High Antenatal Maternal Anxiety Is Related to ADHD Symptoms, Externalizing Problems, and Anxiety in 8- and 9-Year-Olds
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Child Development, July/August 2004, Volume 75, Number 4, Pages 1085 – 1097
High Antenatal Maternal Anxiety Is Related to ADHD Symptoms,
Externalizing Problems, and Anxiety in 8- and 9-Year-Olds
Bea R. H. Van den Bergh and Alfons Marcoen
Associations between antenatal maternal anxiety, measured with the State Trait Anxiety Inventory, and dis-
orders in 8- and 9-year-olds were studied prospectively in 71 normal mothers and their 72 firstborns. Clinical
scales were completed by the mother, the child, the teacher, and an external observer. Hierarchical multiple
regression analyses showed that maternal state anxiety during pregnancy explained 22%, 15%, and 9% of the
variance in cross-situational attention deficit hyperactivity disorder symptoms, externalizing problems, and
self-report anxiety, respectively, even after controlling for child’s gender, parents’ educational level, smoking
during pregnancy, birth weight, and postnatal maternal anxiety. Anxiety at 12 to 22 weeks postmenstrual age
turned out to be a significant independent predictor whereas anxiety at 32 to 40 weeks was not. Results are
consistent with a fetal programming hypothesis.
It is recognized that both genetic and nongenetic phenotype (Gottlieb, 1976; Gottlieb & Halpern, 2002;
factors play an etiological role in commonly diag- Johnson, 1997) has important implications (e.g., for
nosed childhood disorders, such as attention deficit the study of early origins of childhood disorders). It
hyperactivity disorder (ADHD), externalizing prob- has long since been known that neurodevelopmental
lems, and internalizing problems (e.g., see Bradley & impairments seen in some of these disorders (e.g.,
Golden, 2001; Lonigan, Vasey, Phillips, & Hazen, impairments in attention, memory, response inhibi-
2004; Teichner & Golden, 2000). Because of the de- tion, processing speed) are caused by the deleterious
velopment of noninvasive tools to observe the fetus, effect of prenatal factors on developmental processes
techniques for observing and computing neural pro- of the brain. Studies in the field of behavioral
cesses, and concomitant conceptual changes, there teratology have shown that the pattern of resulting
has been an extraordinary progress in genetics, em- behavioral and cognitive effects in the offspring
bryology, and developmental neuroscience (Elman et is dependent on the stage of gestation at which
al., 1998). The knowledge that during postnatal as prenatal exposure occurs, as well as on the type,
well as during prenatal development the critical in- dose, and duration of exposure to the teratogen (e.g.,
teraction of genetic constraints with internal and tobacco, alcohol, cocaine and other drugs, lead, me-
external environmental influences gives rise to the thylmercury, viral and bacterial infections, nutri-
tional deficiencies) and on differences in species and
individual susceptibility (e.g., see Coyle, Wayner, &
Bea R. H. Van den Bergh and Alfons Marcoen, Department of Singer, 1976; Jacobson & Jacobson, 2000; Mayes,
Psychology, Catholic University of Leuven, Belgium. 2002). Because of scientific progress, old questions of
This study was started in the context of a prospective longitu- whether pregnant mothers’ emotional states in any
dinal Ph.D. research project on the effects of maternal stress and
anxiety during pregnancy on children’s behavior and develop- way ‘‘influence the conceptus in the womb’’ (Mon-
ment. This project was initiated at the Center for Developmental tagu, 1962, p. 169) or ‘‘may contribute to the shaping
Psychology of the Catholic University of Leuven supported by a of physical status, behavior patterns, and postnatal
grant from the Fund of Scientific Research Flanders, Belgium, and progress of the children they bear’’ (Sontag, 1941,
was continued at the Population and Family Study Center, Brus-
p. 1002) were in the last decades reframed into the
sels, Belgium, in collaboration with the Center for Developmental
Psychology of the Catholic University of Leuven. general question of whether negative maternal emo-
We are grateful to all of the mothers, children, and teachers who tions during pregnancy should be considered as be-
took part in the study; to Ilse Van Hauwaert and Merel Ackx for havioral teratogens. Can high levels of maternal
their assistance in data collection and processing and in preparing stress, anxiety, or depression, by triggering a cascade
reports, posters, and papers on this data; and to Maarten Mennes of physiological events in the mother, the placenta,
for stimulating discussions and carefully reading the final manu-
script. We greatly acknowledge Thomas O’Connor for his helpful and the fetus, deleteriously affect fetal and postnatal
comments on statistics and previous drafts of the manuscript. neurobehavioral development?
Correspondence concerning this article should be addressed to
Bea Van den Bergh, Department of Psychology, Catholic Univer-
sity of Leuven, Tiensestraat 102, 3000 Leuven, Belgium. Electronic r 2004 by the Society for Research in Child Development, Inc.
mail may be sent to Bea.Vandenbergh@psy.kuleuven.ac.be. All rights reserved. 0009-3920/2004/7504-00081086 Van den Bergh and Marcoen
In the present study we investigated whether high motor development at 3 and 8 months, whereas at 37
maternal anxiety during pregnancy enhances the off- to 38 weeks pma early morning cortisol levels had a
spring’s susceptibility for childhood disorders and smaller effect. In Brouwers, van Baar, and Pop (2001,
whether specific prenatal vulnerability periods exist. N 5 105) maternal anxiety was measured at 32 weeks
Early brain development is defined by a cascade pma and was significantly related to observer report
of orderly timed processes (Levitt, 2003; Nowakow- measures of attention at 3 weeks and 12 months of
ski & Hayes, 2002). Therefore, identifying periods age, as well as to mental development at 24 months
during which the fetus is especially sensitive to of age. Wadhwa and colleagues (Wadhwa et al., 2002;
the deleterious effect of high anxiety might not only Wadhwa, Sandman, & Garite, 2001) mentioned that
have implications for preventive and therapeutic in samples of 24 to 49 mother – infant dyads, mater-
interventions in highly anxious pregnant women but nal stress and stress hormones during pregnancy
also holds the potential of indirectly revealing which had an effect on difficult temperament in children up
processes of early brain development were dis- to 3 years of age. In all of these studies, the associa-
turbed. This might enhance our scant knowledge tions were found after accounting for confounding
(Rutter, 2001) on causal neural processes that con- variables, such as smoking during pregnancy, birth
stitute the biological underpinning of childhood weight, socioeconomic class, and postnatal maternal
disorders. anxiety or depression. We can conclude that, across
the five studies, the effects found are not confined to
a specific gestational period nor are they highly
Prospective Studies Relating Antenatal Maternal Anxiety
specific. However, the observation in two of the
to Infant and Child Behavior, and the Fetal Programming
studies, that the associations were strongest for ear-
Hypothesis in Animal and Epidemiological Research
ly- and mid-gestation anxiety, concur with findings
Up to now, only the Avon Longitudinal Study of of experimental studies with nonhuman primates
Parents and Children (ALSPAC; N 5 6,493; O’Con- (Schneider, Moore, Kraemer, Robert, & DeJesus,
nor, Heron, Golding, Beveridge, & Glover, 2002; 2002; Schneider, Roughton, Koehler, & Lubach,
O’Connor, Heron, Golding, Glover, & the ALSPAC 1999), and the evidence for a link with antenatal
Study Team, 2003) has prospectively studied the ef- anxiety is most conclusive for hyperactivity, inat-
fects of antenatal anxiety into childhood. High levels tention problems, and difficult temperament. Dif-
of maternal anxiety, at 32 weeks postmenstrual age ferences in the results of the five studies might be
(pma), doubled the risk for hyperactivity and inat- due to differences in: (a) the exact timing of the
tention problems, conduct disorder, and emotional anxiety measurements, the period to which they re-
problems at the ages of 47 and 81 months in both fer, the intensity of anxiety and the actual persistence
boys and girls. Maternal anxiety at 18 weeks pma of anxiety throughout pregnancy; (b) the scales used
affected total behavioral and emotional problems, to measure the independent and dependent varia-
but only in girls. This study relied on a short mother bles; (c) biological, psychological, and environmental
report screening instrument in which each of the factors in prenatal and postnatal life period control-
three problem scales was typically defined by five led for (covariates) and not controlled for; and (d)
items. Four research groups found effects on various genetic differences (Heron, O’Connor, Evans, Gol-
behavioral outcomes at earlier postnatal ages. Van ding, & Glover, 2004; Kofman, 2002; Mulder et al.,
den Bergh, Vandenberghe, Daniels, Casaer, and 2002).
Marcoen (1989; N 5 70) found that antenatal anxiety How can the results of the studies cited be ex-
explained 10% to 25% of the variance in neonatal plained? At this point we would like to introduce
movements and behavioral states in hyperactivity, the fetal programming hypothesis (or, more gener-
frequent crying, sleeping and feeding difficulties, ally, perinatal programming hypothesis), which is
and difficult temperament during the first 7 months derived from human epidemiologic and animal
of life. Associations were stronger for anxiety at 12 to model data. As extensively explained by O’Connor
22 weeks pma than for anxiety at 23 to 31 and 32 to (2003), this hypothesis ‘‘suggests a complementary
40 weeks pma (Van den Bergh, 1989, 1990, 1992). In and perhaps competing perspective to predominant
their study, Huizink, Robles de Medina, Mulder, developmental theories explaining effects of early
Visser, and Buitelaar (2002, 2003; N 5 170) perceived experiences on later psychological development’’
stress and pregnancy-related anxieties at 15 to 17 (O’Connor et al., 2003, p. 1034). The fetal program-
weeks pma, and at 27 to 28 weeks pma showed the ming hypothesis states that when disturbing factors
strongest effects on difficult temperament at 3 act during specific sensitive periods of development,
months and attention regulation and mental and they exercise organizational effectsFor programAntenatal Maternal Anxiety and Childhood Disorders 1087
some set pointsFin a variety of systems (e.g., met- observation scale on ADHD and externalizing and
abolic, cardiovascular, or immune system) by dif- internalizing problems in the 8- and 9-year-old child
ferent underlying mechanisms (Barker, 1998, 2002; were completed by the mother and the child, and by
Coe, Lubach, & Karaszweski, 1999; Nathanielsz, the child’s teacher and an external observer, both
1999). Problems will arise if the environment blinded to the level of maternal anxiety during
changes later in life. The organism is then hampered pregnancy. Potential confounders were taken into
because the programmed set points do not readily account, namely, gender of the child, parents’ edu-
readapt to the new environment. The malleability or cational level, smoking during pregnancy, birth
adaptive plasticity of biological systems is develop- weight for pma, and maternal postnatal anxiety. The
mentally constrained, and this produces maladap- findings of this study might bring indirect evidence
tive physiology and ultimately predisposes to for the fetal programming hypothesis and might
disease or disorder (Welberg & Seckl, 2001). For ex- contribute to the knowledge of early origins of
ample, when the body’s capacity to metabolize glu- childhood disorders.
cose was set according to nutritionally deprived
circumstances during prenatal life, the body is una- Method
ble to adjust when dramatic improvements in the
Participants
nutritional environment occur after birth, and dia-
betes results (Barker, 1998). The hypothalamic-pitu- Of the 86 women that participated at the start of
itary-adrenocortical (HPA) axis is particularly the study, 71 participated in the second wave, to-
susceptible to perinatal programming by gluco- gether with their 8- and 9-year-old firstborns, 38 boys
corticoids, at least in rodents. Antenatally stressed and 34 girls (2 girls were twins), with a mean age of 8
animals show prolonged endocrinologic stress re- years 6 months. Eligibility criteria at the start of the
sponses, as well as behavioral symptoms related to study were that the women should be Caucasian,
dysregulation of the HPA axis, such as anxiety, Dutch speaking, between the ages of 18 and 30, not
fearfulness and maladaptive coping in novel or yet have given birth to a viable child, between 12 and
stressful situations, high stress responsivity, declined 22 weeks pregnant, without obstetrical complica-
attention and impaired learning or memory, and tions or medical risks, and not using drugs or med-
neuromotor delays. Associated neurochemical ication with risks to the fetus. To recruit the sample,
changes in limbic areas include reduced glucocorti- the first author approached eligible women about
coid receptor expression in the hippocampus and volunteering for the study while at the Obstetrical
elevated expression of corticotropin releasing hor- and Gynecological Board consultations of the Uni-
mone (CRH) in the amygdala. Programming of de- versity Hospital Gasthuisberg in Leuven. The ethical
veloping neurotransmitter systems, for example, of committee for experiments on human beings of the
the monoaminergic (i.e., noradrenergic, do- Catholic University of Leuven approved the study.
paminergic, and serotonergic) system, which is All participants gave their informed consent.
functionally related to the HPA axis, may also con- Sociodemographic, psychiatric, and obstetric data
tribute to the observed changes. Disturbance in HPA were collected in an interview and by medical chart
axis regulation and brain monoamine levels are as- review. Although there was variation in the socio-
sociated with several disorders in humans (Her- demographic characteristics, the sample mainly con-
lenius & Lagercrantz, 2001; Herman et al., 2003; sisted of healthy, well-educated women, 94% of
Huizink, Mulder, & Buitelaar, 2004; Ladd et al., 2000; whom were married. The following four levels of
Maccari et al., 2003; Weinstock, 2001; Welberg & education were achieved by the parents: without
Seckl, 2001). high school diploma (mothers 5 13.8%, fathers 5
9.2%), high school graduate without additional
schooling (mothers 5 25.1%, fathers 5 33.8%), addi-
Research Aim
tional schooling beyond high school (moth-
The present study represents the second wave of ers 5 30.8%, fathers 5 12.3%), and university degree
the study of Van den Bergh and colleagues and in- (mothers 5 32.3%, fathers 5 44.6%). None of the
vestigates whether high maternal anxiety during women had a psychiatric disorder or had ever been
pregnancy enhances the offspring’s susceptibility for treated for a primary anxiety disorder. All pregnan-
childhood disorders and whether specific vulnerable cies were dated using the last menstrual period or a
periods exist. Mothers completed the State Trait sonographic examination before 20 weeks if dates
Anxiety Inventory (STAI) at 12 to 22, 23 to 31, and 32 had been uncertain. All babies were delivered in the
to 40 weeks pma. Clinical questionnaires and an hospital between 36 and 41 weeks pma with a mean1088 Van den Bergh and Marcoen
birth weight of 3301.48 g (SD 5 490.68), had 5-min than Anx 12 – 22 weeks pma (t 5 3.17, p 5 .00) and
APGAR scores of 9 or 10 (2 babies with score 8) and lower than Anx 32 – 40 weeks pma (t 5 1.56, p 5 .12),
scored high on Prechtl’s list of 52 optimal obstetrical rendering the possibility of midgestation anxiety
conditions (M 5 45.6, SD 5 2.9; Michaëlis, Dopfer, having a less plausible effect in comparison with
Gerbig, Dopfer-Feller, & Rohr, 1979). Anx 12 – 12 weeks pma or Anx 32 – 40 weeks pma. It
also turned out that correlations of Anx 23 – 31 weeks
Study Design pma with the final dependent variables were non-
The participants were included in a prospective significant (ps 4 .05).
follow-up study with six assessments in the first The trait anxiety subscale measures dispositional
wave, at 12 to 22, 23 to 31, and 32 to 40 weeks pma, anxiety or anxiety in general. Women with high
and at 1, 10, and 28 weeks after birth. In Wave 1 a anxiety levels were expected to differ from those
standardized continuous observation, which lasted with low anxiety levels in postnatal contact with
100 to 120 min, was made of behavioral states and their child and in caregiving. We controlled for this
movements at 36 to 38 weeks pma with ultrasound potential confounding effect by including a postnatal
and cardiography, as well as at the fifth or sixth day trait anxiety measure in our model. A principal
after birth (for a complete description of Wave 1, see component analysis (PCA) conducted on all post-
Van den Bergh, 1989, 1990, 1992). In Wave 2 during a natal trait anxiety measures (1, 10, and 28 weeks, and
home visit that lasted 90 to 120 min, a female ob- 8 – 9 years), revealed one component explaining 65%
server, blinded to the level of maternal anxiety dur- of the variance. A variable consisting of the stand-
ing pregnancy, interviewed the mother and assisted ardized component score for each mother was used
the child in completing standardized questionnaires. as a covariate in the analyses.
She also collected completed questionnaires that had
been sent to the mother before the home visit. Socioeconomic Status: Parents’ Educational Level
Questionnaires were sent with the child to his or her
Maternal and paternal education was coded on
teacher, who sent the completed questionnaire back
a 4-point scale (1 5 without high school diploma,
in a prepaid envelope.
2 5 high school graduate without additional school-
Instruments and Measures ing, 3 5 additional schooling beyond high school,
4 5 university degree). A PCA conducted on both
Maternal Anxiety scores yielded one component explaining 85% of the
The anxiety of the mother was measured using a variance.
Dutch version (Van der Ploeg, Defares, & Spiel-
berger, 1980) of the STAI, which is a self-report Obstetric and Antenatal Risk: Smoking During
questionnaire consisting of two subscales of 20 items, Pregnancy and Birth Weight
scored from 1 to 4. State anxiety is conceptualized as Two prenatal covariates likely to be related to
a transient emotional condition, characterized by the children’s behavioral and emotional outcomes
subjectively experienced tension, together with an were included in the statistical analyses: amount
enhanced activity of the autonomous nervous sys- of cigarettes smoked daily during pregnancy, where
tem. Trait anxiety reflects a disposition indicating 0 5 not smoking (n 5 54) and 1 5 smoking (n 5 18;
anxiety proneness. Cronbach’s alphas are .95 and .93 13 smoking less than 5, 3 smoking 10, 1 smoking 15,
for state and trait anxiety scales, respectively (Van and 1 smoking 25 cigarettes), and birth weight for
der Ploeg et al., 1980), and in our sample exceeded pma. The confounding influence of other obstetric
.90 across assessments. STAI data were available at and antenatal risk factors (e.g., pregnancy compli-
all six occasions in Wave 1 and at Wave 2. State cations, parity, medication use) was avoided by us-
anxiety measures provide a valid indication of ing strict eligibility criteria when recruiting the
change in the intensity of transitory anxiety in re- sample.
sponse to real-life stress (Spielberger, 1975). There-
fore, state (and not trait) anxiety measures were
Outcome Measures: Cross-Informant Composites of
included as predictors in our models investigating
Specific Childhood Disorders
the effects of antenatal maternal anxiety. The analy-
ses focused on state anxiety at 12 to 22 weeks pma To obtain robust outcome measures for ADHD
(Anx 12 – 22 weeks pma) and state anxiety at 32 to 40 symptoms and externalizing and internalizing
weeks pma (Anx 32 – 40 pma). Anx 23 – 31 weeks problems, composite measures were constructed by
pma (M 5 34. 94, SD 5 8.79) was significantly lower conducting PCAs on scales of self-administeredAntenatal Maternal Anxiety and Childhood Disorders 1089
clinical questionnaires and standardized observation Ploeg, & Spielberger, 1989), a 3-point anxiety scale,
scales completed by mother, teacher, observer, or with 20 items measuring state anxiety and 20 items
child. Standardized Dutch versions of the scales measuring trait anxiety, was completed by the child
were used; all Cronbach’s alphas were between .78 during the home visit.
and .86 in our sample. Composites. Four composites were constructed: (a)
Mother and teacher report measures. From the Child ADHD symptomsFa PCA on mother and teacher
Behavior Checklist (CBCL/4 – 18 years; Verhulst, van report scales Attentional Problems (CBCL/TRF) and
der Ende, & Koot, 1996) and the Teacher’s Report Hyperactivity/Inattention (CATRS) and on external
Form (TRF/4 – 18 years; Verhulst, van der Ende, & observer report scale (GBO) revealed one component
Koot, 1997), two widely used clinical standardized that explained 66% of the variance; (b) externalizing
formats (120 items scored from 0 to 2) for reporting problemsFa PCA on mother and teacher report
behavioral problems and competencies of children scales Externalizing Problems (CBCL/TRF) and
ages 4 to 18, we used the Attention Problems scale as Acting-Out (CATRS) resulted in one component that
a measure for ADHD symptoms and the broadband explained 66% of the variance; (c) internalizing
dimensions of Externalizing and Internalizing Prob- problemsFa PCA on mother and teacher report
lems. The Externalizing Problems Scale is composed scales Internalizing Problems (CBCL/TRF) and self-
of the Aggressive Behavior and Delinquent Behavior report anxiety scales (STAIC) did not reveal one
subscales. The Internalizing Problems Scale is component; therefore, a PCA was conducted on
composed of the Anxious/Depressed, Somatic mother and teacher scales Internalizing Problems
Complaints, and Withdrawn subscales. Scores are (CBCL/TRF ) separately, which revealed one com-
reported in the form of T scores (M 5 50, SD 5 10). ponent that explained 65% of the variance; and (d)
On all small-band scales of CBCL and TRF, including self-report anxietyFa PCA on the two scales of the
the Attention Problem Scale, T scores above 70 are STAIC revealed one component that explained 71%
considered to be clinically significant and T scores of the variance.
between 67 and 70 are considered borderline. For
Externalizing and Internalizing Problems, T scores
Data Analysis
above 63 are clinically significant and T scores be-
tween 60 and 63 indicate borderline pathology. From We had complete data sets for 52 children. The
the Conners Abbreviated Teacher Rating Scale data for the remaining 20 children were partially
(CATRS; Blöte & Curfs, 1986), a 4-point scoring scale missing at random, resulting in 10.9% of all data that
for measuring ADHD symptoms, we obtained were missing, most of which were teacher and
measures for two clusters: (a) hyperactivity/inat- mother report measures. Observer and child report
tention (four items; e.g., ‘‘is restless and overactive’’ data were complete, except for 1 child. The values of
‘‘is easily distracted’’) and (b) acting-out behavior the missing data were estimated through expectation
(five items; e.g., ‘‘disturbing behavior towards other maximization (EM) algorithm for maximum likeli-
children’’ ‘‘has temper tantrums’’). A score above 16 hood (ML) estimation as provided in SPSS (version
on the total scale and agreement between parent and 11). EM was based on all available data observed in
teacher regarding symptoms of inattention/hyper- Wave 2. In handling missing data, imputation tech-
activity indicate pervasive symptoms for ADHD niques are preferred over case deletion because re-
(van der Meere, Shalev, Börger, & Gross-Tsur, 1995). taining the full sample helps prevent loss of power
Observer report measure. The Groninger Behavior resulting from diminished sample size. Imputation
Observation Scale (GBO; Kalverboer, 1990) consists makes more efficient use of the observed dataFno
of 12 behavior descriptions, scored from 1 to 4, in- cases are sacrificedFand the apparently complete
dicating attention and hyperactivity problems (e.g., data set can be analyzed by standard methods,
‘‘easily gives up’’ ‘‘makes a lot of unnecessary whereas case deletion may induce bias (Schafer &
movements during the execution of the task’’). As Graham, 2002; Wothke, 2000).
children with ADHD are most problematic in their Four hierarchical regressions were performed on
behavior in a context involving task-directed per- the outcome measures. Five covariates (gender of the
sistence and behavioral restraint (Barkley, 2003), we child, parents’ educational level, smoking during
asked an external observer to administer scales to the pregnancy, birth weight, and postnatal maternal trait
child during the home visit and to observe his or her anxiety) were entered in the first step, and two pre-
behavior while completing the scales. dictors (Anx 12 – 22 weeks pma and Anx 32 – 40
Child report measure. The State Trait Anxiety Scale weeks pma) were entered in the second step. Inter-
for Children (STAIC; Bakker, Van Wieringen, Van der actions between gender (coded as 0 5 boys, 1 5 girls)1090 Van den Bergh and Marcoen
and each of the predictors were entered in a third coefficient of covariates and predictors was between
step. Because none of the interactions turned out to 64% and 95%, indicating that between 64% and 95%
be significant in any of the models, they were re- of the explained variance was independent of the
moved in all final models. This approach gave an variance explained by other variables. An alpha level
indication of the unique contribution of the anxiety of .05 was used for all statistical tests.
predictors in predicting childhood disorders. Two
cases were removed in the regression model for ex-
ternalizing problems. Based on their studentized Results
residuals (4 2.65) they were identified as outliers by
Descriptive Analyses
SPSS (see Fox, 1997). When conducting analyses on
the 52 children with complete data, the proportion of Table 1 shows the descriptive statistics of and
the variances explained by the predictors was correlations among predictors, covariates, and de-
slightly higher (about 2%) than in analyses on the pendent variables. Maternal anxiety scores in our
whole sample. sample covered the range of scores observed in a
Multicollinearity can pose serious threats to the nonclinical female Dutch community sample (Van
interpretation of regression coefficients as indexes of der Ploeg et al., 1980). The means of Anx 12 – 22
effects. To mitigate this problem we controlled for weeks pma and 32 – 40 weeks pma were situated,
the tolerance coefficient. In all analyses the tolerance respectively, at Decile 6 and Decile 5 of the Dutch
Table 1
Descriptives and Intercorrelations of Predictors, Covariates and Dependent Variables (N 5 72)
Pearson r
Variable M SD Range 2 3 4 5 6 7 8 9 10
Predictors
1. Anx 12 – 22 weeks pma 38.70 7.75 24.00 to 59 .25 .09 .06 .07 .31 .51 .42 .18 .23w
2. Anx 32 – 40 weeks pma 36.14 8.79 20.00 to 58.00 .33 .21 .17 .57 .02 .22w .28 .10
Covariates
3. Parents’ educational level 1 0 2.21 to 1.69 .07 .06 .50 .19 .03 .16 .06
4. Smoking in pregnancy 1.42 0 0 to 25 .28 .22 .04 .05 .03 .06
5. Child’s birth weight 3301.48 490.68 2245 to 4220 .14 .10 .05 .06 .15
6. Postnatal trait anxiety 1 0 2.09 to 2.24 .13 .14 .27 .01
Dependent variables
7. C ADHD symptoms 1 0 1.26 to 3.31 .68 .35 .15
Attention problems CBCL 3.67 2.91 0 to 11
Attention problems TRF 5.09 5.59 0 to 27
Hyper/inatt CATRS Mo 7.34 2.80 4 to 15
Hyper/inatt CATRS Te 6.08 2.35 4 to 14
GBO observer 22.26 6.18 15 to 39
8. C externalizing problems 1 0 1.19 to 3.70 .33 .11
Externalizing CBCL 7.53 5.70 0 to 24
Externalizing TRF 4.58 7.28 0 to 38
Acting-out CATRS Mo 7.90 2.43 5 to 18
Acting-out CATRS Te 5.45 1.68 4 to 11
9. C internalizing problems 1 0 1.31 to 2.94 .10
Internalizing CBCL 7.01 5.02 0 to 21
Internalizing TRF 4.72 4.24 0 to 21
10. C self-report anxiety 1 0 2.51 to 2.72
State anxiety STAIC 29.60 4.32 20 to 45
Trait anxiety STAIC 32.75 6.31 20 to 48
Note. Anx 5 maternal state anxiety; pma 5 postmenstrual age; C 5 component resulting from principal component analysis;
ADHD 5 attention deficit hyperactivity disorder; CBCL 5 Child Behavior Checklist; TRF 5 Teacher’s Report Form; Hyper/inatt 5
hyperactivity/inattention; CATRS 5 Conners Abbreviated Teacher Rating Scale; Mo 5 mother; Te 5 teacher; GBO 5 Groninger Behavior
Observation Scale; STAIC 5 State Trait Anxiety Scale for Children.
wpo.10. po.05. po.01. po.001.Antenatal Maternal Anxiety and Childhood Disorders 1091
Table 2
Summary of Hierarchical Regression Analysis Predicting Childhood Disorders in 8- and 9-Year-Olds (N 5 72)
C externalizing C internalizing C self-report
C ADHD symptoms problemsa problems anxiety
Childhood disorders
Covariates and predictors b SE B b SE B b SE B b SE B
Step 1: Covariates
Gender child .22 .21 .03 .22 .04 .25 .04 .24
Parents’ educational level .20 .10 .07 .10 .10 .12 .05 .12
Smoking in pregnancy .07 .23 .24 .24 .07 .27 .29 .26
Birth weight child .13 .10 .24 .11 .14 .12 .13 .16
Postnatal trait anxiety .05 .13 .11 .13 .17 .15 .02 .14
R2 for Step 1 .15 .13 .11 .10
Step 2: Predictors
Anx 12 – 22 weeks pma .48 .01 .36 .01 .10 .02 .29 .02
Anx 32 – 40 weeks pma .25 .02 .20 .02 .14 .02 .22 .02
DR2 for Step 2 .22 .15 .02 .09
FModel 5.47 3.40 1.0 1.39 2.18
Note. ADHD 5 attention deficit hyperactivity disorder; Anx 5 maternal state anxiety; pma 5 postmenstrual age; C 5 component resulting
from principal component analysis.
a
n 5 70; two outliers were dropped.
po.05. po.01. po.001.
sample, indicating that the mean level of Anx 12 – 22 The prenatal and postnatal maternal anxiety
weeks pma in our sample was slightly higher than measures were significantly correlated. Correlations
the level of anxiety in normal, nonpregnant women. with and between the covariates showed that Anx
Anx 12 – 22 weeks pma was significantly higher than 32 – 40 weeks pma was negatively correlated with
Anx 32 – 40 weeks pma (t 5 2.11, p 5 .039). In early parents’ educational level, that smoking during
gestation, 40% of the women scored between 40 and pregnancy and child’s birth weight were negatively
59 on Anx 12 – 22 weeks pma; in late gestation 29% of correlated, and that postnatal maternal trait anxiety
the women scored similarly on Anx 32 – 40 weeks was correlated with internalizing problems of the
pma. It is interesting to note that mothers of boys child. All dependent variables but self-report anxiety
(M 5 40.42, SD 5 8.38) scored higher on Anx 12 – 22 were significantly intercorrelated.
weeks pma than did mothers of girls (M 5 36.77,
SD 5 6.57), t(72) 5 2.043, p 5 .045. Of the women who
Maternal Anxiety During Pregnancy and Childhood
scored above the 75th percentile (score 5 43) on Anx
Disorders at Ages 8 and 9
12 – 22 weeks pma, 11 were carrying a boy, and 5
were carrying a girl. Results of the multiple hierarchical regressions
For 11.11% of the children (6 boys, 2 girls) on (see Table 2) revealed that, even after controlling for
either CBCL or TRF, scores concerning attention child’s gender, parents’ educational level, smoking,
problems indicated at least borderline pathology and birth weight, and postnatal maternal anxiety, pre-
6.94% had clinical problems. On the CATRS, 9.7% natal maternal anxiety was significantly associated
met the criteria for pervasive ADHD symptoms, and with ADHD symptoms, externalizing problems, and
when combining CBCL, TRF, and CATRS data, 14% self-report anxiety in the 8- and 9-year-old offspring.
(7 boys, 3 girls) exhibited clinical ADHD symptoms. The proportion of the variance explained uniquely
For externalizing problems, 9.3% (TRF) to 14.8% by the predictors was significant (22%, 15%, and 9%,
(CBCL) of the children showed at least borderline respectively). In our sample, Anx 12 – 22 weeks pma
pathology, and between 6.9% (CBCL) and 7.4% was a significant independent predictor of these
(TRF) scored in the clinical range. For internalizing disorders; this was not the case for Anx 32 – 40 weeks
problems, 9.3% (TRF) to 24.1% (CBCL) of the chil- pma. No significant associations with maternal
dren showed at least borderline pathology, and be- anxiety during pregnancy emerged in the multiple
tween 5.6% (TRF) and 14.8% (CBCL) had clinical regression for internalizing problems, despite a sig-
problems. nificant correlation of Anx 32 – 40 weeks pma (see1092 Van den Bergh and Marcoen
Table 1). Concerning the covariates, boys scored our study are consistent with a fetal programming
higher for ADHD symptoms than did girls, smoking hypothesis in at least two ways.
during pregnancy and being small for postmenstrual First, evidence is found in the timing of antenatal
age at birth were associated with externalizing maternal anxiety enhancing childhood disorder
problems, and smoking during pregnancy was re- susceptibility. Between 12 and 22 weeks pma, the
lated to self-report anxiety in 8- and 9-year-olds. brain is susceptible to programming because many
areas are undergoing active growth and neurons are
still immature (Welberg & Seckl, 2001); that is, the
Discussion
cascade of orderly timed processes of neuron pro-
The present study shows that a high level of mater- liferation; migration; early differentiation; and,
nal state anxiety during pregnancy enhances the sometimes, cell death is taking place. Although re-
offspring’s susceptibility for developing a childhood gion-by-region differences exist in the exact time
disorder. Even after controlling for the child’s gen- course of these processes, in humans most cerebral
der, parents’ educational level, birth weight, smok- cortex neurons are generated and have at least
ing during pregnancy, and postnatal maternal started their migration between Embryonic Days 40
anxiety, maternal state anxiety during pregnancy and 125 (ca 8 – 20 weeks pma or 6 – 18 weeks after
explained between 9% and 22% of the variance in the conception; Levitt, 2003; Nowakowski & Hayes,
childhood disorders investigated. The level of Anx 2002; Rakic, 2002a, 2002b). For example, the limbic
12 – 22 weeks pma was slightly higher than the state system (e.g., the hippocampus, amygdala) and
anxiety level in a normal, nonclinical sample, and limbic regions of the cortex (e.g., anterior cingulate
significantly higher than Anx 32 – 40 weeks pma. In cortex [ACC]) start differentiating already in the 3rd
contrast to Anx 32 – 40 weeks pma, Anx 12 – 22 weeks month of gestation, and the neocortex starts only in
pma turned out to be a significant independent the 6th month (Johnson, 1997). Neurogenesis, neu-
predictor of three of the four childhood disorders ronal migration, and differentiation are controlled by
studied. This result at least suggests that the period processes intrinsic to the cell and processes extrinsic
between 12 and 22 weeks pma is a vulnerable period to the cell, that is, by genes and their products, by
for developing cross-situational ADHD symptoms, interactions between cells, and by interactions of
externalizing problems, and anxiety during child- cells with early neurotransmitters and neuromodu-
hood. However, it cannot be excluded that the vul- lators acting as growth factors (Nowakowski &
nerability period starts even before 12 weeks pma or Hayes, 2002). It is important to note that although
covers only a subperiod of the 12- to 22-week pma before midgestation these developmental processes
period, or that other measures of anxiety (e.g., of are not driven by activity that is modulated by sen-
pregnancy related anxiety) or higher levels of late sory input, they nevertheless can be altered (Bour-
gestation state anxiety than those observed in the geois, 1997). This happens when environmental
present sample, might enhance the susceptibility for factors modulate the influence of intracellular and
the childhood disorders investigated. The results of extracellular developmental signals (Herlenius &
the present study strengthen the evidence estab- Lagercrantz, 2001; Levitt, Reinoso, & Jones, 1998;
lished in the five prospective studies mentioned in Weaver et al., 2001). In general, the earlier the dis-
the Introduction for the association between high turbance occurs, the greater is its potential influence
levels of maternal anxiety during pregnancy and on subsequently occurring events and maturing and,
behavioral, emotional, and cognitive problems in finally, on the mature structure – function relation-
infancy and childhood. In what follows these results ship (Nowakowski & Hayes, 2002). If, for instance, in
are discussed in the framework of the fetal pro- the 3rd or 4th month of gestation, a teratogen mod-
gramming hypothesis and a few strengths, limita- ulates the influence of other developmental signals
tions, and implications of the present study are and disrupts neuronal migration, this may result in
briefly described. abnormality in cell position, which may alter the
laminar structure of the ACC. During the onset of
differentiation, disturbances by teratogens such as
The Fetal Programming Hypothesis Perspective
glucocorticoids (cortisol in humans) can, for exam-
Fetal programming reflects the action of a dis- ple, influence the growth of dendrites, disturb the
turbing factor during sensitive periods of develop- timetable of the expression of several neurotrans-
ment to exercise organizational effects that constrain mitter phenotype (e.g., of the monoaminergic sys-
the malleability of biological systems and ultimately tem) and neuropeptides (e.g., CRH) and their
predispose to disorder (Seckl, 2001). The results of receptors and alter the receptor sensitivity, whichAntenatal Maternal Anxiety and Childhood Disorders 1093
finally alters the balance between excitatory and in- and subcortical) modes of functioning may be per-
hibitory brain circuits (Ladd et al., 2000; Schneider et manently altered as a result of fetal programming by
al., 1998; Weaver et al., 2001; Weinstock, 2001). prenatal cocaine exposure and other early stressful
In humans, subtle aberrations in neuronal migra- experiences. In the same line, we tentatively hy-
tion and early differentiation may play a role in pothesized that in our sample, in children of highly
disorders such as ADHD, dyslexia, epilepsy, autism, anxious women, monoaminergic brain circuits (and
and some forms of schizophrenia (see Fernandez- the HPA axis and limbic system) may have been
Duque & Posner, 2001; Monk, Webb, & Nelson, 2001; prenatally programmed by the disturbing influence
Nowakowski & Hayes, 2002). It is postulated that in of anxiety-related hormones between 12 and 22
general the disturbance of the particular develop- weeks pma. ADHD symptoms (inattention, imp-
mental processes taking place in specific brain areas ulsivity, disinhibition, hyperactivity), externalizing
at the time of the prenatal insult, in interaction with behavior (conduct disorders, aggression), and self-
the genetic susceptibility of the offspring and medi- report anxiety (emotional inhibition and lability, high
ated by later prenatal and postnatal environmental subjective distress in novel situations) may be seen
factors, will determine the way perceptual, cognitive, as different developmental progressive forms of
motor, arousal, and emotional processes are inte- early impaired self-regulation (cf. Posner & Rothbart,
grated, thereby influencing the kind of disorder the 2000; Rubia, 2002).
child might develop (Grossman et al., 2003; Ladd et A second source of evidence for the programming
al., 2000). For instance, it is known that the ACC, in hypothesis is generated when combining the results
which dopamine, noradrenaline, and CRH are pre- of both waves of our study. A strength of the first
sent in abundance, plays an important role in self- wave of our study was that, based on the results of a
regulation by integrating functions of cortical and pilot study (Van den Bergh, Mulder, et al., 1989), we
subcortical areas such as attention, executive func- started to observe the effects of maternal anxiety al-
tioning, motor control, drive, arousal regulation, ready in the prenatal life period. Pad models showed
awareness of emotions, and subjective distress that the effect of antenatal maternal anxiety on fetal
(Damasio, 1998; Paus, 2001, Posner & Raichle, 1994). general movements and length of quiet sleep periods
To our opinion, results of at least two studies make it at 36 to 38 weeks pma was reflected in comparable
plausible that alteration of ACC structure – function behavioral measures at Days 5 to 6 after birth. Neo-
relationships, caused by early teratogens, may have natal maternal anxiety had no direct effect on these
an etiological role in the development of childhood measures (Van den Bergh, 1989, 1990). Clearly, these
disorder observed in our study. First, in humans, analyses suggest that maternal emotions affect fetal
with a diagnosis of bipolar disorders or schizophre- neurobehavioral development, altering the devel-
nia, a change in the density of neurons in particular opmental trajectory from prenatal life onward.
layers of the ACC has been verified in postmortem Moreover, the proportion of variance explained by
cases (Benes, Vincent, & Todtenkopf, 2001). Second, antenatal maternal anxiety was stronger than the
in rabbits early prenatal exposure to cocaine led to proportion explained by postnatal maternal anxiety
anomalous dendritic development of ACC neurons, and did not change from between 36 to 38 weeks
causing impairment of attentional processes (Stan- pma and 8 to and 9 years after birth. These results all
wood & Levitt, 2004). Observation with functional are consistent with a fetal programming hypothesis.
magnetic resonance imaging of the differential pat-
terns of activation of cortical and subcortical areas in
Strengths and Limitations of the Study
children with and without ADHD (e.g., Durston
et al., 2003; Rubia, 2002) at least concur with a hy- First, one strength of our study is the prospective
pothesized role of ACC structure – function rela- design with a retention rate of 84% in the second
tionships in ADHD. However, longitudinal studies wave, over about 9 years. Second, maternal anxiety
linking the patterns of activation to prenatal pro- was measured repeatedly during gestational periods
gramming are lacking. Mayes (2000, 2002) developed that are assumed to be critical windows for brain
a multilevel gate model in which the impaired self- development and at different postnatal periods. The
regulation of children prenatally exposed to cocaine range of anxiety in a normal, nonclinical population
is understood as an imbalance between and within was covered in our sample, and a substantial pro-
dopaminergic and noradrenergic systems in cortical portion of the women scored in the higher range.
and subcortical areas. She hypothesized that the Third, robust cross-informant measures of childhood
threshold for switching between more executive disorders, based on standardized clinical scales and
(prefrontal) and more automatic (posterior frontal an observation scale, were used. This robustness1094 Van den Bergh and Marcoen
makes our results difficult to explain by a simple predominant theories of development and psycho-
rater bias explanation. pathology. Converging evidence from other studies
Some limitations and weaknesses of the study in identifying specific programming windows dur-
must be mentioned. First, the sample was relatively ing gestation could give the impetus to develop
small and the results might be sample specific. Sec- prevention, intervention, and support programs for
ond, the inclusion of physiological measures (e.g., highly anxious pregnant women. These programs
the level of cortisol in the mother) would have ena- could include stress-reduction instructions (e.g.,
bled us to test their assumed causative role in fetal Urizar et al., in press) and cognitive – behavioral
programming. Third, our design was not genetic treatments to reduce anxiety and neuroendocrine
sensitive. It is, however, unlikely that the associa- reactions to stress from early gestation on, or even
tions we found can be explained by genetic media- before conception (e.g., Facchinetti, Tarabusi, & Vol-
tion only because this does not explain why the pe, 2004). They may not only benefit the mother but
associations established in our sample specifically also the child to be born.
involve prenatal maternal anxiety and not postnatal In the future, we plan to investigate the long-
anxiety (cf. O’Connor et al., 2002; O’Connor et al., lasting effects of prenatal maternal anxiety. When the
2003). Of course, it is plausible that highly anxious children are 14 and 15 years old, temperament, be-
mothers are carrying genes for ADHD and that havioral, cognitive, and mood disorders will be
physiological events involved in high antenatal ma- studied and related to neuropsychological measures
ternal anxiety trigger gene regulatory mechanisms of sustained attention, inhibitory control and im-
and the expression of ADHD genes in the offspring. pulsivity, and measures of cortisol and heart rate. If
If this were the case, it would only underline the effects are found, they could broaden our knowledge
importance of this prenatal environmental factor of fetal programming and the origins of disorders in
(Gottlieb & Halpern, 2002; Grossman et al., 2003; adolescence.
Rutter, Silberg, O’Connor, & Simonoff, 1999).
References
Implications and Future Research
Bakker, F. C., Van Wieringen, P. C. W., Van der Ploeg, H. M.,
The present study yielded some of the strongest & Spielberger, C. D. (Eds.). (1989). Handleiding bij de Zelf-
indirect evidence available today that antenatal Beoordelings Vragenlijst voor Kinderen, ZBV-K [Manual for
maternal anxiety may contribute to the shaping of the Self-Evaluation Questionnaire for Children, STAIC].
alteration in the child’s neurobehavioral develop- Lisse, Netherlands: Swets & Zeitlinger.
mentFand, hence, enhance the susceptibility for Barker, D. J. P. (1998). In utero programming of chronic
childhood disordersFby programming some set disease. Clinical Science, 95, 115 – 128.
points in early developing brain structure – function Barker, D. J. P. (2002). Fetal programming of coronary
relationships. Clearly, many questions on exactly heart disease. Trends in Endocrinology & Metabolism, 13,
how fetal programming works in humans, and in 364 – 368.
Barkley, R. A. (2003). Issues in the diagnosis of attention-
what ways the timing, kind, and duration of envi-
deficit/hyperactivity disorder in children. Brain & De-
ronmental disturbances are finally related to altered
velopment, 25, 77 – 83.
neurobehavioral development, are still unresolved. Benes, F. M., Vincent, S. L., & Todtenkopf, M. (2001). The
For example, how can the intriguing links among density of pyramidal and nonpyramidal neurons in
early gestation maternal anxiety, gender of the off- anterior cingulate cortex of schizophrenic and bipolar
spring, and incidence of ADHD found in our sample subjects. Biological Psychiatry, 50, 395 – 406.
be explained? Are mothers of male fetuses more Blöte, A. W., & Curfs, L. M. G. (1986). Het gebruik van de
anxious because boys introduce testosterone and Conners Teacher Rating Scale in Nederland [The use of
corresponding physiological changes into their the Conners Teacher Rating Scale in the Netherlands].
mother’s body? Do these changes enhance feelings Nederlands Tijdschrift voor de Psychologie, 41, 226 – 236.
of anxiety and concomitant release of hormones (e.g., Bourgeois, J. P. (1997). Synaptogenesis, heterochrony and
epigenesist in the mammalian neocortex. Acta Paediat-
cortisol) in the mother? Is it the interaction between
rica, 422, 27 – 33.
testosterone and cortisol (McEwen, 2000; Weinstock,
Bradley, J. D. D., & Golden, C. J. (2001). Biological contri-
2001) that explains the higher incidence of ADHD in butions to the presentation and understanding of at-
boys than in girls? tention-deficit/hyperactivity disorder: A review. Clinical
We agree with O’Connor et al. (2003, p. 1034) that, Psychological Review, 21, 907 – 929.
theoretically, fetal programming offers a comple- Brouwers, E. P. M., van Baar, A. L., & Pop, V. J. M. (2001).
mentary and perhaps competing perspective to the Maternal anxiety during pregnancy and subsequent in-Antenatal Maternal Anxiety and Childhood Disorders 1095 fant development. Infant Behavior and Development, 24, Huizink, A. C., Robles de Medina, P. R., Mulder, E. J. H., 95 – 106. Visser, G. H. A., & Buitelaar, J. K. (2002). Psychological Coe, C. L., Lubach, G. R., & Karaszweski, J. W. (1999). measures of prenatal stress as predictor of infant tem- Prenatal stress and immune recognition of self and perament. Journal of the American Academy of Child and nonself in the primate neonate. Biology of the Neonate, 76, Adolescent Psychiatry, 41, 1078 – 1085. 301 – 310. Huizink, A. C., Robles de Medina, P. G., Mulder, E. J. H., Coyle, I., Wayner, M. J., & Singer, A. G. (1976). Behavioral Visser, G. H. A., & Buitelaar, J. K. (2003). Stress during teratogenesis: A critical evaluation. Pharmacology, Bio- pregnancy is associated with developmental outcome in chemistry & Behavior, 4, 191 – 200. infancy. Journal of Child Psychology and Psychiatry, 44, Damasio, A. R. (1998). Emotion in the perspective of an 810 – 818. integrated nervous system. Brain Research Review, 26, Jacobson, S. W., & Jacobson, J. L. (2000). Teratogenic insult 83 – 86. and neurobehavioural function in infancy and child- Durston, S., Tottenham, N. T., Thomas, K. M., Davidson, M. hood. In C. A. Nelson (Ed.), The effects of early adversity on C., Eigsti, I.-M., Yang, Y., et al. (2003). Differential neurobehavioral development. The Minnesota Symposia on patterns of striatal activation in young children with Child Psychology, 31 (pp. 61 – 112). Mahwah, NJ: Erlbaum. and without ADHD. Biological Psychiatry, 53, Johnson, M. H. (1997). Developmental cognitive neuroscience: 871 – 878. An introduction. Cambridge, MA: Blackwell. Elman, J. F., Bates, E. A., Johnson, M. H., Karmiloff-Smith, Kalverboer, A. F. (Ed.). (1990). Developmental biopsychology. A., Parisi, D., & Plunkett, K. (1998). Rethinking innateness: Ann Arbor: University of Michigan Press. A Connectionist perspective on development. Cambridge, Kofman, O. (2002). The role of prenatal stress in the etio- MA: MIT Press. logy of developmental behavioral disorders. Neurosci- Facchinetti, F., Tarabusi, M., & Volpe, A. (2004). Cognitive- ence and Biobehavioral Review, 26, 457 – 470. behavioral treatment decreases cardiovascular and ne- Ladd, C. O., Huot, R. L., Thrivikraman, K. V., Nemeroff, C. uroendocrine reaction to stress in women waiting for B., Meaney, M. J., & Plotsky, P. M. (2000). Long-term assisted reproduction. Psychoneuroendocrinology, 29, behavioral and neuroendocrine adaptations to adverse 162 – 173. and early experiences. Progress in Brain Research, 122, Fernandez-Duque, D., & Posner, M. I. (2001). Brain imag- 81 – 103. ing and attentional networks in normal and pathological Levitt, P. (2003). Structural and functional maturation of states. Journal of Clinical and Experimental Neuropsycholo- the developing primate brain. Journal of Pediatrics, 143, gy, 23, 74 – 93. S35 – S45. Fox, J. (1997). Applied regression analysis, linear models and Levitt, P., Reinoso, B., & Jones, L. (1998). The critical impact related methods. Thousand Oaks, CA: Sage. of early environment on neuronal development. Pre- Gottlieb, G. (1976). Conceptions of prenatal development: Be- ventive Medicine, 27, 180 – 183. havioral embryology. Psychological Review, 83, 215 – 234. Lonigan, C. J., Vasey, M. W., Phillips, B. M., & Hazen, R. A. Gottlieb, G., & Halpern, C. T. (2002). A relational view of (in press). Temperament, anxiety, and the processing of causality in normal and abnormal development. Devel- threat-relevant stimuli. Journal of Clinical Child and Ado- opment and Psychopathology, 14, 421 – 435. lescent Psychology. Grossman, A. W., Churchill, J. D., McKinney, B. C., Kodish, Maccari, S., Darnaudery, M., Morley-Fletcher, S., Zuena, A. I. M., Otte, S. L., & Greenough, W. T. (2003). Experience R., Cinque, C., & Van Reeth, O. (2003). Prenatal stress effects on brain development: Possible contributions to and long-term consequences: implications of gluco- psychopathology. Journal of Child Psychology and Psychi- corticoid hormones. Neuroscience and Biobehavioral Re- atry, 44, 33 – 63. views, 27, 119 – 127. Herlenius, E., & Lagercrantz, H. (2001). Neurotransmitters Mayes, L. C. (2000). A developmental perspective on the and neuromodulators during early human develop- regulation of arousal states. Seminars in Perinatology, 24, ment. Early Human Development, 65, 21 – 37. 267 – 279. Herman, J. P., Figueiredo, H., Keller, N. K., Ulrich-Lai, Y., Mayes, L. C. (2002). A behavioral teratogenic model of Ostrander, M. M., Choi, D. C., et al. (2003). Central the impact of prenatal cocaine exposure on arousal mechanisms of stress integration: Hierarchical circuitry regulatory systems. Neurotoxicology and Teratology, 24, controlling hypothalamo-pituitatary-adrenocortical re- 385 – 395. sponsiveness. Frontiers in Neuroendocrinology, 24, 151 – 180. McEwen, B. S. (2000). The neurobiology of stress: From Heron, J., O’Connor, T. G., Evans, J., Golding, J., & Glover, serendipity to clinical relevance. Brain Research, 886, V. (2004). The course of anxiety and depression through 172 – 189. pregnancy and the postpartum in a community sample. Michaëlis, R., Dopfer, R., Gerbig, W., Dopfer-Feller, P., & Journal of Affective Disorders, 80, 65 – 73. Rohr, M. (1979). Die Erfassung obstretischer und post- Huizink, A. C., Mulder, E. J. H., & Buitelaar, J. K. (2004). natale Risicofaktoren durch eine Liste optimaler Bedin- Prenatal risk for psychopathology: Specific effects or gungen [The registration of obstetrical and postnatal induction of general susceptibility? Pyschological Bulle- risk factors with a list of optimal conditions]. Mon- tin, 130, 115 – 142. atschrift für Kinderheilkunde, 127, 149 – 155.
1096 Van den Bergh and Marcoen Monk, C. S., Webb, S. J., & Nelson, C. A. (2001). Prenatal Rutter, M., Silberg, J., O’Connor, T., & Simonoff, E. (1999). neurobiological development: Molecular mechanisms Genetics and child psychiatry: I. Advances in quantita- and anatomical change. Developmental Neuropsychology, tive and molecular genetics. Journal of Child Psychology 19, 211 – 236. and Psychiatry and Allied Disciplines, 40, 3 – 18. Montagu, A. (1962). Prenatal influences. Springfield, IL: Schafer, J. L., & Graham, J. W. (2002). Missing data: Thomas. Our view of the state of the art. Psychological Methods, 7, Mulder, E. J. H., Robles de Medina, P. G., Huizink, A. C., 147 – 177. Van den Bergh, B. R. H., Buitelaar, J. K., & Visser, G. H. Schneider, M. L., Clarke, A. S., Kraemer, G. W., Rougthon, A. (2002). Prenatal maternal stress: Effects on pregnancy E. C., Lubach, G. R., Rimm-Kauffman, S., et al. (1998). and the (unborn) child. Early Human Development, 70, Prenatal stress alters brain biogenic amine levels in 3 – 14. primates. Development and Psychopathology, 10, 427 – 440. Nathanielsz, P. W. (1999). Life in the womb. The origin of Schneider, M. L., Moore, C. F., Kraemer, G. W., Robert, A. health and disease. Ithaca, NY: Promethan Press. D., & DeJesus, O. T. (2002). The impact of prenatal stress, Nowakowski, R. S., & Hayes, N. L. (2002). General prin- fetal alcohol exposure, or both on development: Per- ciples of CNS development. In M. H. Johnson, Y. Mu- spectives from a primate model. Psychoneuroendocrino- nakata, & R. O. Gilmore (Eds.), Brain development and logy, 27, 285 – 298. cognition. A reader (2nd ed., pp. 57 – 82). Malden, MA: Schneider, M. L., Roughton, E. C., Koehler, A. J., & Lubach, Blackwell. G. R. (1999). Growth and development following pre- O’Connor, T. G. (2003). Early experiences and psychologi- natal stress exposure in primates: An examination cal development: Conceptual questions, empirical il- of ontogenetic vulnerability. Child Development, 70, lustrations, and implications for intervention. 263 – 274. Development and Psychopathology, 15, 671 – 690. Seckl, J. R. (2001). Glucocorticoid programming of the fe- O’Connor, T. G., Heron, J., Golding, J., Beveridge, M., & tus: Adult phenotypes and molecular mechanisms. Glover, V. (2002). Maternal antenatal anxiety and chil- Molecular and Cellular Endocrinology, 185, 61 – 71. dren’s behavioral/emotional problems at 4 years. British Sontag, L. W. (1941). The significance of fetal environ- Journal of Psychiatry, 180, 502 – 508. mental differences. American Journal of Obstetry and O’Connor, T. G., Heron, J., Golding, J, Glover, V., & the Gynaecology, 124, 996 – 1003. ALSPAC Study Team. (2003). Maternal anxiety and be- Spielberger, C. D. (1975). The measurement of State and havioural/emotional problems in children: A test of a Trait Anxiety: Conceptual and methodological issues. In programming hypothesis. Journal of Child Psychology and L. Levi (Ed.), Emotions: Their parameters and measurement Psychiatry, 44, 1025 – 1036. (pp. 713 – 725). New York: Raven. Paus, T. (2001). Primate anterior cingulate cortex: Where Stanwood, G. D., & Levitt, P. (2004). Drug exposure early in motor control, drive and cognition interface. Nature Re- life: Functional repercussions of changing neurophar- views Neuroscience, 2, 417 – 424. macology during sensitive periods of brain develop- Posner, M. I., & Raichle, M. E. (1994). Networks of atten- ment. Current Opinion in Pharmacology, 4, 65 – 71. tion. In M. I. Posner & M. E. Raichle (Eds.), Images of Teichner, G., & Golden, C. J. (2000). The relationship of the mind (pp. 183 – 179). New York: Scientific American neuropsychological impairment to conduct disorder in Library. adolescence: A conceptual review. Aggression and Violent Posner, M. I., & Rohtbart, M. K. (2000). Developing Behavior, 5, 509 – 528. mechanisms of self-regulation. Development and Psycho- Urizar, G. G. Jr., Milazzo, M., Le, H.-N., Delucchi, K., pathology, 12, 427 – 441. Sotelo, R., & Munoz, R. F. (in press). Impact of stress Rakic, P. (2002a). Intrinsic and extrinsic determinants of reduction instructions on stress and cortisol levels dur- neocortical parcellation: A radial unit model. In M. H. ing pregnancy. Biological Psychology. Johnson, Y. Munakata, & R. O. Gilmore (Eds.), Brain Van den Bergh, B. R. H. (1989). De emotionele toestand van de development and cognition. A reader (2nd ed., pp. 83 – 100). (zwangere) vrouw, obstretische complicaties en het gedrag en Malden, MA: Blackwell. de ontwikkeling van de foetus en van het kind tot de leeftijd Rakic, P. (2002b). Pre- and post-developmental neurogen- van zeven maanden [The emotional state of the (pregnant) esis in primates. Clinical Neuroscience Research, 2, woman, obstetrical complications and the behavior and 29 – 39. development of fetus and child until seven months after Rubia, K. (2002). The dynamic approach to neurodevel- birth]. Unpublished doctoral dissertation, Catholic opmental psychiatric disorders: Use of fMRI combined University Leuven, Belgium. with neuropsychology to elucidate the dynamics of Van den Bergh, B. R. H. (1990). The influence of mater- psychiatric disorders, exemplified with and without nal emotions during pregnancy on fetal and neo- ADHD. Behavioural Brain Research, 130, 47 – 56. namshytal behavior. Pre- and Perinatal Psychology Rutter, M. (2001). Child psychiatry in the era following Journal, 5, 119 – 130. sequencing the genome. In F. Levy & D. Hay (Eds.), Van den Bergh, B. R. H. (1992). Maternal emotions during Attention, genes and ADHD (pp. 225 – 248). Hove, Eng- pregnancy and fetal and neonatal behavior. In J. G. land: Brunner-Routledge. Nijhuis (Ed.), Fetal behavior: Developmental and perinatal
You can also read
