Keywords

Introduction

To introduce his aggregated volume entitled Prenatal determinants of behavior, Joffe (1969, p. ix) wrote: “This book is an attempt to gather together from diverse sources the research which relates events prior to birth to effects on the postnatal behavior of organisms. Such studies are an extension of the widespread and intensive interests in the behavioral effects of events in the early environment to the organism’s earliest environment—the prenatal environment. It is hoped that bringing together studies which arose from a variety of experimental interests and appeared in a wide range of publications will stimulate further interest in a field which only now is being delineated as a research area in its own right.” When he wrote these words, one wonders whether Joffe could have foreseen how much interest the field of prenatal environmental influences was going to gain in the period that lay ahead. Joffe qualified his introductory sentences by stating that, over the centuries, there had been a fluctuation in the belief of prenatal influences and that apparently the status of the area did not depend so much on the available evidence as on the prevailing climate or opinion: “Medical opinion appears to have accepted the general proposition at times and ridiculed at others” (Joffe, 1969, p. 1). Likewise Stott (1958, p. 42) mentioned that effects of psychosomatic stress during pregnancy was “a topic which has fallen under the taboo of “old wives tales.” Although the general proposition of prenatal influences is now widely accepted, the underlying mechanism s of how the prenatal environment influences the developing organism and modulates brain structure–function relations, behavior, health, and disease risk are yet to be fully elucidated (e.g., Crews et al., 2012; Fox, Levitt, & Nelson, 2010; Hofer, 2014; Kolb et al., 2012; Lutz & Turecki, 2014; McEwen & Morrison, 2013; Meaney, 2010; O’Connor, Monk, & Fitelson, 2014; Reul et al., 2015; Schlotz, Jones, Godfrey, & Phillips, 2008; Zannas & West, 2014). Debates on which theoretical frameworks are best able to integrate most of the available research results in a coherent way are ongoing (e.g., Bock, Poeschel et al., 2014; Bock, Rether, Gröger, Xie, & Braun, 2014; Daskalakis, Bagot, Parker, Vinkers, & de Kloet, 2013; Daskalakis & Yehuda, 2014; Del Giudice, 2012; Del Giudice, Ellis, & Shirtcliff, 2011; Hanson & Gluckman, 2014; Lee & Goto, 2013; Lupien, McEwen, Gunnar, & Heim, 2009; Nederhof & Schmidt, 2012; Ortega-Martínez, 2015; Schlotz & Phillips, 2009).

An important reason for our scientific quest, which started some 35 years ago, was the question of when and how individual differences between people arise as well as when and how these processes can be studied. In common with Joffe (1969), we extended our search to the organism’s earliest environment and started to examine such issues as the variation in maternal emotional state during pregnancy and how this might lead to a variation in the behavior of the offspring before and after birth, including potential behavioral problems. To achieve this goal, concepts and methods were borrowed from the fields of developmental psychobiology /developmental behavioral neuroscience (e.g., Blumberg, Freeman, & Robinson, 2010; Gottlieb, 1997; Kolb et al., 2012; Lickliter, 2007; Michel & Moore, 1995), developmental cognitive neuroscience (Johnson, 2011; Johnson & de Haan, 2011), developmental affective neuroscience (e.g., Pollak, 2005; Schechter, 2012), developmental psychology (e.g., Kopp, 1982; Rothbart & Derryberry, 1981; Sameroff, 1975; Sameroff & Chandler, 1975), clinical psychology, and child psychiatry (e.g., Bayley, 2006; Rutter, 1987, 1995, 2002; Verhulst, van der Ende, & Koot, 1996). Our studies contributed to the efforts of interdisciplinary and multidisciplinary researchers examining the Developmental Origins of Health and Disease (DOHaD) hypothesis. This hypothesis originated from the “fetal programming of adult disease”—the hypothesis which states that an adverse fetal environment induces plastic responses that increase the risk of chronic diseases such as type 2 diabetes and coronary heart disease later in life (Barker, 1990, 1995, 2004; Barker & Osmond, 1986). Studies testing the DOHaD hypothesis include early prenatal and perinatal origins of a wide range of diseases and disorders, physical as well as mental, by adverse influences during sensitive periods of development (Gillman, 2005; Gluckman & Hanson, 2004; Gluckman, Hanson, & Beedle, 2007; Hanson & Gluckman, 2014; Nathanielsz, 1999; Meaney, 2010; Meaney, Szyf, & Seckl, 2007; Phillips & Jones, 2006; Seckl, 2007; Seckl & Holmes, 2007; Schlotz et al., 2008). The DOHaD research field was influenced by researchers studying the adaptive and/or maladaptive nature of neural, physiological, and behavioral responses to environmental stressors. Specifically, it was influenced by early life stress (ELS) research, targeting consequences of ELS and individual differences in resilience and vulnerability to stress and adversity later in life and the development of stress-related diseases (e.g., de Kloet, Claessens, & Kentrop, 2014; de Kloet, Joëls, & Holsboer, 2005; de Kloet, Karst, & Joëls, 2008; Gunnar & Quevedo, 2007; Heim & Nemeroff, 2001; Koolhaas et al., 2011; Lupien et al. 2009; McEwen & Morrison, 2013; Meaney, 2010; Meaney, Szyf, & Seckl, 2007; Reul et al., 2015; Swaab, Bao, & Lucassen, 2005). Most of this work demonstrated support for a “three-hit model” (Daskalakis et al., 2013). This is a model in which early life adversity does not directly or inevitably lead to disorder or disease but rather genetic factors (hit 1) in interaction with early life environmental inputs and experience -related factors (hit 2) lead to a certain phenotype with differential susceptibility to later-life challenges. In an individual with a certain phenotype when exposed to a later-life environment (hit-3), his or her mental functions may become compromised and a higher risk of psychiatric symptoms may arise (vulnerability); but when exposed to another type of environment the same individual is expected to be resistant to mental dysfunction (resilience) (Daskalakis et al., 2013, p. 1867).

Recently, we introduced a model to describe adaptation to environments in general rather than limited to early adversity (Van den Bergh, 2010, 2011a, b). Our model is compatible with the latter models (e.g., Daskalakis et al., 2013) in that it attempts to explain the association between early life events (typical as well as atypical; positive as well as negative) and physical and mental health as well as physical and mental health problems in a schematic way. Our model extended the concept of DOHaD by explicitly incorporating the study of behavior and brain-behavior relations in the DOHaD hypothesis (Van den Bergh, 2010, 2011a, b; see Fig. 14.1). The study of the Developmental Origins of Behavior, Health and Disease (DOBHaD) encompasses both short- and long-term consequences of conditions in the environment relevant to behavior, health, and disease risk and addresses research issues related to the interface between developmental, behavioral and medical science.

Fig. 14.1
figure 1

Developmental origins of early brain and behavior development, mental health and mental health problems (adapted from Van den Bergh, 2010, 2011a, b)

Full size image

As demonstrated in this chapter, our DOBHaD model can be used to integrate both new and older data. It integrates the results of human literature on prenatal stress and relates them to results of preclinical, experimental animal studies which associated offspring outcome measures to changes in underlying neural circuits and causally related epigenetic processes (Bock, Poeschel et al., 2014; Bock, Rether et al., 2014; Weaver et al., 2004). The model illustrates that early life events will influence the development of organs , such as the heart, lungs, bones, kidneys and including development of the brain and neural circuitry, influence behavior and evolve into a certain programmed phenotype that finally leads to mental and physical health problems in some later-life environments and to mental and physical health in other environments.

During the prenatal and early postnatal life period the brain is subject to dramatic developmental processes; this period represents a phase of high susceptibility towards environmental influences. The specific physiological, neuroendocrine, and metabolic alterations that enable the individual to adapt to its early environment depend on the timing, duration, type, and magnitude of exposure of the organism to environmental factors and also are influenced by the individual’s genetic and epigenetic susceptibility. The latter implies that individuals differ in their susceptibility to early environmental factors (Daskalidas et al., 2013, Daskalakis & Yehuda, 2014; de Kloet et al., 2014; Nederhof & Schmidt, 2012). A causal pathway often studied in animal research is the alteration of the programming of the stress-regulating system [including epigenetic modifications in the hypothalamic–pituitary–adrenocortical (HPA)-axis and autonomic nervous system ] by early events. Adaptive developmental plasticity leads to changes in the limbic brain structures (hippocampus, amygdala) and the prefrontal cortex, which are involved in (stress) reactivity and regulation patterns, in emotional (e.g., anxiety, anger) and cognitive (e.g., sound perception, appraisal, learning , memory) processing and in temperament al variation in behavior (e.g., fearfulness or negative reactivity; surgency or positive reactivity; harm avoidance, novelty seeking) (Kolb et al., 2012). These changes in underlying circuits may influence how an individual “behaves” (i.e., senses, perceives, appraises, responds) in its environment in general. Moreover, situations of acute and chronic stress and adversity later in life will trigger its stress system in a particular way (Ladd et al., 1999; Lee & Goto, 2013; Seckl, 2007) and express the evolving “programmed phenotype .” Exposure of a person with a certain programmed phenotype to some types of environments may lead to behavioral problems, psychopathology, or more generally mental health problems (vulnerability) while the same phenotype exposed to another type of environment may lead to mental health or resistance to mental dysfunction (resilience).

Having developed a general model of adaptation to the early environment, the focus of this chapter is to demonstrate how research findings from our laboratory over 35 years were integrated into the model, describing four studies in particular which serve as exemplars (Braeken et al., 2014; Van den Bergh, 1989, 1990; Van den Bergh & Mulder, 2012; van den Heuvel, Donkers, Winkler, Otte, & Van den Bergh, 2014). Collectively, the research questions in the four studies focused on psychological and physiological processes and their interplay in examining individual differences in behavior in the offspring. Specifically, these longitudinal studies focused on describing neurobehavioral (Studies 1, 2, and 4), neurocognitive (Study 3), and neurophysiological (Studies 3 and 4) functioning from fetuses (Study 1) to infants (Studies 3 and 4) to adolescents (Study 2). It is important to note that, although all of the studies involved uncomplicated pregnancies in healthy pregnant women, there were a full range of both anxiety and mindfulness scores on the instruments used to measure these maternal variables (e.g., 25 % of women were highly anxious (Study 1); some women had a resolved anxiety disorder (Study 4). To show advances in research thinking and techniques over time, the chapter is divided into two major sections. In the first section, early and later studies from the Leuven prospective follow-up projectFootnote 1 are described (Van den Bergh, 1989, 1990; Van den Bergh & Mulder, 2012). Study 1 examines the influence of maternal emotions on fetal and neonatal behavioral state-related activity and on infant activity (Van den Bergh, 1989, 1990, 1992; Van den Bergh et al., 1989). Study 2 examines the relationship between fetal behavioral states and self-regulation in childhood and adolescence (Van den Bergh & Mulder, 2012). Because other results of the Leuven prospective follow-up project are reviewed elsewhere (see Van den Bergh, 2011b; Van den Bergh, Loomans, & Mennes, 2015), they were not included here. In the second section, recent studies from the Tilburg prospective follow-up project are described. Study 3 explores how variation in both negative emotions (i.e., maternal anxiety) and positive emotions (i.e., maternal mindfulness) influence infant neurocognitive development (van den Heuvel et al., 2014). Study 4, explores the issue of how exposure to a past, resolved maternal anxiety disorder (what is presumed to be an atypical maternal emotional state) influences maternal heart rate variability during pregnancy as well as infant heart rate variability, which in turn influences infant temperament (Braeken et al., 2013). In the final section, we summarize how exposure to maternal emotional state during pregnancy may have changed offspring behavior (i.e., his or her sensation, perception, appraisal, reactivity), use these results to explain the potential use of the DOBHaD model and speculate on potential clinical implications.

Section 1: Maternal Emotions , Fetal and Neonatal Behavioral States, and Child and Adolescent Self-Regulation

Historical Background

In the 1970s, the introduction of real-time ultrasound imaging enabled the direct, standardized, noninvasive study of the human fetus in utero. There was a resurgence of the work of early scientists who attempted to demonstrate classical conditioning (Ray, 1932; Spelt, 1948) and habituation (Peiper, 1925; Sontag & Wallace, 1934) in the fetus and those who examined whether maternal emotions influenced fetal behavior (Sontag, 1941, 1944). As knowledge that had been gained in various fields was brought together and reexamined with new techniques (Eskes, 1992), new paradigms emerged. Groundbreaking insights were reached, such as those on the developmental sequence of fetal movements, starting from 7.5 weeks gestational age (de Vries, Visser, & Prechtl, 1982, 1985) and on the development of fetal behavioral states (Nijhuis, Prechtl, Martin, & Bots, 1982).

Reviewing the literature a decade following this initial surge of research, Hepper (1992) suggested that the renewed interest in the fetal period may have arisen because of a change in the view of the capabilities of the newborn from being poorly developed and unable to adjust to his/her environment to having abilities exquisitely enabling adaptation. He attributed this reappraisal to the ability of scientists to ask the right questions. His ideas were reminiscent of Joffe’s (1969) interpretation of the importance of the prevailing climate or opinion regarding prenatal influences. In the same vein, Prechtl remarked that some writings of fetal behavior from around the turn of the nineteenth to twentieth century “were surprisingly modern, while writings from the 1920s and 1930 often seem extremely limited or obsolete”; the dominance of reflexology and behaviorism during the latter period may account for the prevailing view at that time (Prechtl, 1984, p. 5). One of the clear implications of psychobiological research being (re)-focused on the competences of the neonate and fetus was “the importance of early events in shaping subsequent development” (Smotherman & Robinson, 1995, p. 15), the study of which became a major topic of our research program.

We began our own empirical research program with an analysis and synthesis of the diverse literature on human maternal emotions during pregnancy (Van den Bergh, 1981, 1983, 1989). Many of the studies reviewed showed methodological shortcomings, such as: a failure , to specify the sample characteristics; insufficient control conditions; conclusions based on retrospectively obtained, inadequate assessment of predictor and outcome variables; lack of sound statistical methods; the problem of causation versus correlation; and the problem of rater bias, when the mother completed self-report questionnaires on her own emotional state or her offspring’s behavioral problems. Nevertheless, already at that time, results of the methodologically sound studies (e.g., Farber, Vaughn & Egeland, 1981 and see Carlson & Labarba, 1979 for a review) led to one general conclusion: negative maternal emotions during pregnancy may influence prenatal as well as postnatal behavior in the offspring and lead to behavioral problems and diseases later in life. Indeed, there was evidence that increased levels of negative emotionality and stress in the pregnant women may influence fetal brain development and behavior such as that shown by increased fetal heart rate (FHR) and motility (e.g., Copher & Huber, 1967; Sontag, 1941, 1966) and are associated with pregnancy and birth complications such as hyperemis gravidarum, toxemia, premature birth , and lowered birth weight (e.g., McDonald, 1968). Both negative maternal emotions and pregnancy and birth complications were shown to be associated with developmental irregularities and behavioral problems (e.g., Carlson & Labarba, 1979; Dörner, 1974; Erickson, 1971, 1976a, 1976b; Ferreira, 1960, 1965; Istvan, 1986; Knobloch & Pasamanick, 1966; Pasamanick & Knobloch, 1966; Pasamanick, Rogers, & Lilienfield, 1956; Sameroff & Zax, 1973; Sameroff & Chandler, 1975; Stott, 1958), altered mother–child interactions and adjustment (e.g., Davids & Holden, 1970; Davids, Holden, & Gray, 1963; Farber et al., 1981), childhood diseases (Stott, 1973; Stott & Latchford, 1976), and alterations in adult personality (Sontag, 1966). (Note: The link between maternal stress during pregnancy and childhood disease was recently confirmed in a very large scale (n = 66,203) prospective cohort study in Denmark (Tegethoff, Greene, Olsen, Schafner, & Meinlschmidt, 2012).

It was our aim to study processes that constitute offspring behavior in a prospective and standardized way as this would enable us to identify potential indices [or (bio)markers] of altered programmed phenotype and underlying mechanisms . Identifying indices and mechanisms is interesting from a basic scientific as well as from an applied, clinical point of view, namely to predict the risk of behavioral problems or disease in some environments later in life and, more generally, to set up innovative preventative and intervention strategies.

Study 1: Relationship Between Maternal Emotions and Fetal, Newborn , and Infant Behavior

In the early 1980s our research aims were to study two links which, according to our review, had not been systematically explored with a sound method, design or statistical technique: Can the influence of maternal emotions upon fetal behavior be established in the prenatal period? (first aim) and, is the prenatal influence established in the prenatal period reflected in the neonatal and infant behavior? (second aim)

While previous studies relied on maternal report of fetal movements or on fetal heart rate measures, the introduction of ultrasound in human research facilitated the study of these two links by enabling the direct measurement of fetal behavior for prolonged periods in pregnant women with varying levels of anxiety. Given the groundbreaking results from the work of colleagues in Groningen (de Vries et al., 1982, 1985; Prechtl, 1974, 1984; Visser, Poelman-Weesjes, Cohen, & Bekedam, 1987) and Leuven (e.g.,Casaer, 1979, 1993; Casaer & Devlieger, 1984; Casaer & Eggermont, 1985; Casaer, O’Brien, & Precht, 1973; Deprest et al., 1998; Van Assche, 1997; Van Assche, Holemans, & Aerts, 2001; Vandenberghe & Dewolf, 1990), we selected fetal and neonatal behavioral states and state-dependent movements as primary outcome variables. Fetal behavioral states emerge during the third trimester of pregnancy. They involve multiple interconnected neuronal networks. Functional (re)organization of sleep cycling likely occurs around 28–30 weeks gestational age (GA), 36 weeks GA, and 2 months of age (Nijhuis et al., 1999; Scher, 2008; Visser et al., 1987). From 36 weeks of gestation onward, the low-risk fetus exhibits two states of sleep and two states of wakefulness. (See Chap. 6 by Nijhuis this volume for a description of each state.) Fetuses typically pass through sleep cycles of non-REM (quiet) sleep and REM (active) sleep, lasting about 70–90 min (Visser, Mulder, & Prechtl, 1992). The time spent in wakefulness is usually less than 10 %. Typical fetal sleep states show a concordant (uninterrupted) association between the state parameters for a prolonged time and a simultaneous (synchronized) change of state parameters (≤3 min) at their beginning and end (transitions). The degree of sleep state stability and the duration of transitions into and out of a particular state are considered measures of neurophysiological development, integrity, and maturity (Mulder, Morssink, Van Der Schee, & Visser, 1998; Visser et al., 1992).

Our study was designed to examine both the issue of the influence of maternal emotions on fetal behavior (aim 1) and the effects of alterations in fetal behavior on neonatal and infant behavior (aim 2) in the same population. The sample included 86 Dutch women, 18–30 years of age, in their first pregnancy which was singleton, low-risk, with no medication or drug use. All pregnancies were dated using the last menstrual period and/or an ultrasonographic examination before 14 weeks. The course of pregnancy remained unremarkable with delivery from 36 to 41 weeks of gestation. All infants had a birth weight above the 10th percentile, a 5-min Apgar score of 9 or 10, and no postnatal medical complications while in hospital. Maternal pregnancy anxiety was measured at 12–22, 23–31, and 32–40 weeks GA and at 1 week, 10 weeks, and 7 months after delivery using the self-report state anxiety subscale of the Spielberger State Trait Anxiety Inventory (Spielberger, Gorsuch, & Lushene, 1970; Van der Ploeg & Defares, 1980). In a subsample of 37 women [Mean age (SD) = 26.6 (2.55)], simultaneous recordings of fetal heart rate (FHR), fetal generalized body movements (GM) and fetal rapid eye movements (REM) were made continuously for 2 h at 36–38 weeks GA. FHR recordings were collected using a cardiotocograph and scored visually into episodes of heart rate pattern (HRP) A, B, C, or D (Nijhuis et al., 1982). Fetal generalized body and eye movements were observed and video-recorded using real-time ultrasound scan. The presence of states (or coincidence) 1F–4F was identified according to predefined criteria (see Mulder, Visser, Bekedam, & Prechtl, 1987; Nijhuis et al., 1982). For some analyses, the data of state 2F and 4F, were combined (in order to have one measure of states in which fetuses are actively making movements); in other ones, state 2F, 3F, and 4F were combined to have one measure of state other than 1F/quiet sleep. On day 5–6 after birth, a comparable 2-h observation was carried out on the newborn. At 7 months of age, maternal report of the infant activity was dichotomized as problematic or not problematic. To examine our two research aims, we constructed and tested several sets of nested linear structural relation (LISREL)-models, including state and/or trait anxiety in each pregnancy trimester as predictor variable(s). To select the fetal behavioral measures to be included in the LISREL-models, component analyses were performed, with variamax rotation of components with eigenvalue >1 and the measures with the highest values on each of the components were selected; comparable neonatal measures were then introduced in the models. Fetal sex was introduced as a second predictor variable since our data showed that male fetuses were more active than female fetuses. Maternal anxiety postpartum had no significant association with neonatal behavioral measures and was not introduced in the neonatal LISREL model. However, as maternal anxiety 7 months after delivery was significantly associated with the infant activity measure, we introduced it as a second predictor to examine the link with infant activity at 7 months. The models were tested on n = 28 fetuses/neonates for whom all data were available.

As can be seen in Fig. 14.2—Model A, LISREL modelling showed that fetuses of high anxious mothers made more general movements and head movements and that male fetuses made more general movements than females. (Note: Only the final model of nested LISREL models are shown.) The prenatal influence of maternal State anxiety also was reflected in neonatal behavior; infants who made more head movements as fetuses, made more general movements and more head movements as neonates. The observation that the percentage of fetal head movements (rather than the percentage of fetal general movements) was significantly related to neonatal head and general movements may indicate that the newborns had difficulties in adjusting to gravidity, which may have more effect on making body movements than on making head movements (Michel & Moore, 1995; Prechtl, 1984). Model B illustrates that maternal Trait anxiety had a negative influence on the mean duration of epochs of coincidence of fetal State 1, and that infants who have shorter epochs of State 1F as fetuses also have shorter duration of epochs of State 1 as neonates . Model C reveals that at 7 months after birth, maternal anxiety during pregnancy had an indirect effect on activity level of the infant (i.e., via influencing fetal general movements) and maternal anxiety measured at 7 months after delivery had a direct influence on infant activity level. By this time, the mother and , infant would have had much more opportunity to interact than at four or five days after birth. Of course, these differences may have resulted because infant activity was reported by the mother and reflect her bias. In future research, direct observation of maternal–infant interactions at this age might be useful in untangling these effects.

Fig. 14.2
figure 2

LISREL model A,B,C concerning the relation between maternal anxiety during pregnancy, fetal and neonatal behavioral states and motor activity, and infant activity (adapted and translated from Van den Bergh, 1989). Legend: 7 months = 7 months after delivery; df = degrees of freedom; Epoch C1F-MD: Mean duration of epochs of coincidence State 1 in the fetus; Epoch 1N-MD: Mean duration of epochs of State 1 in the neonate ; % GM = percentage of general movement s; % HM = percentage of head movements; NoC1F = epochs of coincidence other than State 1F; C24F = epochs of coincidence 2F and 4F; Trim2 = Second trimester of pregnancy; Trim3 = Third trimester of pregnancy

Full size image

In summary, the LISREL modelling, together with other results from this cohort (e.g., those showing that infants of high anxious pregnant mothers had a more difficult temperament at 10 weeks and cried more, were hungrier, had more stomach cramps at 7 months than infants of mothers who were less anxious, see Van den Bergh, 1989, 1990, 1992) answered our two research aims cited above in a positive way. Results of Study 1 can be integrated into our DOBHaD model (see Fig. 14.1) in the following way: prenatal exposure to maternal anxiety (early life events) in interaction with supposed (epi)genetic factors, may have an enduring influence on (or programs) fetal/neonatal brain development and behavioral functioning as reflected in higher fetal and neonatal reactivity. This higher reactivity may evolve into a programmed phenotype including seeking out arousal-inducing events. Moreover, the induced behavioral alterations observed in the offspring of mother’s with high anxiety levels may influence the quality of the interaction between mother and child (i.e., the caregiving environment) in a negative way, increasing the risk of subsequent offspring behavioral or mental health problems. [Further explanation of the (use of the) DOBHaD model is given in the final section which clarifies the i ntegration of the results of this study.]

A link between the level of stress and anxiety of the mother during pregnancy and ultrasonographically observed fetal behavior and fetal heart rate is now well established (for a review see DiPietro, Costigan, Pressman, & Doussard-Roosevelt, 2000; Kafalí, Derbent, Keskí, Símavlí, & Gözdemír, 2011; Monk et al., 2011; Van den Bergh, Mulder, Mennes, & Glover, 2005). Most studies have reported that increased maternal anxiety was associated with increased fetal arousal/wakefulness and increased FHR variability and % of body movements during states 2F and 4F . As an example, DiPietro and colleagues (2000) observed that fetuses of women with a positive versus negative attitude toward pregnancy exhibited different overall levels of motor activity (reduced versus increased, respectively). Although positive (pleasant, optimistic) emotions and negative stressors are believed to be regulated by the same physiological system [hypothalamic–pituitary–adrenal (HPA) axis and autonomic nervous system (ANS)], the negative emotions may have reflected chronic negative conditions, which were both unpredictable and uncontrollable and triggered a stress response, involving cortisol release (Koolhaas et al., 2011), while positive emotions (dispositional optimism) have been linked to lower levels of cortisol responses under stress (Jobin, Wrosch, & Scheier, 2014).

Study 2: Relationship Between Fetal Behavioral State and Child and Adolescent Self-Regulation

Having observed in Study 1 that maternal anxiety did influence fetal behavioral state and was associated with alterations in state-related fetal activity level, we turned our attention to the influence of fetal states on self-regulation. Specifically, we addressed the question: Is fetal behavioral state organization a biological precursor of child and adolescent self-regulation? (Van den Bergh & Mulder, 2012)

Background

Sleep plays a critical role in early brain development, arousal regulation, attention and cognition (Graven & Browne, 2008; Mirmiran, Maas, & Ariagno, 2003; Mulder, Ververs, de Heus, & Visser, 2011; Peirano, Algarín, & Uauy, 2003) and the study of sleep ontogeny (i.e., behavioral state organization) can be used to identify patterns of brain maturation (Scher, 2008). For example, in one study (Scher, Steppe, & Banks, 1996), sleep measures of both the healthy preterm infant (assessed at term equivalent age) and the healthy full-term newborn were predictive of performance on the Bayley scales of mental development at 12 and 24 months. In another study (Holditch-Davis & Edwards, 1998), in high-risk premature infant s born at gestational ages from 27 to 29 weeks onwards, the degree of sleep state control after birth was associated with postnatal neurodevelopmental status at term equivalent age. The predictive value of these measures for behavioral developmental outcome in later life has remained unexplored due to a lack of long-term follow-up studies. Therefore, in a nonclinical sample (see Study 1 for a description of this sample), we examined whether differences in sleep state organization in the near term fetus could account for differences in child and adolescent self-regulation (Van den Bergh & Mulder, 2012).

Theories of self-regulation presume that human s, from prenatal life or birth onward, display individual differences in reactivity and regulation that have implications for subsequent development and adaptation (Calkins & Fox, 2002; Gunnar, Talge, & Herrera, 2009; Henrichs & Van den Bergh, 2015; Kochanska, Coy, & Murray, 2001; Kopp, 1982; Posner & Rothbart, 2000; Pruessner et al., 2010). Reactivity is understood as the arousability of physiological and behavioral systems, while self-regulation refers to neural and behavioral processes which function to modulate this reactivity. Interestingly, in some theories, temperament has been defined as constitutionally based individual differences in reactivity and regulation (Rothbart & Ahadi, 1994; Rothbart & Bates, 1998; Rothbart & Derryberry, 1981; Rothbart, Sheese, Rueda, & Posner, 2011). As the infant and child mature, later-developing neural structures become integrated into the existing neural organization, which involves reorganization of circuits (Michel & Moore, 1995). Due to this patterned reorganization, initial reactive forms of regulation are supplemented by an increasing capacity for volitional, effortful control or self-regulation (Derryberry & Rothbart, 1997). Much of the self-regulation development results from increasing volitional control over attention al processes and enhanced inhibitory control over motor behavior (Calkins & Fox, 2002). Starting in childhood and continuing throughout adolescence , executive function s such as attention al focusing, maintenance and shift of focusing, and inhibitory control become integrated into complex emotional and behavioral regulatory processes. These processes, in turn, are involved in planning and goal setting, responsible decision making, emotional and motivational changes, and interpersonal relationships (Nelson et al., 2002; Rothbart & Bates, 1998; Van den Bergh & Mulder, 2012, p. 585). Failure of self-regulation in one way or another is a characteristic feature of behavioral problems and mental disorders (Henrichs & Van den Bergh, 2015).

At the time of our study, we could find no empirical work on individual differences in typical fetal brain maturation processes, such as expressed in fetal behavioral state organization or in relation to the long-term consequences for self-regulation. Thus, the aim of this prospective longitudinal study was to examine which measures of fetal behavioral state organization in the normal, near-term fetus are predictors of measures of self-regulation obtained from the same individuals when 8–9 and 14–15 years of age.

A total of 73/86 offspring participated in this second study. Twenty-five mother–offspring pairs who had participated in the fetal observation part of the Leuven study detailed above and had complete data for both the fetal behavioral observation session at the end of pregnancy and a follow-up study on the offspring at ages 8–9 or 14–15 were included. The reference (i.e., comparison) group consisted of 48 mothers and their , children/adolescents who participated only in the follow-up study but not in the fetal observation study.

For the follow-up study reported here, the mothers completed Dutch versions of temperament questionnaires, measuring concepts of reactivity (i.e., positive reactivity (or surgency) and negative reactivity) and of self-regulation (i.e., effortful control). The Children’s Behavior Questionnaire (CBQ; Ahadi, Rothbart, & Ye, 1993; Rothbart, Ahadi, Hershey, & Fisher, 2001 translated and validated for a Dutch-speaking sample by Van den Bergh & Ackx, 2003) was used when their children were 8–9 years of age and the revised Early Adolescence Temperament Questionnaire (EATQ-R; Capaldi & Rothbart, 1992; Ellis & Rothbart, 2001; translated and validated for a Dutch-speaking sample by Hartman, Oldehinkel, De Winter, & Ormel, 2002) when the children were 14–15 years old. Only the temperament data concerning self-regulation are used in this study. Statistical modelling of the fetal–child–adolescent data demonstrated that one behavioral state measure, namely the time a typically developing fetus takes to pass from quiet sleep (S1F) to active sleep (S2F) in the last month before birth, is associated with her/his degree of self-regulation in childhood and adolescence . In particular, fetuses exhibiting sharp, synchronous transitions from quiet sleep into active sleep , compared with fetuses showing non-synchronized transitions (lasting >3 min) reached a higher level of effortful control (i.e., higher than the reference group but within normal ranges) both at 8–9 years and 14–15 years. Although the mechanism s underlying fetal state transitions are yet unknown and in need of future study, our results demonstrate that studies of sleep ontogeny can provide insights into fetal brain maturational processes which have implications for later environmental adaptation as well as developmental consequences for behavior. The results of Study 2 can be integrated in our DOBHaD model (Fig. 14.1). The supposed interaction between fetal environmental and (epi)genetic factors is reflected in synchronous fetal state transitions from quiet into active sleep in some fetuses and in asynchronous ones in other fetuses; these types of transitions are one element of early brain-behavior processes. These early differences may evolve into a programmed phenotype implying optimal self-regulation in the former group and implying suboptimal self-regulation in the latter ones.

Section 2: Maternal Heart Rate Variability and Emotions Are Associated with Infant Cognition and Heart Rate Variability

Background

While the participants in the following studies were pregnant women and their offspring, the offspring were only measured after birth; however, the studies are included here because they have implications for fetal psychobiological development. The studies described below were part of the Prenatal Early Life Stress (PELS)-project, a multinational, European (i.e., Belgium, Netherlands, UK) project. The PELS-project was one of four projects of the “Stress and Mental Health” program (EuroSTRESS) in which the research questions aimed at increasing our knowledge of the basic mechanism s of stress-related mental disorders as well as advancing our understanding of how early life experience s, genetic makeup, and repeated traumatic events in adulthood might predispose a person to the development of mental health disorders. The ultimate objective of the EuroSTRESS-project was the use of this knowledge for the development of new treatment strategies and the prevention and/or amelioration of such disorders (for more information see: http://www.esf.org/coordinating-research/eurocores/completed-programmes/eurostress.html).

To identify specific maternal risk (and resiliency) factors during pregnancy having an influence on offspring neurodevelopment, a total of 151, 170, 190 pregnant women was recruited in the UK, Belgium and the Netherlands respectively. Each country gathered information on stress, anxiety and depression levels by having the pregnant women complete self-report questionnaires in addition to providing saliva samples for cortisol measures in each pregnancy trimester and at 2–4 and 9–12 months after delivery. For the infants, birth outcome data were collected; infant saliva cortisol was measured at 2–4 and 10–12 months-of-age. The Bayley Scales of Infant Development as well as a behavioral inhibition task were administered at 9–12 months after birth. In the Netherlands, infant event related potentials (ERP) and heart rate variability were measured at 2–4 and 9 months of age. Epigenetic analyses limited to one cohort (Belgium) showed that prenatal maternal emotional state, particularly pregnancy related anxiety, was associated with the methylation state of the NR3C1 gene in the child (Hompes et al., 2013; Hompes, 2014).

Study 3: Maternal Anxiety and Mindfulness During Pregnancy and Infant Neurocognitive Function

Electroencephalography (EEG), in general and event-related brain potentials (ERP ) specifically, are unique tools which can be employed to assess cognitive functions such as attention , habituation and memory in early infancy. Indeed, ERP recordings in infants and the analysis of the responses have become well developed in the past 25 years (e.g., Alho, Sainio, Sajaniemi, Reinikainen, & Näätänen, 1990; for a review, see Kushnerenko, Van den Bergh, & Winkler, 2013). Cognitive abilities may be shown through infant responses to auditory stimuli (e.g., such as those used in auditory oddball paradigms) that mimic important features of the postnatal environment (Smotherman & Robinson, 1995). We examined how exposure to variation in maternal emotional state during pregnancy might influence the early neurocognitive development of the offspring using auditory stimuli, basing our studies on the work of Winkler and collaborators (Winkler, 2007; Winkler, Háden, Ladinig, Sziller, & Honing, 2009; Winkler et al., 2003).

Auditory attention is a key aspect of early neurocognitive function as it is a prerequisite of important skills, such as learning to speak and communicate with others. Moreover, some non-speech sounds also require one’s attention because they may signal an opportunity or some danger and need further processing. Some sounds may be irrelevant for the current behavioral goals and their processing should be stopped or suppressed (Kushnerenko et al., 2013). Obviously, it is important to be able to differentiate between these types of sounds. While being vigilant is adaptive in a new and/or hostile environment, being constantly alert and vigilant even if the environment is more favorable may be maladaptive. The finding that individuals exposed to early life stress are more vigilant (which is a key characteristic of anxious individuals) is consistently found in animal studies (Gunnar & Quevedo, 2007; Lutz & Turecki, 2014; Weinstock, 2005, 2008), while the assocation between maternal anxiety during pregnancy and childhood anxiety or emotional problems is found in some human studies (O’Donnell, Glover, Holbrook, & O’Connor, 2014; O’Connor, Heron, Golding, & Glover, 2003; Van den Bergh & Marcoen, 2004). Thus, we looked for early markers of these traits or problems in infants exposed to varying levels of maternal anxiety during prenatal life. Moreover, as human DOBHaD studies have almost exclusively focused on the effects of negative maternal emotions during pregnancy on child neurocognitive function, the focus of this study was expanded to include exposure to positive maternal emotions to determine whether and how they might also influence fetal (brain) development. Such a focus was both theoretically interesting and clinically relevant. For instance, Lobel, DeVincent, Kaminer, and Meyer (2010) had shown that in women with high-risk pregnancies , optimism was a key protective factor against adverse pregnancy outcomes. Thus, a good candidate for our study was a positive trait, such as mindfulness (Keng, Smoski, & Robins, 2011). Being mindful refers to a state of mind consisting of two key elements: (1) An alert mode of perceiving all mental contents (i.e., perceptions, sensations, cognitions, and emotions) and (2) a friendly, accepting, and nonjudgmental attitude towards those mental contents (Kohls, Sauer, & Walach, 2009). During pregnancy, experiencing positive emotions due to these two factors may enhance a pregnant woman’s resilience against stress and adversity occuring during the pregnancy and hence constitute a prenatal environment that positively influences fetal brain development.

To study the relationship between maternal anxiety and mindfulness and offspring outcome, data from 79 Dutch mother–infant pairs were employed (van den Heuvel et al., 2014). At 20 weeks of gestation, women reported anxiety using the Symptom Checklist (SCL-90; Arrindell & Ettema, 1981, 2003) and mindfulness using the Freiburg Mindfulness Inventory (FMIs-14, Walach, Buchheld, Buttenmuller, Kleinknecht, & Schmidt, 2006). When their infants were 9 months-of age, EEG and auditory elicited ERPs were recorded using a passive auditory oddball paradigm. The stimulus sequences consisted of four different types of 200 ms sound events with an interstimulus interval of 300 ms, namely the standard sound and three deviant sounds (i.e., a white noise segment, a unique environmental sound such as slamming a door, and the same sound as the standard sound but with an interval of 100 ms). The frequent standard had a probability of 0.70 and the three types of deviants each had a probability of 0.10. A total of 1500 stimuli were delivered.

Mixed-mode ANOVAs were employed in two separate analyses, including the infants’ mean ERP amplitudes (elicited by each of the four types of sound events) and either maternal anxiety or maternal mindfulness as a predictor. Preliminary analyses showed no effects of gestational age, birth weight or maternal anxiety at 9 months after delivery and they were not included in subsequent modelling. The results showed that higher maternal mindfulness (during the second trimester) was associated with smaller infant N250 and higher infant P150 ERP amplitudes to the standard sound while higher maternal anxiety (during the second trimester) was associated with larger N250 amplitudes to the standard sound. No effects were found for the three deviant sound stimuli.

From these results, we concluded that infants prenatally exposed to higher levels of maternal mindfulness devote less in-depth processing to repeated sounds with low information content, suggesting fast habituation to these sounds. In contrast, infants prenatally exposed to higher levels of maternal anxiety processed such uninformative sounds more extensively and/or they habituated more slowly to these stimuli. We speculate that the 9 month-old infant ERP directional differences observed here to higher maternal mindfulness and anxiety during pregnancy might stem from infants prenatally exposed to higher maternal mindfulness pre-attentively forming more accurate perceptual representations, as reflected in higher P150 amplitudes to the standard sound. If so, a mindfulness intervention for pregnant women suffering from anxiety may be a desirable alternative or adjunct to pharmacological interventions. Clearly, firm conclusions await future research (van den Heuvel et al., 2014). The findings in relation to higher maternal anxiety during pregnancy are consistent with the results of other studies, namely with those showing that children prenatally exposed to high maternal anxiety have poorer language acquisition (King & Laplante, 2005; Laplante et al., 2004; Laplante, Brunet, Schmitz, Ciampi, & King, 2008) and are more anxious (O’Connor et al., 2003; Van den Bergh & Marcoen, 2004).

The results of Study 3 can be integrated in the DOBHaD model in the following way: prenatal exposure to maternal anxiety (early life events) in interaction with supposed (epi)genetic factors lead to altered emotion and (neuro)cognition and may evolve into increased vigilance (constituting the programmed phenotype ).

As well as investigating the effects of anxiety during pregnancy on later offspring development, we also examined the effects of resolved anxiety disorders during pregnancy on infant development.

Study 4: Heart Rate Variability in Pregnant Women and Their Infants

At the time that we began this research, from our own studies (see above) and that of others (e.g., Alder, Fink, Bitzer, Hösli, & Holzgreve, 2007; Ross & McLean, 2006), it was known that active anxiety disorders and experiencing a high anxiety level had long-term detrimental effects on pregnant mothers and their offspring. However, it was unknown if a resolved, nonactive, maternal anxiety disorder had similar effects. Anxiety-related conditions, such as reduced autonomic cardiac control, indicated by reduced heart rate variability (HRV) could persist despite disorder resolution, with long-term health implications for mothers and children (Braeken et al., 2013). The autonomic nervous system of the fetus seems to be susceptible to the influence of maternal cardiac characteristics (Young, 2002). Indeed, it has been shown that HRV of the developing fetus is altered in the offspring of mothers with a number of psychiatric conditions, including anxiety disorders, and these differences persist postnatally (Dierckx et al., 2009; DiPietro et al., 2000; Monk et al., 2004). Thus, we designed this study to test the hypothesis that pregnant mothers with a history of, but not current anxiety disorder, and their children have low HRV, predicting offspring anxiety-like temperament (Braeken, 2014; Braeken et al., 2013).

To test the hypothesis, a case-control study including 56, 1st trimester Dutch women (n = 22 with a history of anxiety disorder; n = 34 with no history of psychopathology determined using the Mini-International Neuropsychiatric Interview 6.0, Sheehan & Lecrubier, 2010) and their offsprings was carried out. Anxiety was measured with the State Trait Anxiety Inventory (Spielberger et al., 1970; Van der Ploeg & Defares, 1980) and maternal ECG (to obtain maternal HR and HRV) was measured continuously during rest and mental stress. Stress was induced during a mental task. Each mother participated in a 25-min task that consisted of five testing phases, lasting 5 min each (Vlemincx, Taelman, De Peuter, Van Diest, & Van Den Bergh, 2011). Stress was induced in the second and fourth phases, with the remainder being relaxation phases. The stress consisted of mentally solving a complex mathematical problem such as (361 + 11) ÷ (3 × 4) + 137 without verbalization and selecting the answer from three choices presented on a computer screen. The relaxation phases consisted of viewing pictures considered peaceful and listening to music considered restful. At 2–4 months of age, infant ECGs were recorded. At 9–10 months of age, infant fearfulness was assessed using the unpredictable mechanical toy paradigm of the fear subscale of the Laboratory Temperament Assessment Battery (Lab-TAB)-Locomotor Version (Goldsmith & Rothbart, 1999) (Braeken et al., 2013, p. 2–3).

Repeated measures ANOVA controlling for mother’s age and prepregnancy BMI, showed that HRV was lower in women in the past anxiety group compared to controls on both the root mean square of successive differences (RMSSD) and the high frequency (HF) measures of HRV. Regression analysis indicated that there was a significant relationship between maternal HRV measures and child HRV measures only in the anxiety group. Simple effects analysis showed that children of mothers with a past anxiety disorder had lower HRV (for both RMSSD and HF measures) than those born to mothers without a past anxiety disorder history. For all children, low HRV measures at 2–4 months were associated with a higher chance of fearful behavior at 9–10 months.

These results revealed that pregnant women with a past anxiety disorder had autonomic alterations (reduced parasympathetic function, indexed by HRV) early in pregnancy which may have influenced a subsequent physiological (reduced parasympathetic function, indexed by HRV) and/or psychological (fearful temperament ) attribute of their offspring. The findings were independent of variations in maternal state-anxiety, age, sex, or body mass index. Additionally, mother–child associations were not explained by the children’s birth weight or gestational age. The mechanism s by which a previous maternal anxiety disorder and/or HRV become associated with parasympathetic nervous system function in the offspring are unknown and a matter of speculation. It could be that altered autonomic function in pregnant women modulates their fetus’ development. There is some support for this postulate as reduced HRV has been shown to be associated with dysregulation of several allostatic systems, including glucose regulation , hypothalamic–pituitary–adrenal axis function and inflammatory processes (Thayer & Lane, 2007; Thayer & Sternberg, 2006; Thayer, Yamamoto, & Brosschot, 2010) all of which may modulate fetal development (Lupien et al., 2009; Matthews & Phillips, 2010; Meyer et al., 2006; Van den Bergh, 2011; Van den Bergh, Mennes, et al., 2005; Young, 2002). However, whether altered maternal ANS function is causative or simply the result of shared underlying processes is unknown. Alternatively, given that mothers and their children share genes and environmental exposures, maternal behavior also may be an important factor in the observed associations (Rutter, 2002; Stern, 2009; Weaver et al., 2004) as well as shared genes. For example, research has shown that the combination of a brain-derived neurotrophic factor (BDDNF) V/V genotype and early life stress predicts changes in brain structure that are associated with lower HRV and higher anxiety (Gatt et al., 2009). These findings may explain , in part, why the women with a history of an anxiety disorder in this study demonstrated lower HRV. Whether it may account for our observation of a relationship between maternal–infant HRV is unknown (Braeken et al., 2013, p. 6).

Our DOBHaD model (Fig. 14.1) shows the effects of prenatal exposure to resolved maternal anxiety and altered ANS function (early life events) leading to altered ANS function and fearful temperament , which may evolve into a pro-anxiety phenotype (constituting the programmed phenotype ).

Conclusion: The Use and Strength of the DOBHaD Model and Clinical Implications

In the past 30 years and especially in the last decade, an increasing number of studies have provided continuing evidence for an association between prenatal exposure to maternal stress, anxiety and depression , and altered behavior in the offspring. This body of evidence indicates that it may indeed be the case that events prior to birth, such as maternal emotions during pregnancy, influence the way offspring respond to their postnatal environment as demonstrated by correlations with infant outcome in the studies described above. The child of a highly anxious pregnant mother, by adapting to the early exposures when he/she was a fetus, reacts differently than the child of a low anxious pregnant mother, as reflected in newborn activity and EEG-responses, HRV, and temperament in infancy (see above studies), a delay in language development in toddlers and an enhanced risk for behavioral and emotional problems in childhood, specific cognitive problems, anxiety and depression in adolescence and young adulthood (for recent reviews see: Beydoun & Saftlas, 2008; Bock, Poeschel et al., 2014; Braeken, 2014; Charil, Laplante, Vaillancourt, & King, 2010; Glover, 2011, 2014, 2015; Glover, O’Connor, & O’Donnell, 2010; O’Donnell, O’Connor, & Glover, 2009; Graignic-Philippe, Dayan, Chokron, Jacquet, & Tordjman, 2014; Henrichs & Van den Bergh, 2015; Lewis, Galbally, Gannon, & Symeonides, 2014; Lewis & Olive, 2014; Loomans, 2013; Loomans et al., 2011, 2013; Mennes, 2008; O’Connor et al., 2014; Otte, 2013; Räikkönen, Seckl, Pesonen, Simons, Van den Bergh, 2011; Schlotz & Phillips, 2009; Van den Bergh & Henrichs, 2015; Van den Bergh et al., 2015; Van den Bergh, Mulder, et al., 2005; Weinstock, 2008).

To understand how exposure to prenatal (and early postnatal) environmental events may influence later behavior, health and disease, several models have been developed (see for example Bock, Rether et al., 2014; Nederhof & Schmidt, 2012). However, no firm conclusion can yet be drawn about the validity of the different models (Daskalakis et al., 2013). In recently generated models (Bock, Rether et al., 2014; Daskalakis et al., 2013; de Kloet et al., 2014; Hanson & Gluckman, 2014; Lewis et al., 2014), including our own (Van den Bergh, 2010, 2011a), early life events typically are seen as “conditional determinants” rather than as determinants which always/invariably lead to behavioral problems, disorder, or disease. Put simply, this means that an organism that was prenatally programmed (or organized) to be adapted to a particular environment, will gradually be behaving in an altered “biased” way. The “bias” constitutes his or her programmed phenotype and reflects the way the organism was adapted during its early development. Although early adversity will in some environments finally lead to disorder or disease, in other environments early adversity may constitute a possible source of adaptation (Daskalakis et al., 2013). For instance, according to the “mismatch hypothesis” only a mismatch between the early environment and later postnatal environment will lead to disorder and disease, while a match will not (Gluckman & Hanson, 2004; Hanson & Gluckman, 2014). Biological sensitivity (Boyce & Ellis, 2005) or differential susceptibility to the environment (Belsky & Pluess, 2009) models predict that some individuals are more susceptible than others to both the adverse and beneficial effects of, respectively, unsupportive and supportive environments. This genetic difference in sensitivity or susceptibility and the nature of the environment will influence how mental health or mental health problems are shaped; these processes covary with physical health and health problems.

The results of the four studies described above exemplified how “conditional determinants” (i.e., prenatal environmental events such as exposure to maternal anxiety, maternal mindfulness, a resolved maternal anxiety, and altered maternal ANS) might have influenced the course of fetal development, resulting in alterations in the function of the brain and motor systems as evidenced during fetal life and, gradually, in a “biased” response/altered phenotype to subsequent environmental inputs later in life. Using the results of Study 1 to illustrate, being a highly active fetus could be seen as an adaptation to the prenatal environment shaped by a highly anxious mother. The subsequent observation of an increase in neonatal activity might signify that the brain of the infants became shaped to facilitate a higher level of arousal during fetal life. It also could indicate an effect on self-regulation such that later in life, the infant, toddler and child would seek out arousal-inducing events. A consequence of arousal-seeking is that it could lead to hyperactivity and/or impulsivity , which places the child at risk for behavioral problems or attention deficit hyperactivity disorder (ADHD) in some, but not in other environments. For instance increased motor activity will in a (school) environment that requires restraint of impulsivity and motor activity be seen as inadequate behavior. However, in an environment that is stress-inducing and potentially harmful, or in a novel environment, increased motor activity (hyperactivity) may be adaptive (e.g., to explore the environment for threats and opportunities) (Jensen et al., 1997). In Study 3, infants of highly anxious pregnant women were more vigilant than those of low anxious pregnant women (i.e., ERPs indicated that they reacted stronger to a repeated, uninformative sound and seem to habituate less). In Study 4, infants of pregnant women with a resolved anxiety disorder who had lower HRV measures of RMSSD and HF during pregnancy, also showed similar lower HRV measures at 2 months of age, which predicted an anxious-like temperament at 9 months of age. When these infants, with an altered phenotype [i.e., ‘a bias in neurocognitive function (Study 3) and in sympathetic activity/fearfulness (Study 4)] encounter an anxiety or fear-inducing, environment such behavior may have an adaptive value. According to some authors (e.g., Hanson & Gluckman, 2014; Lee & Goto, 2013; Lewis et al., 2014; O’Connor et al., 2014) this bias, which also can be interpreted as a higher degree of stress reactivity, may in fact be promoted by maternal anxiety during pregnancy. The mother signals, with physiological changes accompanying anxiety, aspects of the environment to which the fetus adapts because it may have adaptive value if he or she encounters a similar (i.e., a matched) anxiety-inducing environment after birth/later in life (i.e., early life experience may program the brain for life to come). However if an individual is more vigilant in all types of environments (e.g., also in more favorable, safe ones), this biased (prenatally acquired) behavior may tax or compromise emotional and/or neurocognitive functioning and anxiety symptoms and poorer language acquisition may result. In order to develop adequate self-regulation skills, it will be vital that parents (and other educators) understand this fearfulness and/or heightened stress reactivity and try to induce changes in this behavior. Neural circuitry, molecular profiles, and neurochemistry can be (positively) changed by experiences ; these changes will in turn influence subsequent behavior (Bai & Repetti, 2015; McEwen & Morrison, 2013). If therapeutic interventions are needed, they might target the underlying mechanism s that produce heightened stress reactivity as well as strategies that might prevent subsequent related behavioral problems or psychopathology (Schechter, 2012). Importantly, the fact that phenotype s are programmed by adaptation to early life environments does not imply that they cannot be changed! To refer to the amazing plasticity of the developing brain, Seymour Levine once said “Nothing is written in stone” (Levine, 2005); this statement has frequently been cited in early life stress literature (e.g., Daskalakis et al., 2013).

A strength of the studies used to generate our DOBHaD model, including the ones detailed above, is that they focused on processes (constituting elements) of behavior that could be measured in an objective, standardized way (e.g., changes in fetal behavioral state; event related potentials during sensory stimulation ; heart rate variability measures). The importance of the findings and the model based on this line of research are their potential for identifying those maternal–offspring factors that could serve as markers of later mental health issues. A recent Danish population based study indicates that a person exposed to prenatal stress makes more use of primary health care than a person not exposed to prenatal stress (Li, Yang, Guldin, Vested, & Vestergaard, 2015). However, only when markers and underlying mechanism s are identified, may interventions be developed that are targeted to prevention and/or amelioration of specific health issues. Such interventions may focus on the care given to the mother during pregnancy, the mother herself, or the offspring. Recent randomized controlled trial studies of interventions to treat anxiety and depression during pregnancy show not only improvements in maternal mental health but in infant outcome (e.g., better self-regulation and stress reactivity) as well (e.g., Milgrom et al., 2015).

The human and economic toll of mental health issues in the population is substantive. A recent UK report (Bauer, Parsonage, Knapp, Iemmi, & Adelaja, 2014) calculated the costs of mental health problems (i.e., depression , anxiety, psychosis, post-traumatic stress disorder) during pregnancy and the first year after childbirth at about £8.1 billion for each 1-year cohort of births and the equivalent of just under £10,000 for a single birth. About 72 % of this cost relates to adverse effects on the child rather than the mother. It seems that, not only for fundamental scientific studies, but also for clinical studies and for society as a whole, the study of prenatal environmental influences on offspring outcome is critical. It has considerable potential for improving behavior and health outcomes because maternal anxiety, stress and lifestyle are modifiable.