Extensive evidence suggests that prenatal stress is a risk factor, undermining child well-being, as reflected in chronic health, behavioral, and cognitive problems (for a review, see Entringer, Buss, & Wadhwa, Reference Entringer, Buss and Wadhwa2015; Glover, Reference Glover2014; van den Bergh, Mulder, Mennes, & Glover, Reference van den Bergh, Mulder, Mennes and Glover2005). For example, prenatal stress is associated in prospective studies with preterm birth and low birth weight (for review, see Wadhwa et al., Reference Wadhwa, Glynn, Hobel, Garite, Porto, Chicz-DeMet and Sandman2002), deficiencies in intellectual and language functioning (Laplante et al., Reference Laplante, Barr, Brunet, Du Fort, Meaney, Saucier and King2004), attention-deficit/hyperactivity disorder symptoms (Grossman et al., Reference Grossman, Churchill, McKinney, Kodish, Otte and Greenough2003), externalizing and anxiety problems (Glover, Reference Glover2011), and motor and mental developmental disorders (Kofman, Reference Kofman2002). Although such findings are routinely interpreted in the human literature as evidence that prenatal stress disrupts “optimal” development, herein we review evidence for a radically different interpretation of how and why prenatal stress is associated in observational studies with the negative developmental phenotypes to which it has been repeatedly related. We build the case that prenatal stress programs postnatal developmental plasticity, further developing an argument first advanced by Pluess and Belsky (Reference Pluess and Belsky2011).
We begin by outlining the theoretical framework of differential susceptibility, which has been used to guide much recent research on individual differences in environmental sensitivity (Belsky, Reference Belsky1997, Reference Belsky, Ellis and Bjorklund2005; Belsky, Bakermans-Kranenburg, & van IJzendoorn, Reference Belsky, Bakermans-Kranenburg and van Ijzendoorn2007; Belsky & Pluess, Reference Belsky and Pluess2009, Reference Belsky and Pluess2013; Ellis, Boyce, Belsky, Bakermans-Kranenburg, & van IJzendoorn, Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and van IJzendoorn2011; Ellis, Shirtcliff, Boyce, Deardorff, & Essex, Reference Ellis, Shirtcliff, Boyce, Deardorff and Essex2011). Within this first major section, we highlight empirical evidence that infant negative emotionality and physiological reactivity, two well-documented sequelae of prenatal stress, are markers of increased susceptibility to both positive and negative developmental experiences and environmental exposures. In the paper's second major section, we review a separate line of research that has consistently linked prenatal stress to these two susceptibility markers, heightened negative emotionality and physiological reactivity. This leads us to return to Pluess and Belsky's (Reference Pluess and Belsky2011) hypothesis that prenatal stress programs postnatal plasticity, sharing recent evidence consistent with this proposition, including new experimental research in which prenatal stress is manipulated, as is postnatal rearing. In so doing, we will highlight the many different ways in which prenatal stress has been operationalized in the developmental literature. After reviewing this work, we outline future directions for research, focusing on mechanisms that could instantiate enhanced plasticity and potential moderators of prenatal programming affects. After considering, then, how prenatal stress may promote postnatal plasticity and for whom this may be more and less likely, we conclude by considering the ultimate, evolutionary issue, namely, why such prenatal programming in response to prenatal stress may have evolved.
Differential Susceptibility
By applying an evolutionary analysis to human development, Belsky (Reference Belsky1997, Reference Belsky, Ellis and Bjorklund2005; Belsky & Pluess, Reference Belsky and Pluess2009, Reference Belsky and Pluess2013) proposed that individuals should vary in their susceptibility (i.e., developmental plasticity) to environmental influences and especially those of the rearing environment (Boyce & Ellis, Reference Boyce and Ellis2005). This proposition was based on appreciation that the future is, and always has been, inherently uncertain. Thus, to maximize the likelihood of genetic material being passed from one generation to the next (i.e., reproductive fitness), natural selection should have crafted offspring to vary in their susceptibility.
The reasoning for this claim becomes apparent when we consider the case of an environmental mismatch between the rearing environment and the future context in which the developing individual finds him/herself. If, for instance, an environmental mismatch occurred whereby the rearing environment did not match the adult environment, it would prove more costly for the individual whose development was heavily influenced by his or her early environment than, perhaps, the individual who was not, and might better fit that future environment. In an effort to mitigate this ever-present risk of a potentially changing environment, Belsky (Reference Belsky1997, Reference Belsky, Ellis and Bjorklund2005) theorized that nature should have selected for humans to vary in their susceptibility to parental as well as other environmental influences. This way, not every individual would end up developmentally mismatched to his/her future environment when the rearing environment and the future environment turned out to be rather different (i.e., mismatched).
On the basis of this theoretical analysis, it follows that individuals should vary in their susceptibility to environmental influences. One can think, typologically, then, of two developmental strategies: “plastic—or conditional—strategists” are those whose development is heavily shaped by their developmental experiences, whereas “fixed—or alternative—strategists” are those whose development is relatively unaffected by their early environment and whose development is more rather than less canalized (Belsky, Reference Belsky, Rodgers, Rowe and Miller2000). It should be appreciated that the kind of variation in developmental plasticity just illustrated may be best conceptualized in dimensional rather than typological terms, with some being more and some less susceptible to environmental factors and forces rather than some being highly susceptible and others not at all susceptible.
Having delineated the theoretical logic underlying differential-susceptibility thinking, attention is now turned to organismic factors associated with greater developmental plasticity. We consider first negative emotionality and, thereafter, physiological reactivity.
Negative Emotionality as a Phenotypic indicator of Plasticity
Some of the earliest evidence documenting differential susceptibility to environmental influences emerged from research on Temperament × Parenting interaction (Belsky, Reference Belsky1997, Reference Belsky, Ellis and Bjorklund2005; Belsky et al., Reference Belsky, Bakermans-Kranenburg and van Ijzendoorn2007), a long-standing focus of developmental inquiry (Rothbart & Bates, Reference Rothbart, Bates, Damon, Lerner and Eisenberg2006; Slagt, Dubas, Deković, & van Aken, Reference Slagt, Dubas, Deković and van Aken2016). Appreciation of the role that temperament might play in making some children more susceptible to environmental influences than others was not the result of any theoretical analysis or expectation but rather emerged as an empirical observation once evidence consistent with differential-susceptibility theorizing was sought. In reviewing relevant evidence, Belsky (Reference Belsky, Ellis and Bjorklund2005) observed that the effect of rearing experience on a variety of psychological and behavioral outcomes was consistently greater for a subgroup of infants and toddlers who could be characterized as highly negatively emotional (e.g., irritability, fearfulness, and inhibition) or as having a difficult temperament. Even if such temperamental styles conferred developmental risk under aversive contextual conditions (e.g., maternal depression and harsh parenting), as long appreciated, they also predisposed children to benefit more than others from benign or especially supportive developmental circumstances (e.g., sensitive parenting and high-quality child care). Such enhanced susceptibility to effects of both positive and negative contextual conditions has been referred as increased likelihood of being affected “for better and for worse” (Belsky et al., Reference Belsky, Bakermans-Kranenburg and van Ijzendoorn2007).
In their reviews of the differential-susceptibility-related literature, Belsky and Pluess (Reference Belsky and Pluess2009, Reference Belsky and Pluess2013) highlighted a range of evidence indicating that negative emotionality functioned as a plasticity factor. This included work documenting the heightened environmental sensitivity (“for better and for worse”) of children with high levels of negative emotionality in studies linking maternal empathy (Pitzer, Jennen-Steinmetz, Esser, Schmidt, & Laucht, Reference Pitzer, Jennen-Steinmetz, Esser, Schmidt and Laucht2011) and anger (Poehlmann et al., Reference Poehlmann, Hane, Burnson, Maleck, Hamburger and Shah2012) with externalizing problems; mutual responsiveness observed in the mother–child dyad with effortful control (Kim & Kochanska, Reference Kim and Kochanska2012); intrusive maternal behavior (Conway & Stifter, Reference Conway and Stifter2012) and poverty (Raver, Blair, & Willoughby, Reference Raver, Blair and Willoughby2012) with executive functioning; sensitive parenting with social, emotional, and cognitive–academic development (Roisman et al., Reference Roisman, Newman, Fraley, Haltigan, Groh and Haydon2012); teacher–child conflict with change in symptomology during the primary-school years (Essex, Armstrong, Burk, Goldsmith, & Boyce, Reference Essex, Armstrong, Burk, Goldsmith and Boyce2011); mother's depressive symptoms with child adjustment (Dix & Yan, Reference Dix and Yan2014); maternal responsiveness with adolescent allostatic load (Dich, Doan, & Evans, Reference Dich, Doan and Evans2015); and of coercive parenting with adolescent alcohol use (Rioux et al., Reference Rioux, Castellanos-Ryan, Parent, Vitaro, Tremblay and Séguin2016). Perhaps qualifying some of these findings are the results of a recent meta-analysis of research on Parenting × Temperament interaction, as it revealed that the “for better and for worse,” differential-susceptibility-related effect was restricted to investigations that assessed negative emotionality in infancy, not later in life. When, meta-analytically, negativity was examined as a moderator of parenting effects at older ages, results proved consistent with diathesis–stress thinking (Slagt et al., Reference Slagt, Dubas, Deković and van Aken2016).
It is well appreciated that rearing effects chronicled in observational studies like those just cited may actually be the result of third variables (e.g., genetics) and not capture true causal influence. This makes experimental research particularly important (Bakermans-Kranenburg & van IJzendoorn, Reference Bakermans-Kranenburg and van IJzendoorn2015). Especially notable, then, are findings from a recent randomized control trial evaluating the effects of an intervention designed to enhance children's language development (van den Berg & Bus, Reference van den Berg and Bus2014). In line with differential-susceptibility thinking, highly reactive children whose parents received the intervention showed the greatest increase in language development skills, and the poorest performance when randomized to the control group, with the intervention proving entirely ineffective for children who were not highly reactive. Thus, findings from both observational and experimental studies prove consistent with the proposition that negative emotionality is a behavioral indicator of enhanced developmental plasticity, “for better and for worse.” No longer, then, should negativity be regarded solely as a development risk factor. It would seem to be just as much an “opportunity” factor.
Physiological reactivity as an endophenotypic indicator of plasticity
Boyce and Ellis's (Reference Boyce and Ellis2005) also advanced an evolutionary-inspired differential-susceptibility model of environmental influences, referred to as the biological sensitivity to context (BSC) framework. In contrast to Belsky's (Reference Belsky, Ellis and Bjorklund2005; Belsky et al., Reference Belsky, Bakermans-Kranenburg and van Ijzendoorn2007; Belsky & Pluess, Reference Belsky and Pluess2009, Reference Belsky and Pluess2013) theorizing, the BSC model was based on a biological mechanism instantiating differential susceptibility to environmental influence, namely, physiological reactivity. Children with heightened physiological reactivity, Boyce and Ellis (Reference Boyce and Ellis2005) theorized, would be more affected by their environment, in a “for better and for worse” manner, than those not as physiologically reactive. Of note, this theorizing was post hoc and emerged in attempt to explain unanticipated findings emanating from work carried out a decade earlier by Boyce et al. (Reference Boyce, Adams, Tschann, Cohen, Wara and Gunnar1995).
Empirical support for BSC thinking emerged in the years since the theory was promulgated. Evidence consistent with the claim that more physiologically reactive children would prove more susceptible to environmental effects, “for better and for worse,” than other children has been detected in research evaluating effects of actual marital conflict (Obradović, Bush, & Boyce, Reference Obradović, Bush and Boyce2011) and simulated interparental aggression (Davies, Sturge-Apple, & Cicchetti, Reference Davies, Sturge-Apple and Cicchetti2011) on externalizing problems; of family adversity on school achievement (Obradović, Bush, Stamperdahl, Adler, & Boyce, Reference Obradović, Bush, Stamperdahl, Adler and Boyce2010); of attachment security, presumed to itself reflect rearing experience, on problem behavior (Conradt, Measelle, & Ablow, Reference Conradt, Measelle and Ablow2013); of changes in paternal depressive symptoms on child internalizing behavior (Laurent et al., Reference Laurent, Leve, Neiderhiser, Natsuaki, Shaw, Harold and Reiss2013); of family aggression on posttraumatic stress symptoms/antisocial behavior (Saxbe, Margolin, Spies Shapiro, & Baucom, Reference Saxbe, Margolin, Spies Shapiro and Baucom2012); of the family environment on pubertal development (Ellis, Boyce, et al., Reference Ellis, Shirtcliff, Boyce, Deardorff and Essex2011); of teacher–child conflict on change in symptom severity (Essex et al., Reference Essex, Armstrong, Burk, Goldsmith and Boyce2011); of harsh discipline on externalizing problems (Chen, Raine, et al., Reference Chen, Nolte and Schlötterer2015); and of family income on early executive function (Obradović, Portilla, & Ballard, Reference Obradović, Portilla and Ballard2016).
Given concerns already raised about the limits of observational research when it comes to inferring causation, it is also notable that there is some experimental evidence documenting the plasticity-enhancing role of elevated physiological reactivity. Specifically, highly reactive children benefited from a psychotherapeutic intervention designed to reduce problem behavior, whereas the same was not so for other children (van de Wiel, van Goozen, Matthys, Snoek, & Engeland, Reference van de Weil, van Goozen, Matthys, Snoek and van Engeland2004). Heightened physiological reactivity would also seem to function, then, as both a risk and an opportunity factor.
Prenatal Stress and Emotional/Physiological Reactivity
Evidence just summarized indicating that highly negatively emotional and physiologically reactive infants, toddlers, and perhaps children as well evince greater developmental plasticity than do others becomes especially intriguing when juxtaposed to independent evidence linking prenatal stress with both of these plasticity markers. It is well documented that prenatal stress, measured in a variety of ways (e.g., maternal anxiety and cortisol), predicts greater behavioral and physiological dysregulation in infancy and childhood. With regard to behavioral dysregulation, exposure to prenatal stress, measured in a variety of ways (e.g., maternal psychological distress and maternal cortisol) at different gestational times, is associated with increased displays of sadness, frustration, and fear, as well as a stable disposition of heightened (negative) emotional reactivity (Huizink, De Medina, Mulder, Visser, & Buitelaar, Reference Huizink, De Medina, Mulder, Visser and Buitelaar2002; van den Bergh et al., Reference van den Bergh, Mulder, Mennes and Glover2005). Research also documents associations linking maternal psychological stress during late pregnancy with the increased behavioral reactivity of 4-month-olds (Davis et al., Reference Davis, Snidman, Wadhwa, Glynn, Schetter and Sandman2004) and maternal psychological distress, during early pregnancy, to irregular sleeping and eating patterns of 6-month-olds and heightened inhibition and negative emotionality of 5-year-olds (Martin, Noyes, Wisenbaker, & Huttenen, Reference Martin, Noyes, Wisenbaker and Huttenen1999). Relatedly, higher levels of maternal cortisol in late pregnancy forecast fussier infant behavior, including more negative facial expressions and increased frequency of crying at 7 weeks of age (de Weerth, van Hees, & Buitelaar, Reference de Weerth, van Hees and Buitelaar2003). Especially noteworthy is research showing that elevated levels of both maternal cortisol in late pregnancy and psychosocial problems (i.e., anxiety and depression) in middle and late pregnancy predict greater infant negativity at 2 months of age even when controlling for maternal postnatal psychological state (Davis et al., Reference Davis, Glynn, Schetter, Hobel, Chicz-Demet and Sandman2007).
Just as notable, perhaps even more so, is a recent prospective study exploring effects of prenatal stress, indexed via amniotic cortisol during the second trimester of pregnancy, on child's birth weight and temperament at 3 months of age (Baibazarova et al., Reference Baibazarova, van de Beek, Cohen-Kettenis, Buitelaar, Shelton and van Goozen2013). Results revealed that higher levels of amniotic cortisol predicted more negative temperament via reduced birth weight (i.e., cortisol → birth weight → temperament). In addition, low birth weight, which has been consistently linked to prenatal stress, even in genetically informed work (Rice et al., Reference Rice, Harold, Boivin, van den Bree, Hay and Thapar2010), is associated with negative emotionality (Pluess & Belsky, Reference Pluess and Belsky2011). Notable, too, is work showing that pregnant women exposed to a natural disaster (i.e., the 1998 Canadian ice storm), who experienced greater subjective distress or illness/infection at various time points in their pregnancy, had infants with more difficult temperaments; and these relations, too, remained significant after controlling for postpartum depression and major life events (Laplante, Brunet, & King, Reference Laplante, Brunet and King2015).
Turning to physiological functioning, research reveals that prenatal-stress exposure is associated with dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis in infants and children, as reflected in greater maternal depression in middle pregnancy predicting elevated basal cortisol concentrations in newborns (Field et al., Reference Field, Diego, Dieter, Hernandez-Reif, Schanberg, Kuhn and Bendell2004) and higher maternal cortisol in middle and late pregnancy predicting greater cortisol response to a heel-prick 24 hr after birth (Davis, Glynn, Waffarn, & Sandman, Reference Davis, Glynn, Waffarn and Sandman2011). These latter effects appear to be at least partly mediated via epigenetic changes in the glucocorticoid receptor gene (NR3C1), which encodes for glucocorticoid receptor, a major component of the stress response (Oberlander et al., Reference Oberlander, Weinberg, Papsdorf, Grunau, Misri and Devlin2008). Such effects on children's cortisol levels as a function of expectant-mothers’ heightened pregnancy-specific fears and cortisol levels measured at multiple times throughout pregnancy extend to even the first day of school (Gutteling, de Weerth, & Buitelaar, Reference Gutteling, de Weerth and Buitelaar2005). The results of a natural experiment in humans positioned near the World Trade Center on 9/11 also documents prenatal-stress effects on infant stress physiology; pregnant mothers present or near the 9/11 terrorist attacks who subsequently developed posttraumatic stress disorder had infants with dysregulated diurnal cortisol rhythms at 1 year of age relative to infants of other mothers (Yehuda et al., Reference Yehuda, Engel, Brand, Seckl, Marcus and Berkowitz2005). These results are consistent with experimentally documented findings in rodent studies indicating that prenatal stress induced by restraint stress or social stress is associated with higher baseline and reactive-corticosterone levels in offspring (Maccari, Krugers, Morley-Fletcher, Szyf, & Brunton, Reference Maccari, Krugers, Morley-Fletcher, Szyf and Brunton2014). In summary, then, diverse approaches to measuring prenatal stress, ranging from maternal psychological distress to maternal cortisol levels, highlight its effects on children's emotional and physiological dysregulation postnatally.
Prenatal programming of postnatal plasticity
Consideration of both sets of evidence summarized through this point, one indicating that prenatal stress is associated with elevated behavioral and physiological dysregulation and the other that such phenotypic functioning is associated with heightened susceptibility to positive and negative environmental influences, raises the intriguing hypothesis first advanced by Pluess and Belsky (Reference Pluess and Belsky2011) that prenatal stress fosters, promotes or “programs” postnatal developmental plasticity. If true, this hypothesis could account for many of the adverse, later developing phenotypes long associated with prenatal-stress exposure, including behavioral problems and academic difficulties: perhaps the reason that prenatal stress is associated with problematic functioning in childhood and adolescence in observational research is because the very forces that engendered stress in pregnancy (e.g., poverty, unemployment, marital conflict, and maternal depression) continue postnatally for many children whose prenatal experience fostered heightened developmental plasticity. Thus, when these children are exposed, postnatally, to conditions of adversity that persist beyond pregnancy, they prove especially susceptible to their influence.
Notably, the same prenatal-programming process could also account for why beneficial effects of prenatal stress have sometimes been detected in studies of well-resourced families. Consider in this regard DiPietro, Novak, Costigan, Atella, and Reusing's (Reference DiPietro, Novak, Costigan, Atella and Reusing2006) work showing that prenatal stress, measured via maternal psychological distress during middle pregnancy, predicted better infant mental scores in a well-educated, mostly white and married sample. Quite conceivably, the prenatally stressed infants who postnatally encountered supportive rearing environments proved especially sensitive and responsive to the psychological and behavioral “nutrients” available to them and thus disproportionately flourished due to their prenatally induced and enhanced developmental plasticity. In summary, would-be prenatal-stress effects may not so much be directly the result of the prenatal experience but rather reflect the enhanced influence of the postnatal environment on children especially susceptible to both supportive and unsupportive developmental experiences and environmental exposures.
When Pluess and Belsky (Reference Pluess and Belsky2011) first postulated their prenatal programming of postnatal plasticity hypothesis, based on the two independent literatures highlighted in the opening paragraphs of this paper, they provided accompanying empirical evidence to support their claims. One relevant investigation relied on data from the NICHD Study of Early Child Care (NICHD Early Child Care Research Network, 2005) and linked prenatal stress, indexed via low birth weight, to infant negative emotionality, which, in turn, was associated with infants being more susceptible to “for better and for worse” parenting effects on behavioral and cognitive functioning (Pluess & Belsky, Reference Pluess and Belsky2011). More recently, longitudinal work by Sharp, Hill, Hellier, and Pickles (Reference Sharp, Hill, Hellier and Pickles2015) revealed that maternal prenatal anxiety, measured during late pregnancy, increased children's developmental responsiveness to postnatal maternal stroking during the first few weeks of life with regard to later anxious/depressive symptoms. In this case, children exposed to high levels of prenatal anxiety evinced greater anxious/depressive symptoms when they experienced limited maternal stroking postnatally, yet very little symptomology when exposed to a great deal of maternal stroking, an effect found to be especially pronounced in girls. The same was not true of children whose mothers experienced little anxiety during pregnancy. In both cited works, regression slopes linking the environmental-exposure predictor with the measured outcome revealed that those exposed to high levels of prenatal stress manifest both the highest and lowest levels of all study members of the outcomes measured.
Further evidence of prenatal programming of postnatal plasticity comes from research comparing preterm and full-term babies. There is a substantial body of work showing psychosocial stress to be an etiological risk factor for preterm birth (Shapiro, Fraser, Frasch, & Seguin, Reference Shapiro, Fraser, Frasch and Séguin2013), even when controlling for other well-known risk factors (e.g., twin pregnancy, tobacco use, infection, and premature contractions; Lilliecreutz, Laren, Sydsjo, & Josefsson, Reference Lilliecreutz, Larén, Sydsjö and Josefsson2016). Thus, preterm birth can be considered a marker of prenatal stress. Pertinent to the issue of prenatal programming of postnatal plasticity, then, is an investigation that examined the differential effects of the caregiving environment on infant cognitive and social functioning in preterm and full-term infants (Gueron-Sela, Atzaba-Poria, Meiri, & Marks, Reference Gueron-Sela, Atzaba-Poria, Meiri and Marks2015). Results revealed that preterm infants were more developmentally responsive to their caregiving environment, evincing the greatest social and cognitive functioning when exposed to a high-quality caregiving environment but the lowest social and cognitive functioning when they experienced a low-quality caregiving environment. Caregiving quality did not, however, predict social and cognitive development in the case of full-term infants. These findings are in line with those of earlier work that chronicled stronger associations between maternal responsiveness and cognitive growth in the case of preterm infants than full-term ones (Landry, Smith, Swank, Assel, & Vellet, Reference Landry, Smith, Swank, Assel and Vellet2001). An intervention designed to promote maternal responsiveness proved successful in doing so, but when it came to effects on children's development, the benefits of being in the experimental group rather than the control group proved greater in the case of children born preterm rather than full-term (Landry, Smith, & Swank, Reference Landry, Smith and Swank2006).
Beyond observational evidence of prenatal programming of postnatal plasticity
Even if all the findings reviewed through this point appear consistent with the claim that prenatal stress promotes enhanced susceptibility to postnatal experiences, via heightened negative emotionality and physiological reactivity (or preterm birth/low birth weight), the work cited is not without limits. As already noted, observational studies in particular do not provide a basis for strong causal inference. After all, a mother could carry certain genes that increase her chances of becoming anxious or depressed during pregnancy, genes that she could pass on to her child, which, in turn, could make him or her more susceptible to postnatal environmental influences. Were that the case, we would have misinterpreted much of the evidence reviewed in discussing the claim that prenatal stress programs postnatal plasticity. One obvious scientific solution to this empirical conundrum would involve experimentally increasing the stress of pregnant women in order to determine if this affects infant emotional and/or physiological reactivity. However, even if this proved to be the case, were such unethical research undertaken, there would still be the issue of differential susceptibility to postnatal environmental influences.
In circumstances such as this, one way to proceed to further the empirical evaluation of a hypothesis of interest, in this case the prenatal-programming hypothesis, is to conduct an animal experiment. This is what we proceeded to do, using prairie voles (Microtus ochrogaster) as our experimental subjects (Hartman, Freeman, Bales, & Belsky, Reference Hartman, Freeman, Bales and Belsky2018). We chose prairie voles as study animals because they display key characteristics of social monogamy and selective social behavior, including preference for a familiar partner, an emotional attachment to the pair-mate, and male care of offspring. Social attachments are a key aspect of the early environment for humans and many other mammalian species (Mason & Mendoza, Reference Mason and Mendoza1998). Other common rodent models, such as rats and mice, do not form selective social attachments (except filial attachment) as adults, whereas prairie voles, like humans, do so. Furthermore, prairie voles naturally vary, in traitlike fashion across multiple litters, in the amount of care they display toward their pups (Perkeybile, Griffin, & Bales, Reference Perkeybile, Griffin and Bales2013). Thus, prairie voles are an optimal animal to use in cross-fostering paradigms, which afford the contrasting effect of more and less supportive parenting, when testing hypotheses based on findings from human studies.
Our study design involved, in its first stage, assigning pregnant voles on a random basis to a social-stress or a no stress condition during the last week of pregnancy. Those assigned to the experimental group were exposed to an unfamiliar and lactating (hence, aggressive) female vole for 10 min/day for 5 consecutive days, using a plexiglass divider to keep the animals separate (and physically unharmed). This paradigm is known to increase stress reactivity in offspring, both behaviorally and physiologically (Brunton & Russell, Reference Brunton and Russell2010). Those in the control condition were left undisturbed.
The second stage of our investigation occurred postnatally when the offspring born to both experimental and control mothers were cross fostered, again on a random basis, to either high- or low-quality (unrelated) rearing parents. We felt confident in characterizing the two groups of parents this way because we utilized a previously established method of quantification that has been shown to be effective in distinguishing high- and low-quality parents in prairie voles (Perkeybile et al., Reference Perkeybile, Griffin and Bales2013). Specifically, we recorded parenting behaviors (e.g., nursing, contact, licking, and grooming) before the start of the experiment to quantify each pair's natural level of parenting (Perkeybile et al., Reference Perkeybile, Griffin and Bales2013). These parenting scores were summed and the top-ranked quartile became the high-quality parental group, and the bottom quartile the low-quality parental group, in the cross-fostering phase of our experiment.
In sum, the research we undertook used a 2 (Prenatal Stress: Yes vs. No) × 2 (Postnatal Rearing: High vs. Low quality) research design. Based on everything stipulated through this point, we predicted that large differences would emerge in the development of the prenatally stressed voles reared under high- and low-quality conditions due to their heightened susceptibility to rearing effects but that the same would not be true of those voles not exposed to stress prenatally. Moreover, we hypothesized that group differences would take the “for better and for worse,” differential-susceptibility-related form: the prenatally stressed voles would score highest and lowest of all four groups of voles on the outcome variables measured (see next paragraph), with the scores of the unstressed voles falling in between.
For the most part, results of our experiment proved consistent with the prenatal programming of postnatal plasticity hypothesis. That is, prenatally stressed voles were more developmentally responsive to the rearing environment than voles not prenatally stressed. Specifically, voles cross-fostered to high-quality rearing environments displayed, as adults, the least behavioral and physiological reactivity when subjected to a stressor (i.e., forced swim), but the most if they were exposed by low-quality rearing environments. In the case of voles in the control condition that were not prenatally stressed, rearing environmental quality exerted no effect whatsoever on later reactivity. In an attempt to illuminate brain processes that might mediate the effects of prenatal stress on postnatal plasticity, we discovered that voles prenatally stressed and cross-fostered to high-quality rearing environments had the most vasopressin 1a receptor density in the amygdala. We chose this potential mediating factor to study because it was previously shown to be related to anxiety behavior and social functioning (Carter, Grippo, Pournajafi-Nazarloo, Ruscio, & Porges, Reference Carter, Grippo, Pournajafi-Nazarloo, Ruscio and Porges2008).
Future Research Directions
The fact that our experimental animal study generated results strikingly consistent with what has been found in human research provides strong evidence that prenatal stress programs postnatal plasticity, at least in voles. Even so, research to date documenting the potential beneficial effects of prenatal stress, when matched with a supportive postnatal environment, remains limited. This dearth of research is likely due to the almost exclusive focus on the adverse effects of prenatal stress with little consideration of postnatal experiences, and this itself is due to the fact that even when the interaction of prenatal and postnatal environments is considered, it is usually examined in terms of the “risk and resilience” or diathesis–stress framework (Zuckerman, Reference Zuckerman1999). This results in an exclusive focus on pathological outcomes (e.g., anxiety, depression, cognitive disorders, and poor health), which leaves little opportunity to illuminate the (postnatal) conditions under which prenatal stress may actually promote more rather than less competent development. Clearly, further research should consider the interaction between the quality prenatal environment and postnatal environment on outcomes that can range from positive (i.e., high functioning) to negative (i.e., low functioning).
Having said that, there are many other ways that future inquiry could seek to illuminate the prenatal programming of postnatal plasticity. In what follows, we consider first a variety of study designs with humans that could be used to determine the effects of prenatal stress on susceptibility to postnatal environmental influences. Thereafter, we turn attention to potential mechanisms instantiating postnatal plasticity via prenatal stress, as these too merit future attention. Finally, we entertain the prospect that some individuals may be more susceptible than others to prenatal-stress effects in hopes of encouraging future work on moderators of the enhanced-plasticity programming process under consideration.
Human research designs
As described previously, a major limitation of prenatal-stress research is genetic similarity of mother and fetus, which confounds prenatal-stress effects with genotypic effects. Fortunately, one may address this limitation using different study designs, some of which include adoption, gestational cross-fostering, interventions, and natural experiments.
Utilizing adoption studies is a potentially fruitful avenue of research considering that the prenatal environment would be unrelated to the postnatal one, much akin to cross-fostering experiments in animals (presuming adoptive and biological mother are themselves unrelated). Such research would, of course, necessitate gathering measurements of the stress the biological mother experienced during pregnancy, which may present formidable challenges. Nevertheless, by using adoption studies, one could effectively disassociate the prenatal effects from the postnatal ones while controlling for genetic influence. This empirical direction would seem to be especially worth pursuing because pregnant mothers who place children for adoption may be under a greater amount of distress than the average population, potentially leading their infants to being especially developmentally plastic. However, another potential challenge that may be encountered with such adoption research is that children often experience several caregiving settings (e.g., multiple foster homes and institutional care) prior to a stable placement (Rubin, O'Reilly, Luan, & Localio, Reference Rubin, O'Reilly, Luan and Localio2007). Hence, these children may not only be exposed to several rearing conditions that may vary in quality but also experience these contexts at different time periods, which may, in turn, be more or less influential in programming their development (i.e., timing effects). Therefore, investigators pursuing this line of research should explore how both the differing quality and the timing of exposure to these various settings may influence children's developmental trajectories.
Similar to adoption studies, a gestational cross-fostering research design innovated and employed by Rice et al. (Reference Rice, Harold, Boivin, Hay, van Den Bree and Thapar2009) may also help disentangle the effects of genetics and the prenatal environment. Specifically, Rice et al. (Reference Rice, Harold, Boivin, Hay, van Den Bree and Thapar2009) studied prenatal effects on child development by examining mothers who were either biologically related or unrelated to their child as a product of in vitro fertilization. By comparing these pairs, Rice et al. (Reference Rice, Harold, Boivin, Hay, van Den Bree and Thapar2009) were able to determine the influence of the prenatal environment independent of genetic continuity. Future work may also utilize this novel design in order to distinguish prenatal stress effects from genetic ones.
Another desiderata of future research should be to determine whether effective treatments for prenatal anxiety and depression (essentially experiments that downregulate prenatal stress) reduce infant's susceptibility to postnatal environmental influences. It is quite conceivable that any random control trials seeking to reduce prenatal stress may already provide experimental evidence as to whether the postnatal rearing environments of women randomized to control/no-treatment conditions actually exert more influence (or at least predictive power) than those of women successfully treated for their anxiety and depression prenatally. We hypothesize that the association between postnatal experiences (e.g., parenting quality) and child development would be weaker for experimental mothers who received (and responded positively to) stress-reducing treatment during pregnancy and stronger for the control group whose stress was not downregulated. It should be noted, however, that interventions aimed at reducing stress during pregnancy may also affect the postnatal environment. For example, an intervention designed to reduce anxiety during pregnancy by providing the mother with coping skills and/or emotion regulation strategies may very well influence mother–child interactions postnatally. Thus, one would need to account for any intervention effects on measurements of postnatal environmental quality when interpreting the effects of prenatal stress interventions on child susceptibility.
Natural experiments, including exposure to natural disasters, might also afford insight into prenatal-stress effects on postnatal plasticity due to their random nature. Specifically, one benefit of utilizing these types of investigations is that the stressor is an objective hardship, in which duration and intensity can be measured, that is randomly distributed in the population. Hence, experiencing a natural disaster is independent of the mother's personality, behavior, and genetic predisposition, unlike other forms of stressors such as interpersonal conflict (e.g., Jaffee & Price, Reference Jaffee and Price2007). This type of work could, potentially, further illuminate prenatal-stress effects by reducing the amount of maternal confounding factors. Having said this, investigators would be wise to entertain the possibility that some mothers may be more sensitive to the adverse experience of a natural disaster than others (i.e., differential susceptibility).
Proposed mechanisms of plasticity
In turning to consider candidate biological mechanisms potentially instantiating developmental plasticity resulting from prenatal stress, we draw heavily, even if not exclusively, on ideas advanced by Boyce and Ellis (Reference Boyce and Ellis2005) and Moore and Depue (Reference Moore and Depue2016). Given the ubiquitous effects of prenatal stress and thus numerous possible mechanisms, we should make clear that we will be limited in our focus. While acknowledging that prenatal stress has significant effects on neural activation and connectivity (e.g., Buss, Davis, Muftuler, Head, & Sandman, Reference Buss, Davis, Muftuler, Head and Sandman2010), epigenetic machinery (e.g., noncoding RNAs and DNA methyltransferases; Cruceanu, Matosin, & Binder, Reference Cruceanu, Matosin and Binder2017), and inflammation processes (e.g., Coussons-Read, Okun, & Nettles Reference Coussons-Read, Okun and Nettles2007), all of which could be potential biological mechanisms, these will not be considered in detail in this report.
Physiological reactivity
As described previously, heightened reactivity of the HPA system is the key mechanism proposed by Boyce and Ellis (Reference Boyce and Ellis2005) responsible for enhanced environmental sensitivity. Recall, also, that increased physiological reactivity has consistently been linked to prenatal-stress exposure. Thus, it would follow that prenatal stress would foster greater physiological reactivity and, thereby, increased developmental plasticity (i.e., prenatal stress → greater physiological reactivity → increased plasticity). Although portions of this process have been studied in isolation, the entirety of this potential mechanistic pathway has yet to be evaluated empirically. In addition, by examining this candidate pathway, we are likely to identify additional biological processes that contribute to the instantiation of environmental sensitivity (e.g., epigenetics and neural connectivity).
Consider in this regard the growing interest in the mediating role of epigenetics with respect to effects of prenatal stress on physiological reactivity. Most epigenetic studies have focused on the programming effects of early postnatal life with the seminal study by McGowan et al. (Reference McGowan, Suderman, Sasaki, Huang, Hallett, Meaney and Szyf2011) showing that, in rats, early postnatal stress influences hippocampal DNA methylation in the promoter region of NR3C1, the gene coding the glucocorticoid receptor, which regulates the stress response. Research in both humans and animals suggests that prenatal stress may induce the same epigenetic modifications in homologous promoter regions of NR3C1. For example, Mueller and Bale (Reference Mueller and Bale2008) found that, in mice, prenatal stress increased stress reactivity and hypothalamic methylation in the promoter region of NR3C1. Several human studies using neonatal cord blood have found that prenatal anxiety (Hompes et al., Reference Hompes, Izzi, Gellens, Morreels, Fieuws, Pexsters and Verhaeghe2013), maternal exposure to interpartner violence (Radtke et al., Reference Radtke, Ruf, Gunter, Dohrmann, Schauer, Meyer and Elbert2011), and depressive symptoms (Conradt, Lester, Appleton, Armstrong, & Marsit, Reference Conradt, Lester, Appleton, Armstrong and Marsit2013; Oberlander et al., Reference Oberlander, Weinberg, Papsdorf, Grunau, Misri and Devlin2008), all indisputable markers of prenatal stress, are associated with differential methylation patterns in the promoter region of NR3C1. One recent investigation examining pregnant mothers exposed to chronic stress in Democratic Republic of Congo showed that infants had differential methylation patterns across several genes (i.e., CRH, CRHBP, NR3C1, and FKBP5) shown to regulate the HPA axis (Kertes et al., Reference Kertes, Kamin, Hughes, Rodney, Bhatt and Mulligan2016). These methylation patterns were associated with infant birth weight.
Even if most epigenetic work has focused primarily on methylation of the candidate gene NR3C1, prenatal-stress effects on stress reactivity undoubtedly involve a cascade of multiple genetic, endocrine, and epigenetic factors. Thus, even if less well studied than the HPA system, it should be appreciated that the sympathoadrenomedullary (SAM) system is another crucial component of the stress response, one involved in the release of catecholamines such as norepinephrine (NE) and epinephrine (E). Even if most catecholamines are metabolized by enzymes in the placenta, results of several studies suggest that reduced amounts are still transferred from mother to fetus; moreover, fetuses can produce their own catecholamines in response to maternal stress (for a review, see Merlot, Couret, & Otten, Reference Merlot, Couret and Otten2008). Although the effect of prenatal stress on fetal exposure to catecholamines and later postnatal development remains unclear, one investigation did find that maternal E and NE levels during pregnancy predicted infant soothability, and thus negative emotionality, thereby raising the possibility that maternal activation of the SAM system may be linked to postnatal plasticity (Wroble-Biglan, Dietz, & Pienkosky, Reference Wroble-Biglan, Dietz and Pienkosky2009).
However limited the research thus far, NE was highlighted by Moore and Depue (Reference Moore and Depue2016) as a key regulator of environmental reactivity. These scholars hypothesized that high levels of NE would modulate environmental effects, “for better or for worse.” Specifically, high levels of NE under stressful conditions would produce hypervigilance and impaired cognition whereas higher levels of NE under supportive circumstances could yield ideal levels of attention to facilitate exploration and ability to take advantage of opportunities in the environment. Therefore, NE could have an important role in regulating developmental plasticity.
In sum, prenatal stress appears to have significant effect on programming the HPA system, via epigenetic mechanisms and, potentially, the SAM system. Relatedly, prenatal stress is known to affect brain areas such as the prefrontal cortex, hippocampus, and amygdala that also regulate the HPA axis (Lupien, McEwen, Gunnar, & Heim, Reference Lupien, McEwen, Gunnar and Heim2009). While outside the scope of this review, it is likely that the functioning and connectivity of these regions has a major role in developmental plasticity (Moore & Depue, Reference Moore and Depue2016). Therefore, their relation to prenatal stress should be explored further.
Serotonin
Although prenatal stress involves a cascade of complex and diverse endocrine actions, serotonin may be of particular importance when considering programming effects. Serotonin, a neurotransmitter that is widely distributed throughout the brain, is crucial for neuronal development early in life, operating in two major ways: as a growth factor regulating development of neural systems (Whitaker-Azmitia, Druse, Walker, & Lauder, Reference Whitaker-Azmitia, Druse, Walker and Lauder1996) and as a trophic factor regulating synaptogenesis and dendritic pruning (Gaspar, Cases, & Maroteaux, Reference Gaspar, Cases and Maroteaux2003). It seems likely, therefore, that if prenatal stress affected these processes during fetal development, then the developing child could be influenced in lasting ways. After all, serotonin activity is known to play a role in regulating, perhaps most notably, stress reactivity later in life (Canli & Lesch, Reference Canli and Lesch2007).
As it turns out, there is ample evidence in animal studies that prenatal stress produces lasting alterations in the serotonin system (Miyagawa et al., Reference Miyagawa, Tsuji, Fujimori, Saito and Takeda2011; Mueller & Bale, Reference Mueller and Bale2008; van den Hove et al., Reference van den Hove, Lauder, Scheepens, Prickaerts, Blanco and Steinbusch2006). For example, prenatally stressed mice evince lower serotonin transporter levels and a depressive-like phenotype (Mueller & Bale, Reference Mueller and Bale2008). In humans, increased maternal depressive mood during the second trimester of pregnancy is associated with reduced methylation in the promoter region of maternal and infant SLC6A4, the locus of the serotonin gene that codes for the serotonin transporter (Devlin, Brain, Austin, & Oberlander, Reference Devlin, Brain, Austin and Oberlander2010). Thus, it appears that prenatal stress exerts programming effects on the serotonin system, which is not surprising given the evidence that the HPA and serotonin systems are cross-regulated (see St.-Pierre, Laurent, King, & Vaillancourt, Reference St.-Pierre, Laurent, King and Vaillancourt2016, for review).
In addition to prenatal-stress effects, the serotonin system has been linked to variation in developmental plasticity. The serotonin transporter linked polymorphic region (5-HTTLPR) of SLC6A4 is one of the most well-studied genetic polymorphisms found to be associated with individual differences in susceptibility to environmental influences. Consider in this regard that individuals carrying one or more short alleles evince “for better or for worse” plasticity when the rearing predictor and child outcome are, respectively, maternal responsiveness and moral internalization (Kochanska, Kim, Barry, & Philibert, Reference Kochanska, Kim, Barry and Philibert2011); child maltreatment and antisocial behavior (Cicchetti, Rogosch, & Thibodeau, Reference Cicchetti, Rogosch and Thibodeau2012); stressful life events and preschool-onset depression (Bogdan, Agrawal, Gaffrey, Tillman, & Luby, Reference Bogdan, Agrawal, Gaffrey, Tillman and Luby2014); and supportive parenting and positive affect (Hankin et al., Reference Hankin, Nederhof, Oppenheimer, Jenness, Young, Abela and Oldehinkel2011). Just as significantly, 5-HTTLPR short alleles have been linked to greater negative emotionality and physiological reactivity, outcomes associated with prenatal stress as previously reviewed, in both humans and nonhuman primates (e.g., Champoux et al., Reference Champoux, Bennett, Shannon, Higley, Lesch and Suomi2002; Lakatos et al., Reference Lakatos, Nemoda, Birkas, Ronai, Kovacs, Ney and Gervai2003).
Given evidence that prenatal stress produces alterations in the serotonin system and that the serotonin system appears to be systematically related to variation in developmental plasticity, it stands to reason that serotonin should be a key mechanism for instantiating prenatal-programming effects. Investigators examining use of selective serotonin reuptake inhibitors on women during pregnancy may thus want to consider effects on offspring susceptibility to environmental influences.
Oxytocin and vasopressin
Oxytocin (OT) and arginine vasopressin (AVP) are two closely related nonapeptides thought to influence, among other things, the regulation of social behavior (e.g., attachment, affiliation, social dysfunction; Carter, Reference Carter2014; Carter et al., Reference Carter, Grippo, Pournajafi-Nazarloo, Ruscio and Porges2008). In addition, OT and AVP play a critical role in regulating the HPA axis. Specifically, OT can attenuate the stress response by downregulating the sympathetic nervous system (Carter, Reference Carter2014), while AVP mRNA expression plays a critical role in regulating anxious and depressive behaviors (Wigger et al., Reference Wigger, Sánchez, Mathys, Ebner, Frank, Liu and Landgraf2004).
OT, though not AVP, was highlighted by Moore and Depue (Reference Moore and Depue2016) as a mechanism for instantiating environmental responsivity. Some evidence indicates that single nucleotide polymorphisms in the OT receptor gene (OTR) moderate environmental effects in a differential-susceptibility-like fashion. Specifically, single nucleotide polymorphisms in OTR moderate effects of perceived threat on charitable behavior (Poulin, Holman, & Buffone, Reference Poulin, Holman and Buffone2012); socioeconomic status on obesity risk (Bush et al., Reference Bush, Allison, Miller, Deardorff, Adler and Boyce2017); alcohol use on aggressive behavior in men (Johansson et al., Reference Johansson, Bergman, Corander, Waldman, Karrani, Salo and Westberg2012); supportive parenting on adolescent social anxiety (Olofsdotter, Åslund, Furmark, Comasco, & Nilsson, Reference Olofsdotter, Åslund, Furmark, Comasco and Nilsson2017); and harsh parenting on young adult allostatic load (Brody, Miller, Yu, Beach, & Chen, Reference Brody, Miller, Yu, Beach and Chen2016).
Even though variations in OT have been primarily studied with respect to effects of maternal care and other early postnatal experiences and exposures, there is some evidence to suggest that it may also be subject to prenatal programming. A study by Unternaehrer et al. (Reference Unternaehrer, Bolten, Nast, Staehli, Meyer, Dempster and Meinlschmidt2016) found that maternal cortisol during the second trimester predicted greater OT-receptor methylation in neonatal cord blood. In rats, the negative effects of prenatal stress on social behavior were found to be reversed by OT administration (Lee, Brady, Shapiro, Dorsa, & Koenig, Reference Lee, Brady, Shapiro, Dorsa and Koenig2007). However, given its critical role in quality of early maternal care, it will be imperative for future research to distinguish the effects of the prenatal versus postnatal environment on differences in OT.
As compared to OT, AVP has received far less empirical attention with respect to either early life effects or variation in susceptibility to environmental influences. Nevertheless, there is reason to believe that AVP has a central, and perhaps even greater, role than OT when it comes to prenatal-programming effects. Consider in this regard the aforementioned vole study by Hartman et al. (Reference Hartman, Freeman, Bales and Belsky2018); it found that vasopressin 1a receptor density in the amygdala helped account for the effect of high-quality rearing in the case of prenatally stressed animals. Although OT-receptor binding was also examined as a possible mediator of such prenatal-stress effects, no evidence for such a role emerged.
Further evidence of the special significance of vasopressin relative to OT is research indicating (a) that effects of prenatal stress on social memory in rats is mediated by vasopressin 1a receptor mRNA expression but not OT receptors (Grundwald, Benítez, & Brunton, Reference Grundwald, Benítez and Brunton2016) and (b) that prenatal exposure to AVP or caffeine, but not OT, alters learning in female rats (Swenson, Beckwith, Lamberty, Krebs, & Tinius, Reference Swenson, Beckwith, Lamberty, Krebs and Tinius1990). Of significance also is that whereas OT is first detected a few days following birth, AVP can be detected in the prenatal and perinatal periods in the fetal brain and is thought to play a significant role in central nervous system maturation (Bloch et al., Reference Bloch, Guitteny, Chouham, Mougin, Roget and Teoule1990; Tribollet, Goumaz, Raggenbass, Dubois-Dauphin, & Dreifuss, Reference Tribollet, Goumaz, Raggenbass, Dubois-Dauphin and Dreifuss1991). Thus, alterations in AVP may be a prime target of inquiry in investigations of mechanisms instantiating prenatal programming of postnatal plasticity.
With respect to susceptibility to environmental effects, there is extremely limited work investigating whether variations in AVP are associated with differences in susceptibility. However, data have indicated the relevance of the AVPR1A polymorphism, the gene coding for vasopressin 1a receptor, on human behavior, with studies documenting main effects of AVPR1A variants on autism (Kim et al., Reference Kim, Young, Gonen, Veenstra-vanderWeele, Courchesne, Courchesne and Insel2002), age of first sexual intercourse (Prichard, Mackinnon, Jorm, & Easteal, Reference Prichard, Mackinnon, Jorm and Easteal2007), and pair-bonding behavior in men (Walum et al., Reference Walum, Westberg, Henningsson, Neiderhiser, Reiss, Igl and Lichtenstein2008). Furthermore, there is some evidence to suggest that variation in the AVPR1A is related to differences in environmental sensitivity. At a neurological level, AVPR1A variants differentially predict amygdala reactivity to faces (Meyer-Lindenberg et al., Reference Meyer-Lindenberg, Kolachana, Gold, Olsh, Nicodemus, Mattay and Weinberger2009). In addition, a study by Poulin et al. (Reference Poulin, Holman and Buffone2012) found that the AVPR1A polymorphism interacted with perceived threat to predict commitment to civic duty in a “for better and for worse,” differential-susceptibility-related manner. Specifically, individuals who carried the short/long genotype had the highest commitment to civic duty under low perceived threat but the lowest commitment under high perceived threat conditions. For other genotypes, there was no association between perceived threat and civic commitment. Likewise, research by Tabak et al. (Reference Tabak, Meyer, Castle, Dutcher, Irwin, Han and Eisenberger2015) showed that administration of intranasal AVP, but not OT, increased empathic concern but only if individuals were exposed to high levels of childhood paternal warmth. There was no association between intranasal AVP and empathetic concern under conditions of low paternal warmth; thus, this study documented variation in sensitivity to the positive environment only, a phenomenon referred to as vantage sensitivity (Pluess & Belsky, Reference Pluess and Belsky2013). In sum, research indicates that the AVP system is sensitive to prenatal effects and appears to be linked to human social behavior and environmental sensitivity. Future work should consider variations in AVP as a candidate mechanism by which prenatal stress may instantiate postnatal plasticity.
Worth considering as well is that prenatal stress effects on AVP may be mediated through increases in fetal androgen exposure. The vasopressin system is sexually dimorphic and highly steroid responsive. For example, in rats, castration results in a significant decrease of vasopressin expression while testosterone replacement ameliorates such effects (Devries, Buijs, van Leeuwen, Caffe, & Swaab, Reference DeVries, Buijs, van Leeuwen, Caffe and Swaab1985). In humans, prenatal stress is tied to higher fetal cortisol, and unlike adults, fetal cortisol and testosterone are positively correlated (Gitau, Adams, Fisk, & Glover, Reference Gitau, Adams, Fisk and Glover2005). Likewise, multiple studies document effects of prenatal stress on masculinization of brain and behavior, especially in females (e.g., Anderson, Rhees, & Fleming, Reference Anderson, Rhees and Fleming1985). Findings such as these led Del Giudice et al. (Reference Del Giudice, Barrett, Belsky, Hartman, Martel, Sangenstedt and Kuzawa2018) to hypothesize that fetal androgen exposure may increase developmental plasticity, a proposition that also seems worthy of empirical attention.
Role of the placenta
Recent investigations of prenatal programming have begun to explore the role of the placenta as a key mediator of prenatal-stress effects on fetal development. The placenta is an organ that serves as the interface between mother and fetus and can quickly adapt to changes from the maternal environment (e.g., prenatal stress). The role of the placenta is well known in actively modulating vital functions of the fetus, such as nutrient and oxygen exchange (Jansson & Powell, Reference Jansson and Powell2007), but also plays a pivotal role in the production and modulation of glucocorticoids and amines.
In particular, the placenta affects HPA axis regulation in both the mother and fetus. Specifically, the placenta produces corticotropin-releasing hormone in response to cortisol, which modulates the maternal HPA axis in a positive loop. In addition, the placenta plays a protective role against maternal cortisol by inactivating it using the placental barrier enzyme 11β-hydroxysteroid dehydrogenase Type II (11β-HSD2). This results in only 10%–20% of the cortisol from maternal circulation reaching the fetus (Gitau, Cameron, Fisk, & Glover, Reference Gitau, Cameron, Fisk and Glover1998). In rats, prenatal stress induced by restraint stress not only increased maternal cortisol but also was linked to a reduction in the expression and activity of the placental 11β-HSD2 (Peña, Monk, & Champagne, Reference Peña, Monk and Champagne2012). In turn, these epigenetic changes in placental 11β-HSD2 were themselves related to DNA methylation in the fetal brain as well as increases in fetal corticosterone levels (Peña et al., Reference Peña, Monk and Champagne2012). Furthermore, in humans, greater maternal anxiety measured 1 day prior to birth predicted lower gene expression of placental 11β-HSD2 (O'Donnell et al., Reference O'Donnell, Jensen, Freeman, Khalife, O'Connor and Glover2012) and decreased activity of placental 11β-HSD2 is associated with early development, including fetal growth restriction (Börzsönyi et al., Reference Börzsönyi, Demendi, Pajor, Rigó, Marosi, Ágota and Joó2012), prematurity (Demendi et al., Reference Demendi, Börzsönyi, Pajor, Rigó, Nagy, Szentpéteri and Joó2012), and low birthweight (Green et al., Reference Green, Babineau, Jolicoeur-Martineau, Bouvette-Turcot, Minde, Sassi and Steiner2017).
Considered together, it appears that prenatal stress may alter the transplacental barrier via epigenetic changes in 11β-HSD2, thereby resulting in increased fetal exposure to maternal cortisol, with consequences for phenotypic outcomes. Evidence to such an effect comes from a study by Glover, Bergman, Sarkar, and O'Connor (Reference Glover, Bergman, Sarkar and O'Connor2009), which examined women at various stages of their pregnancy ranging from early to late. They found that the correlation between maternal and amniotic fluid cortisol levels was greater in women with elevated anxiety compared to less anxious women (Glover et al., Reference Glover, Bergman, Sarkar and O'Connor2009). Similarly, prenatal stress may increase placental permeability, and thus fetal exposure, to other hormones. In humans, prenatal stress indexed by maternal psychological distress during late pregnancy has been associated with increased levels of serotonin and NE transporters as well as a downregulation of monoamine oxidase in placental cells, which would lead to increased intrauterine availability of these hormones (Blakeley, Capron, Jensen, O'Donnell, & Glover, Reference Blakeley, Capron, Jensen, O'Donnell and Glover2013; Ponder et al., Reference Ponder, Salisbury, McGonnigal, Laliberte, Lester and Padbury2011). Thus, a major mechanism by which prenatal stress may affect the fetus is through alterations to the placental barrier, which increase fetal exposure to select hormones (Aye & Keelan, Reference Aye and Keelan2013; Seckl & Holmes, Reference Seckl and Holmes2007). Of special significance to this paper, these changes in the placenta have been tied to aspects of infant temperament, with higher levels of placental mRNA in serotonin and glucocorticoids being associated with greater behavioral dysregulations in infants (Räikkönen et al., Reference Räikkönen, Pesonen, O'reilly, Tuovinen, Lahti, Kajantie and Reynolds2015).
Overall, then, the work cited suggests that prenatal stress may increase fetal sensitivity to maternal influences via greater placental permeability. Consequently, one might begin to consider whether all placentas are equally reactive to fluctuations in maternal physiology or whether there might be differences in how sensitive the placenta is, thereby moderating maternal effects on the fetus. One might imagine that some placentas may be very sensitive to changes in maternal physiology, such as greater stress, thus rapidly adjusting accordingly, whereas other placentas may be more resilient and need stronger or more consistent maternal signals to respond. This would have consequences for the fetus with some being more protected than others from the placental changes induced by prenatal stress.
As it turns out, placentas do appear to differ in their sensitivity to maternal signals. One recent study of rats revealed that the placental response of 11β-HSD2 to prenatal stress in the form of social and restraint stress administered daily throughout pregnancy depends on the genetic makeup of the mother (Lucassen et al., Reference Lucassen, Bosch, Jousma, Krömer, Andrew, Seckl and Neumann2009). Specifically, rats selectively bred for high anxiety and exposed to prenatal stress showed a greater reduction in placental 11β-HSD2 compared to their low-anxiety counterparts. Furthermore, there is variation in the placental response to stress based on the sex of the fetus. For instance, in response to prenatal stress, placentas of male fetuses tend to become insensitive to glucocorticoid levels, with females remaining sensitive (reviewed in St.-Pierre et al., Reference St.-Pierre, Laurent, King and Vaillancourt2016). This observation suggests that males and females have opposing adaptions to prenatal stress, with males increasing growth at the risk of decreased survival while females experience reduced growth to promote survival (St.-Pierre et al., Reference St.-Pierre, Laurent, King and Vaillancourt2016).
Given this emerging research, future studies should investigate the role of the placenta in prenatal programming of postnatal plasticity. It is clear that the placenta plays a major role in transmitting maternal signals, including stress, to the fetus. It may be the case, as already suggested, that some placentas are more responsive to maternal stress than others, which may either attenuate or amplify the effects of prenatal stress.
Intestinal microbiota
An additional way that prenatal stress may affect a child's susceptibility to environmental influence is through the colonization of intestinal microbiota. It has become increasingly clear that intestinal microbiota influence brain development and behavior via the microbiome–gut–brain axis. For example, alterations in the microbiome have been linked to psychological disorders including depression and anxiety (see Sherwin, Rea, Dinan, & Cryan, Reference Sherwin, Rea, Dinan and Cryan2016, for a review). Specifically, animal studies have shown that germ-free and antibiotic-treated mice exhibit anxiety-like and depressive-like behavior as well as alterations in the serotonergic, neurotrophic, and HPA systems (Bercik et al., Reference Bercik, Denou, Collins, Jackson, Lu, Jury and Verdu2011). If treated with probiotics, mice showed a reduction in anxiety-like and depressive-like behavior (Bravo et al., Reference Bravo, Forsythe, Chew, Escaravage, Savignac, Dinan and Cryan2011), an effect that has been replicated in humans (Messaoudi et al., Reference Messaoudi, Lalonde, Violle, Javelot, Desor, Nejdi and Cazaubiel2011).
Particularly relevant to this report, microbiome patterns have been linked to temperament and stress physiology, two established markers of developmental plasticity. A number of animal studies indicate that the microbiome regulates activation of the HPA axis with germ-free mice and rats showing elevated stress responses (see Sherwin et al., Reference Sherwin, Rea, Dinan and Cryan2016, for a review). In human infants, microbiota patterns have been linked to negative temperament with lower diversity and stability of microbiota during the first weeks of life predicting greater crying, fussiness, and colic (de Weerth, Fuentes, Puylaert, & de Vos, Reference de Weerth, Fuentes, Puylaert and de Vos2013; Pärtty, Kalliomäki, Endo, Salminen, & Isolauri, Reference Pärtty, Kalliomäki, Endo, Salminen and Isolauri2012). Moreover, a study by Christian et al. (Reference Christian, Galley, Hade, Schoppe-Sullivan, Dush and Bailey2015) found that patterns of bacterial diversity were related to sociability and activity levels during early childhood.
A separate line of work has established prenatal stress as a predictor of infant intestinal microbiota. Take, for example, research by Bailey, Lubach, and Coe (Reference Bailey, Lubach and Coe2004) indicating that in rhesus monkeys, prenatal stress adversely affected the intestinal microbiota of offspring. Another investigation, using mice, revealed that prenatal stress predicted offspring intestinal microbiota as well as anxiety-like behavior in adults (Gur et al., Reference Gur, Shay, Palkar, Fisher, Varaljay, Dowd and Bailey2017). These findings extend to humans with both subjective reports of stress and cortisol exposure during pregnancy predicting differences in infant microbiota diversity, which, in turn, are linked to infant health (Zijlmans, Kopela, Riksen-Walraven, de Vos, & de Weerth, Reference Zijlmans, Korpela, Riksen-Walraven, de Vos and de Weerth2015).
Taken together, this literature calls attention to another potential mechanistic pathway instantiating enhanced developmental plasticity: prenatal stress affects infant intestinal microbiota, which, in turn, influences environmental susceptibility, perhaps through temperamental negativity. This suggests that there may be utility in evaluating whether intake of probiotics during infancy and early childhood is linked to reduced plasticity via easier temperament.
One important consideration to this proposition, however, is the growing literature that breastfeeding influences infant intestinal microbiota (Penders et al., Reference Penders, Thijs, Vink, Stelma, Snijders, Kummeling and Stobberingh2006). Specifically, breastfed infants at 1 month of age show a different intestinal microbiota profile than formula-fed infants even when accounting for other various factors known to affect infant intestinal microbiota (e.g., method of delivery; Penders et al., Reference Penders, Thijs, Vink, Stelma, Snijders, Kummeling and Stobberingh2006). In addition, there is evidence that prenatal maternal depression may affect whether and how long mothers chose to breastfeed and also that engaging in breastfeeding may reduce postpartum depression (Figueiredo, Canário, & Field, Reference Figueiredo, Canário and Field2014). Thus, future research aimed at identifying whether changes in infant intestinal microbiota is a mechanism by which prenatal stress influences postnatal plasticity should also consider the maternal influence of breastfeeding on infant microbiota composition and the quality of the postnatal environment.
Potential moderators of plasticity
Having highlighted candidate mechanisms that may link prenatal stress and enhanced developmental plasticity, attention is now turned to potential moderators that may enhance or reduce the effect of prenatal stress on susceptibility to postnatal environmental influences.
Genetic moderation of prenatal stress effects
As previously noted, there is evidence, in humans, that prenatal stress appears to increase postnatal plasticity. Some Gene × Environment interaction work calls attention to the possible genetic moderation of such prenatal programming. As previously noted, the serotonin transporter linked polymorphic region (5-HTTLPR) is a genetic variant that has been consistently identified as a genetic marker of plasticity with the short allele carriers showing greater variation in response to postnatal environmental exposures (Belsky & Pluess, Reference Belsky and Pluess2009, Reference Belsky and Pluess2013; van IJzendoorn, Belsky, & Bakermans-Kranenburg, Reference van IJzendoorn, Belsky and Bakermans-Kranenburg2012). Pluess et al. (Reference Pluess, Velders, Belsky, van IJzendoorn, Bakermans-Kranenburg, Jaddoe and Tiemeier2011) tested the hypothesis that it would be fetuses carrying the 5-HTTLPR short allele who, if exposed to elevated levels of maternal anxiety prenatally (measured via self-reported anxiety during middle pregnancy), would be most likely to develop negatively emotional temperaments; results proved consistent with this proposition. A study by Babineau et al. (Reference Babineau, Green, Jolicoeur-Martineau, Bouvette-Turcot, Minde, Sassi and Lydon2015) extended this work upon examining the interaction of prenatal depression and 5-HTTLPR in predicting infant and early childhood behavioral dysregulation. These investigators observed that greater prenatal depression measured during middle or late pregnancy predicted more infant and early childhood dysregulation from 3 to 36 months of age but, like Pluess et al. (Reference Pluess, Velders, Belsky, van IJzendoorn, Bakermans-Kranenburg, Jaddoe and Tiemeier2011), only for short-allele carriers of 5-HTTLPR (Babineau et al., Reference Babineau, Green, Jolicoeur-Martineau, Bouvette-Turcot, Minde, Sassi and Lydon2015). Of note, the detected genetic moderation took the form of differential susceptibility, because when children with short alleles were exposed prenatally to maternal depression, they had the highest levels of dysregulation, but when exposed to lower or little prenatal depression, they had the lowest levels. Finally, a recent inquiry by Green et al. (Reference Green, Babineau, Jolicoeur-Martineau, Bouvette-Turcot, Minde, Sassi and Steiner2017) revealed that prenatal depression measured during middle to late pregnancy interacted with a polygenic profile score that was, in part, based on 5-HTTLPR in predicting infant negative temperament. After compositing a number of “susceptibility” alleles (i.e., genetic variants shown to make individuals more environmentally responsive) from 5-HTTLPR and the dopamine-receptor D4 (DRD4) gene, results indicated that prenatal depression only predicted greater infant negative emotionality for those carrying more susceptibility genotypes.
Given this work documenting genetic moderation of prenatal-stress effects on infant temperament, we would encourage future investigators not only to expand their genetic focus beyond the two candidate genes just highlighted but also to consider maternal genotype. After all, genetic makeup of the fetus may not only moderate prenatal-stress effects on the infant, but maternal genotype might affect whether mothers differ in their stress-related responses to potentially stress-inducing experiences and exposures. It is possible, after all, that mothers with more susceptible genotypes experience greater subjective stress than do others even when exposed to the same would-be stressor. Because the fetus and mother are biologically related, it will be important to disentangle maternal-genotypic effects from fetal-genotypic effects, possibly through adoption studies, gestational cross-fostering (Rice et al., Reference Rice, Harold, Boivin, Hay, van Den Bree and Thapar2009), or in the case of animal studies, cross-fostering. It is certainly conceivable that fetal and maternal genotype could interact when it comes to prenatal stress influencing the postnatal plasticity of offspring.
Sex differences
Sex differences in response to prenatal stress are often of specific interest especially in animal studies. In rats, some commonly investigated outcomes of prenatal stress, such as anxiety-like behavior and hippocampal neuroplasticity, have been found to be sex dependent such that females display greater anxiety-like behavior while males show decreased hippocampal neuroplasticity as a result of prenatal stress (Zuena et al., Reference Zuena, Mairesse, Casolini, Cinque, Alemà, Morley-Fletcher and Nicoletti2008). Other prenatal stress outcomes, including depression-like behavior, appear to be not sex dependent (van Waes et al., Reference van Waes, Darnaudery, Marrocco, Gruber, Talavera, Mairesse and Maccari2011). In human research, it is even less clear whether and how sex interacts with prenatal stress. One inquiry, using data from pregnant mothers who were exposed to flooding, revealed that higher levels of hardship during pregnancy predicted greater infant irritability, but only for boys (Simcock et al., Reference Simcock, Elgbeili, Laplante, Kildea, Cobham, Stapleton and King2017). Other research finds girls to be more sensitive to prenatal-programming effects (e.g., Sharp et al., Reference Sharp, Hill, Hellier and Pickles2015). These mixed findings may partly be due to sex-dependent differences in the outcome of interest. For example, boys present more frequently with intellectual impairment and childhood behavioral disorders related to prenatal stress whereas girls may develop subtler, later-onset anxiety and affective disorders (Davis & Pfaff, Reference Davis and Pfaff2014). Thus, whether one discerns prenatal-stress effects for only males or only females may depend on whether one is investigating, respectively, male- or female-biased phenotypes.
It seems quite possible that there may be different biological mechanisms in males and females that are activated by prenatal stress. As stated previously, the response of the placenta due to prenatal stress appears to differ for male and female. In addition, work with rodents indicates that prenatal stress increases hedonic preferences in males but such preferences are reduced in females due to lower estrogen levels (Reynaert et al., Reference Reynaert, Marrocco, Mairesse, Lionetto, Simmaco, Deruyter and Morley-Fletcher2016). Taken together, this work suggests that sex may have a significant role in moderating prenatal stress effects; however, more research is needed to determine whether there are consistent sex differences in response to prenatal stress and, specific to this paper, whether, should that be the case, this translates into sex-based differences in postnatal plasticity resulting from prenatal stress.
Timing and type of prenatal stress
It is clear from the work reviewed herein that many different types of stressors can influence child development, including maternal anxiety and depression (O'Connor, Heron, Golding, & Glover, Reference O'Connor, Heron, Golding and Glover2003; van den Bergh, van Calster, Smits, van Huffel, & Lagae, Reference van den Bergh, van Calster, Smits, van Huffel and Lagae2008), pregnancy-specific anxiety (Huizink et al., Reference Huizink, De Medina, Mulder, Visser and Buitelaar2002), and exposure to acute disasters such as a Canadian ice storm (Laplante, Brunet, Schmitz, Ciampi, & King, Reference Laplante, Brunet, Schmitz, Ciampi and King2008) and 9/11 terrorist attacks (Yehuda et al. Reference Yehuda, Engel, Brand, Seckl, Marcus and Berkowitz2005). This diversity of stressors also extends to animal work, some of which include repeated restraint (e.g., Henry, Kabbaj, Simon, Moal, & Maccari, Reference Henry, Kabbaj, Simon, Moal and Maccari1994), electric shock (e.g., Takahashi & Kalin, Reference Takahashi and Kalin1991), chronic unpredictable stress (e.g., Mueller & Bale, Reference Mueller and Bale2008), and social stress (e.g., Brunton & Russell, Reference Brunton and Russell2010). These different types of stressors are likely to vary in intensity, duration, and predictability, all of which may result in divergent effects on the mother and fetus.
In particular, evidence indicates that the intensity of prenatal stress may matter with respect to its postnatal consequences. In the previously mentioned work of DiPietro et al. (Reference DiPietro, Novak, Costigan, Atella and Reusing2006), mild prenatal stress was found to positively affect infant motor development and cognitive ability, at least in the advantaged sample they were studying. Such results led the authors to propose a curvilinear response to prenatal stress with the greatest negative effects emerging under intense and chronically stressful conditions and the most positive effects resulting from conditions of mild to moderate stress. Of note is that this hypothesis was empirically confirmed using data on pregnant women who experienced the aforementioned Canadian ice storm (Laplante et al., Reference Laplante, Brunet, Schmitz, Ciampi and King2008). Hence, the cited theorizing and research make clear that the intensity of a stressor should be considered when seeking to understand its effects on the child. Future work should also seek to determine whether the distinctive effects of varying intensity of stress applies equally to different types of stressors (e.g., depression vs. anxiety vs. daily hassles vs. bereavement) and why that might be the case. Perhaps, different stressors may be linked to unique physiological profiles in mothers and therefore exert varying effects on their fetus. Important to note, though, is that multiple stressors frequently co-occur, thus making this research proposition somewhat difficult to address. However, other work, examining the unique influence of particular components in a stressful environment, has shown that specific experiences, even if related, may be more or less salient in directing child development (Hartman, Sung, Schlomer, Simpson, & Belsky, Reference Hartman, Sung, Simpson, Schlomer and Belsky2017).
Relatedly, the timing of prenatal stress may also be important to consider. For example, it has been suggested that perturbations early in pregnancy are likely to produce more severe neurological insults than later stressors, perhaps via effects on placental functions and neural organization (Watson & Cross, Reference Watson and Cross2005). Notable, then, is evidence that exposure to stress in the first trimester rather than later in gestation heightens the risk of schizophrenia (Khashan et al., Reference Khashan, Abel, McNamee, Pedersen, Webb, Baker and Mortensen2008). In addition, work by Davis and Sandman (Reference Davis and Sandman2010) shows that the effects of maternal cortisol on infant cognitive development is dependent on timing of exposure. Whereas higher maternal cortisol levels early in gestation predicted lower mental development scores in offspring, the very same physiological condition predicted better mental development when it occurred late in gestation. Yet the opposite seems true when it comes to prenatal-stress effects on emotional and behavioral problems during childhood (O'Connor, Heron, Golding, Beveridge, & Glover, Reference O'Connor, Heron, Golding, Beveridge and Glover2002). Clearly, it should not be assumed that when it comes to the timing of prenatal stress, effects will be most pronounced when stress occurs early in pregnancy.
Conclusion
Having cited evidence, including new experimental research, consistent with the proposal that prenatal stress programs postnatal plasticity and considered in some detail how such programming might be biologically instantiated and which children might be most susceptible to such programming effects, in concluding this report, we turn attention to the ultimate question central to evolutionary analysis: why should prenatal stress influence postnatal plasticity? Before addressing this issue we should make clear that what we offer is a post hoc argument. Unlike the notion of differential susceptibility to environmental influences, which was based on theoretical first principles rather than existing evidence (Belsky, Reference Belsky1997, Reference Belsky, Ellis and Bjorklund2005; Belsky & Pluess, Reference Belsky and Pluess2009), the basis of Pluess and Belsky's (Reference Belsky and Pluess2009), the hypothesis that prenatal stress programs postnatal plasticity was derived from consideration of two independent sets of evidence, as made clear in the opening paragraphs of this paper.
Prenatal-stress research is often framed in terms of the fetal-programming hypothesis, which stipulates that the fetus adapts its phenotype to the anticipated postnatal environment based on maternal cues regarding the quality of the extrauterine ecology (Barker, Reference Barker1998; Bateson et al., Reference Bateson, Barker, Clutton-Brock, Deb, D'udine, Foley and McNamara2004; Gluckman, Hanson, Spencer, & Bateson, Reference Gluckman, Hanson, Spencer and Bateson2005). The evolutionary bio-logic here is that such a “predictive adaptive response” (PAR) would increase the likelihood of the developing individual fitting, both biologically and behaviorally, the specific environment in which he or she will live postnatally, thereby increasing reproductive fitness (see Belsky, Reference Belsky2012; Belsky, Steinberg, & Draper, Reference Belsky, Steinberg and Draper1991, for related theorizing and evidence about PAR). According to the original formulation of the prenatal-programming hypothesis, a fetus exposed to prenatal stress should develop a “thrifty” phenotype (i.e., small body size) because it would be advantageous in a food-limited and harsh environment. After all, a larger body would require greater nutritional resources than a smaller one to remain healthy.
Although this PAR-related view seems to make intuitive sense, it would seem to disregard the fact that the future is inherently uncertain and thus the prenatal environment may not accurately map on to future postnatal conditions. Were that so, a developmental mismatch would occur, as previously noted. Hence, it may not always be beneficial to canalize development according to the intrauterine environment. Given this, we still need to address the question of why, from an evolutionary perspective, prenatal stress would foster greater responsiveness to the postnatal environment.
Insight into this issue would seem to come from considering the mother herself. One possibility, suggested by M.B. Hennessy (personal communication, November 8, 2015) is that if a woman is stressed during pregnancy, she is not fitting her environment very well. Should that be the case, it would seem advantageous for the fetus to adopt a wait-and-see approach, being especially sensitive to the postnatal environment before committing to a developmental trajectory (see Frankenhuis & Del Giudice, Reference Frankenhuis and Del Giudice2012). After all, the fetus might not necessarily “know” whether the stress experienced by the mother is state or trait dependent, meaning temporary or enduring. If it is enduring, there will always be time for the plasticity-enhanced child to regulate development in accordance with a stressful postnatal world. Recall that this was the very reason why we think that so much evidence exists linking prenatal stress with compromised development.
Now consider the case of the pregnant mother who experiences very little stress. It seems likely that this would be because she fits the environment well, has long done so, and expects to continue to do so well into the future. Under such conditions, it would seem advantageous for the fetus to canalize its development, based on maternal cues, earlier, in pregnancy, rather than later, following parturition.
In addition to this proposition, another possibility, suggested by D.W. Belsky (personal communication, January 26, 2018), is that prenatal stress decanalizes development, just as do some postnatal stressors (Burrows & Hannan, Reference Burrows and Hannan2013; Chen, Nolte, & Schlotterer, Reference Chen, Raine, Rudo-Hutt, Glenn, Soyfer and Granger2015). The evolutionary reasoning behind this is that stress conveys to the developing organism that its otherwise canalized development is not likely to enable it to succeed (in reproductive-fitness terms), and so the likelihood of reproductive success might be increased if the organism deviated from its previous canalized path. One way of doing so could be to “take instructions” from the developmental environment and thus be particularly susceptible to postnatal experiences and exposures.
In conclusion, then, we have advanced—and extended—herein the claim that prenatal stress promotes developmental plasticity by increasing susceptibility to postnatal environmental experiences and exposures (Pluess & Belsky, Reference Pluess and Belsky2011). In addition to reviewing the Pluess and Belsky (Reference Pluess and Belsky2011) proposal and citing new evidence consistent with it, we have considered how such enhanced plasticity might be instantiated, which children might be most susceptible to such prenatal-stress effects, and even why development may operate in the way we have hypothesized that it does. In so doing we have further developed a view of prenatal stress profoundly different from the prevailing one, which considers only adverse effects of such early life experience. It is our hope that the evidence and ideas advanced herein will stimulate further research by encouraging other investigators to look at the potential “upside” of prenatal stress when infants experience supportive rearing milieus postnatally. Even if more empirical support is needed, the work we have cited and the argument we have advanced has the potential to inform policy and intervention. Radically, one might even consider, should further support emerge for the view central to this report, promoting prenatal stress when there is every reason to believe that the postnatal environment will be highly supportive of developmental well-being.
At the same time, one should not lose sight of the developmental consequences of prenatal stress when the postnatal world remains highly stress inducing. Not only would the child exposed to such a “double whammy” be adversely affected by it, but because of his or her enhanced plasticity, this sequelae would be radically different from how this very child might have developed had the postnatal world proved supportive. In such cases, the human capital development cost could be huge: a child highly susceptible to postnatal developmental exposures and experiences would have its development severely compromised rather than enhanced. Just imagine a world in which such children could experience just the opposite kind of life than many no doubt will.
