of nonsense syllables vs. ecologically valid stimuli), the format through which tasks are presented (e.g., words vs. pictures) and responded to (e.g., spoken vs. written), and the test setting (e.g., lab vs. community context) match the experiences of stress-adapted people.

Addressing these issues will require continued research efforts from behavioral scientists coupled with increased dialogue and collaboration with teachers, workforce development specialists, and other community-engaged professionals who work with stress-adapted children and youth. Such cross-disciplinary interaction should promote a better understanding of how classroom environments, instructional strategies, job training, and related domains can be designed to support stress-adapted individuals in ways that converge with their life experiences to leverage their unique skills and abilities. Although much research is needed, this work holds the promise for transformative interventions that work with, instead of against, the skills and abilities of individuals from a diverse range of life circumstances.