Content uploaded by David P Laplante
Author content
All content in this area was uploaded by David P Laplante
Content may be subject to copyright.
The effects of prenatal maternal stress on children’s cognitive
development: Project Ice Storm
SUZANNE KING
1,2
& DAVID P. LAPLANTE
2
1
Department of Psychiatry, McGill University, Que
´
bec, Canada, and
2
Douglas Hospital Research Centre, Verdun, Que
´
bec,
Canada
Abstract
There exists considerable research on the effects of prenatal maternal stress on offspring. Animal studies, using random
assignment to experimental and control groups, demonstrate the noxious effects of prenatal maternal stress on physical,
behavioural and cognitive development. The generalizability of these results to humans is problematic given that cognitive
attributions moderate reactions to stressors. In humans, researchers have relied upon maternal anxiety or exposure to life
events as proxies for the stressors used with animals. Yet, the associations between maternal anxiety or potentially non-
independent life events and problems in infants are confounded by genetic transmission of temperament from mother to child.
We summarize the literature on prenatal maternal stress and infant cognitive development, leading to the conclusion that the
human literature lacks the ability to separate the effects of the objective exposure to a stressor and the mother’s subjective
reaction. We then describe our prospective Project Ice Storm in which we are following 150 children who were exposed in utero
to a natural disaster. We demonstrate significant effects of the objective severity of exposure on cognitive and language
development at age two years with important moderating effects of the timing during pregnancy. The implications of our
findings are discussed.
Keywords: Cortisol, foetal programming hypothesis, human, hypothalamo–pituitary– adrenal axis, language development,
longitudinal study
Introduction
Retrospective studies on humans, and experimental
research with animals, suggest that psychosocial
stressors during pregnancy can influence the physical,
behavioural and cognitive outcomes of the offspring.
There are important gaps in the existing literature,
however.
Research on prenatal maternal stress is hampered
by methodological constraints. Although animal
studies are ideal in their control of prenatal and
postnatal environments, the application of findings to
humans is not always clear. Human studies, on the
other hand, do not allow for stressful events to be
randomly assigned during pregnancy by an experi-
menter and may be compromised by a host of
potential confounds.
Natural and man-made disasters act as “natural
experiments”, randomising the distribution of stress
exposure. The Que
´
bec Ice Storm of January 1998
qualifies as such an event. The storm resulted in
electrical power failures for more than three million
individuals for anywhere from 6 h to more than
5 weeks. As such, large numbers of pregnant women
in various stages of pregnancy were randomly
exposed to varying degrees of storm-related
hardship.
The purpose of this paper is to review the existing
literature on the effects of prenatal maternal stress,
particularly with respect to infant cognitive develop-
ment, and to present an introduction to our on-going
longitudinal study of children exposed to the January
1998 Quebec Ice Storm while in utero. In addition, we
will present early findings with respect to effects of
ISSN 1025-3890 print/ISSN 1607-8888 online q 2005 Taylor & Francis Group Ltd
DOI: 10.1080/10253890500108391
Correspondence: S. King, Douglas Hospital Research Centre, 6875 LaSalle Blvd., Verdun, Quebec H4H 1R3, Canada. Tel: 1 514 761 6131.
Ext. 2353. Fax: 1 514 762 3049. E-mail: suzanne.king@douglas.mcgill.ca
Stress, March 2005; 8(1): 35–45
the ice storm stress on children’s cognitive and
language development at age 2 years.
Effects of prenatal maternal stress (PNMS)
on foetal and child development
There is a sizeable literature on the effects of prenatal
maternal stress (PNMS) in rodents, non-human
primates, and in humans. In humans, studies of
prenatal maternal anxiety (Crandon 1979a, 1979b,
Field et al. 1985, Grimm 1961) and of prenatal
exposure to severe life events (Lou 1993) suggest that
one of the first effects is on perinatal outcomes, with
PNMS associated with a variety of severe obstetric
complications, but especially with preterm birth
(Paarlberg et al. 1999, 1995).
Weinstock reviewed the literature on prenatal stress,
particularly rodent studies, and reports wide-ranging
effects (Weinstock 2001). Rodent studies typically use
various forms and intensities of restraint or noise to
stress pregnant dams one to three times per day,
usually for several days at some point between days 10
and 22 of gestation. Behavioural effects of prenatal
exposure to various forms of stress are seen in social
deficits and behaviours in rodent offspring suggestive
of internalizing problems such as anxiety and
depression. Controlled studies with non-human
primates also indicate that PNMS results in beha-
vioural alterations suggestive of internalizing problems
(Clarke and Schneider 1993, Schneider 1992b).
Naturalistic studies with humans find similar effects
on behaviour (Huizink et al. 2004). Studies of prenatal
maternal anxiety (O’Connor et al. 2002, Van den
Bergh and Marcoen, 2004), prenatal exposure to
stressful life events (Stott 1973), and even prenatal
exposure to dexamethasone (a synthetic glucocorti-
coid) (Trautman et al. 1995) are associated with
children who are more withdrawn, anxious and
depressed. A handful of retrospective studies have
examined associations between independent stressors
and mental health outcomes. A classic study by
Huttunen and Niskanen (1978) examined rates of
mental illness in samples from Finland in which the
father had died either while the person was in utero or
during the person’s first year of life. Significantly
greater rates of schizophrenia and other mental illness
were found in the prenatal stress exposure group. Van
Os and Selton (1998) also found a significant increase
in rates of schizophrenia in Holland as a function of
the German invasion during World War II. These
studies show a specific effect of the timing of the
stressor during the pregnancy, with the most noxious
effects associated with exposure during the second
trimester.
A number of other studies in non-human primates
suggest that social stressors during gestation have
negative effects on neurological functioning in the
offspring, including reduced motor activity and
maturity, less muscle tone and co-ordination, slower
reaction time, and poorer balance (Schneider and Coe
1993, Schneider et al. 1999). In humans, significantly
lower Prechtl neurological inventory scores 4–14 days
after birth were found when mothers had experienced
severe life events by mid-pregnancy (Lou et al. 1994).
PNMS studies have also discovered cognitive
sequelae. In rodents, PNMS has resulted in impair-
ment in maze learning (Nishio et al. 2001), reversal of
learning set (Weller et al. 1988), and short and long-
term memory (Gue 2004), while some studies show
no effect on, for example, acquisition, discrimination
or extinction in an operant conditioning task (Weller
et al. 1988) or on object memory (Bowman 2004).
Studies of non-human primates have found that both
chronic stress during pregnancy (Schneider and Coe
1993, Schneider et al. 1999) and two-weeks of
adrenocorticotrophic hormone administration at
mid-gestation (Schneider 1992a) predict poorer
attention, greater distractibility, and delayed object
permanence (Schneider 1992a), an early indicator of
cognitive development. Similar findings have been
made in humans when linking mothers’ self-reported
stress levels (McIntosh et al. 1995) or anxiety
(Van den Bergh and Marcoen 2004) during pregnancy
to risk of attention-deficit and hyperactivity disorder
(ADHD) in children. In a longitudinal study of
anxiety in pregnant mothers and outcomes in their
babies in England, O’Connor et al. showed that, even
after controlling for a host of potential confounding
factors, higher levels of anxiety experienced by the
mother at weeks 12–22 of pregnancy significantly
predicted more severe attention problems in the
children at 18 and 32 months of age (O’Connor et al.
2002). Another study found that at 9 years of age,
children of women who were highly anxious during
pregnancy showed more attention deficits than
children of low anxiety women (Mulder et al. 2002).
Chronic stress resulting from moderate daily hassles
(Stott 1973) and state and/or trait anxiety (Brouwers
et al. 2001) are also associated with delayed language
development and lowered intellectual functioning,
respectively.
The effects of PNMS on cognitive development
may be due to an effect on foetal brain development.
PNMS in rodents predicts the degeneration of
hippocampal neurons (Uno et al. 1989, 1990) and
reductions in the size of the hippocampus by as much
as 30% (Uno et al. 1994), which might be manifest in
reduced memory functioning. Lower scores on
cognitive tests may be due, in part, to the indirect
effects of prenatal stress on attentional functioning.
Timing of teratogens
It has been hypothesised that it may be less the type of
disruption of foetal neural development than the
timing that is critical in determining risk for negative
S. King & D. P. Laplante36
outcomes since the timing of an insult may indicate
which developmental processes are likely to be
affected (Mednick et al. 1988). A common theme
among many studies of PNMS is that the worst
outcomes are associated with stressors timed at mid-
gestation (Glynn et al. 2001, Huizink et al. 2004,
Huttenen and Niskanen 1978, O’Connor et al. 2002,
2003, Schneider et al. 1992, Watson et al. 1999),
a critical period for brain development (Andreason
1999, Weinberger 1995). Moreover, if outcomes are
mediated through foetal hypothalamus– pituitary–
adrenal (HPA) axis activity, its influence should only
be seen in mid-gestation after the foetal HPA axis
becomes functional (Gitau et al. 2001). Furthermore,
reports suggest that mothers may become more
immune to the potential negative effects of stress as
their pregnancy advances into the 3rd trimester of
pregnancy (Glynn et al. 2001, Kammerer et al. 2002).
However, some researchers have found effects of 3rd
trimester stress on obstetric complications (Crandon
1979b). Also, it has been reported that early gestation
stress in non-human primates has a more detrimental
effect on birth weight and neuromotor functioning
than mid-late gestational stress (Schneider et al.
1999). Thus, the question of timing of prenatal stress
is intimately associated with the outcome factor of
interest.
Critique of the literature
There is no shortage of studies of prenatal maternal
stress. Yet, to date, there have been no ideal studies of
PNMS that can determine the extent, longevity and
mechanisms of the effects in humans. In animal
research, “subjects” can be randomly assigned to
various stress conditions. While these studies can
provide working hypotheses on the mechanisms by
which PNMS influences outcomes, there are limi-
tations to animal studies. First, rats are not born at the
same stage of development as humans are. Secondly,
rat and primate studies are unable to take into account
the kinds of cognitive appraisals that occur in humans
between an external stressor and the response to the
stressor which may well be the critical variable in
determining a pregnant woman’s physiological
response to a stressor (Lazarus 1991) and subsequent
damage to the foetus.
The human studies that do assess “stress” during
the pregnancy are also unable to tease apart the
various aspects of the process. Studies of anxiety
during pregnancy cannot separate state versus trait
anxiety, the latter of which may be passed on to babies,
in part, genetically (Rowe 1994) and through
parenting style and modelling, as well as through the
complicated mechanisms of gene-by-environment
correlations (Rutter and Silberg 2002). Studies of
life events during pregnancy are equally hampered in
that such events are far from being randomly assigned
to women; divorce and job loss are not always
“independent” life events but may, instead, be
influenced by heritable personality traits. Most
human studies of stress also have limited statistical
power: An extremely large sample of pregnant women
is required to include a sufficient number who had
experienced an independent life event. And the large-
scale retrospective studies of independent life events
(death of father, hurricane, tornado, foreign invasion)
are too far removed from the incident to assess the
pregnant woman’s appraisal of the event. What is
needed to permit us to tease apart the objective and
the subjective components of the stress process is a
man-made or natural disaster that affects large
numbers of pregnant women in a quasi-random
fashion.
Disaster research
Disasters provide unique opportunities to study the
effects of stress in people. Disasters are characterised
by “disruption exceeding the adjustment capacity of
the affected community” (Lechat 1979, p. 11) and
range from oil spills to hurricanes and earthquakes.
Important dimensions of disasters include loss
(of persons or property), threat to life or physical
integrity, scope and duration, blame for the event
(man versus nature), familiarity with similar events,
speed of onset and amount of displacement or change.
Since events differ along many dimensions, one must
tailor-make disaster questionnaires in order to assess
degree of exposure.
Standardised instruments, however, have proven
useful in estimating the psychological impact of
disasters. One review of 52 studies concluded that, on
average, sudden onset disasters are associated with a
17% increase in the prevalence of psychological
disorders (Rubonis and Bickman 1991) including
post-traumatic stress disorder (Palinkas et al. 1993).
When response is measured along a psychopathology
continuum, disasters have been shown to be associated
with significant increases in psychological symptoms.
One review uncovered 7 so-called “natural exper-
iments” in which a disaster struck a community that
had recently been surveyed in an epidemiological
study, providing estimates of pre- and post-disaster
symptoms on the same subjects. These studies confirm
the disaster effect (Bromet and Dew 1995). In some
instances, the emotional effects of a disaster may last
for many years as shown in a recent 11-year follow-up
of women evacuated to Kyev following the Chernobyl
nuclear power plant explosion (Adams et al. 2002).
Questions remaining about prenatal maternal stress
The available literature on prenatal maternal stress
suggests that a large gap exists: Understanding how
the various elements of the stress process
Effects of prenatal maternal stress 37
(in particular, the objective exposure versus the
subjective reaction) operate, individually and in
combination, to influence cognitive and other out-
comes in the unborn human child. In order to
successfully disentangle the objective elements of the
stress process from the mother’s own temperament
(which can be passed on to her children through a
combination of genetics, the prenatal environment,
and parenting style, not to mention the influence of
the mother’s own temperament on her response style
as she completes questionnaires rating her child) one
requires an independent stressor that affects large
numbers of pregnant women and which is distributed
in a relatively random fashion.
Project Ice Storm
Project Ice Storm was initiated soon after a series of
freezing rain storms hit Southern Quebec in Canada
between January 5 and 9, 1998. The build-up of ice on
utility lines toppled more than 1000 electrical pylons
and more than 25,000 wooden transmission poles
resulting in electrical power failures during the dead of
winter for more than 1.4 million households in more
than 700 municipalities. More than three million
individuals experienced power outages ranging from a
few hours to more than 6 weeks. Four hundred and fifty
emergency shelters were established which slept as
many as 17,000 people in a single night. In Quebec, at
least 27 people died as a result of hypothermia,
accidents, or carbon monoxide poisoning or fire
associated with unconventional heating methods.
Thousands of people were injured or hospitalized as a
result of the storm. The agricultural sector suffered
major financial losses associated with the power outage:
Thousands of animals were lost when ventilation and
heating systems or milking machinery could not be
kept going. The maple syrup industry was nearly
obliterated due to the near total loss of branches on
maple trees. There were $1 Billion worth of insurance
claims, $3 Billion of lost income to businesses and $1
Billion dollars required to repair hydroelectric infra-
structure. More than 46,000 people were laid off work
as a direct or indirect consequence of the crisis.
Everyone had to deal with 5-inch ice build-up on cars,
streets, trees and roofs. Several buildings collapsed
under the weight of ice and snow and more than 200
people were hospitalized with severe injuries as a result
of falling off their roofs while removing snow and ice; 3
people died in this way. (Statistics provided by Ministe
`
re
de la Se
´
curite
´
Publique.)
Thus, the 1998 Quebec ice storm provided a unique
opportunity. The objective of Project Ice Storm is to
determine the nature and duration of the effects of an
independent stressor during pregnancy on the unborn
child in a prospective design with a relatively
large sample of families. By conducting repeated
assessments of women affected by the ice storm, as
well of their children, over several years, we are
determining the effects of objective stress exposure
and subjective stress reaction on perinatal outcomes,
maternal postpartum depression, and the behavioural,
physical and cognitive development of the children.
Method
Sample development
The area of Que
´
bec that was worst hit by the storm
was the region called the Monte
´
re
´
gie, an area
southeast of Montreal that is mainly suburban and
rural and which came to be known as “the Black
Triangle”. To identify potential subjects, we contacted
the four hospitals in the region and obtained the
names of physicians authorized to deliver babies there.
Approximately 20 doctors identified more than 1400
women who met our inclusion criteria: They were
pregnant on January 9, 1998 or became pregnant
during the 3 months following the storm, were 18
years of age or older, and spoke fluent French.
On June 1, 1998, we delivered the appropriate
number of pre-stamped questionnaires to each
participating clinic where clinic staff addressed the
envelopes and mailed them. Of the 1440 question-
naires sent out, 224 were returned for a response rate
of 15.5% which is considered to be the norm for
unsolicited postal questionnaires. This response rate
may be explained also by the fact that during the
summer of 1998 the potential subjects were either
pregnant or caring for a newborn, as well as for any
older children, and may also have been dealing with
ice storm-related damage to the home and financial
sequelae. Of the 224 women who responded to the
first questionnaire, 178 provided their name and
address and agreed to further contact. We have
continued to follow as many of these families as
possible; many have not been followed since because
their ice storm pregnancy ended in miscarriage or
stillbirth, while others have been lost to follow-up, and
a small number have refused continued involvement,
leaving 141 families who are continuing their
participation in the study.
Instruments
In total, we have made contact with the families 7
times between June 1998 and the children’s 6th
birthday with an eighth contact on-going and a ninth
in the planning stage for an evaluation at age 7
1
2
years.
Questionnaire #1—Reactions to the Storm, was sent
on June 1, 1998. The primary goal of the first
questionnaire was to evaluate the extent of the
women’s objective exposure to the ice storm, as well
as their subjective reaction to the storm. In order to
assess objective exposure, we examined the literature
on natural disasters. We formulated questions (Table I)
S. King & D. P. Laplante38
that would reflect the participant’s experiences related
to 4 categories of exposure used in other disaster
studies: Threat, Loss, Scope and Change (Bromet and
Dew 1995). Each dimension was scored on a scale of
0–8, ranging from no exposure to high exposure.
A total objective stress score was calculated by
summing scores from all four dimensions using
McFarlane’s approach (McFarlane 1988). Because
there was no theoretical basis to believe that any one of
the four dimensions of our scale was more distressing
than the other dimensions, and based on McFarlane’s
study of Australian fire fighters (McFarlane 1988),
each dimension was weighted equally to obtain the
total score of our scale which we dubbed STORM32.
Questionnaire #1 also included an assessment of the
women’s subjective stress reaction to the ice storm. In
the literature on natural disasters and trauma, one
instrument stands out as being the gold-standard
measurement of subjective reactions: The Impact of
Event Scale– Revised (Weiss and Marmar 1997). The
22-item scale describes symptoms from three cat-
egories relevant to post-traumatic stress disorder:
Intrusive thoughts, hyperarousal and avoidance.
A French-version of the scale was developed and
validated by our team (Brunet et al. 2003) to reflect the
mothers’ symptoms relative to the ice storm crisis.
Participants respond on a 5-point Likert scale, from
Not at all to Extremely, the extent to which the behaviour
describes how they felt over the preceding seven days.
To assess the severity of maternal psychiatric and
stress-related symptoms in general, we have used, at
various times, either the General Health Question-
naire (Goldberg 1972) which includes scales reflec-
ting anxiety, depression, somatic symptoms and
dysfunction, or the Edinburgh Postnatal Depression
Scale (Cox et al. 1987). Mothers have also indicated,
since the birth of their baby, their experiences with
any other life events in the preceding 12 months using
a modified version of the Life Experiences Survey
(Sarason et al. 1978). We have also assessed the
mothers’ personality using the short form (60-item)
NEO Five Factor Inventory (Costa and McCrae
1992) which includes scales for neuroticism, extro-
version, openness to experience, agreeableness and
conscientiousness. Using data on maternal and
paternal education and occupation, we estimated
socioeconomic status using the Hollingshead scale
(Hollingshead 1973).
Laboratory assessment. When the children were 2 years
old, a subset of 61 families participated in a laboratory
assessment of cognitive (Bayley Scales of Infant
Development, 2nd Edition Bayley 1993), language
(MacArthur Communicative Development Inventory
(Fenson et al. 1993)) and functional play (Zelazo and
Kearsley 1980) development. These families were
selected to represent the lowest and highest thirds in
severity of objective ice storm stress, and the 3
trimesters of exposure. Children with severe low birth
weight, Caesarian-section delivery, or prenatal
smoking or other severe prenatal stressor (e.g. death
of close relative) were excluded.
Results: The ice storm and cognitive
development
Here, we provide an overview of findings on the effects
of prenatal exposure to the ice storm on cognitive
development at age two years. These results have been
Table I. Questions used to assess the four dimensions (Threat, Loss, Scope, and Change) of our objective stress questionnaire that the
mothers completed shortly after the ice storm.
Threat Loss Scope Change
1. Were you injured? 1. Did your residence suffer
damage as a result
of the ice storm?
1. How many days were
you without electricity?
1. Did your family stay
together for the duration
of the ice storm?
2. Was anyone close to
you injured?
2. Did you experience a
loss of personal income?
2. How many days were
you without the use
of your telephone?
2. Did you spend any
time in a temporary
shelter?
3. Were you ever in
danger due to:
3. Did you suffer a
loss of business income?
3. How often were you
required to change residence
during the ice storm?
3.1. ...the cold 4. Did you take in
guests during the ice
storm?
3.2. ...exposure to downed electrical
power lines
5. Did you experience an
increase in physical work
during the ice storm?
3.3. ...exposure to carbon monoxide
3.4. ...lack of potable water
3.5. ...lack of food
3.6. ...falling branches & ice
Effects of prenatal maternal stress 39
published (Laplante et al. 2004), or are under review
(Laplante et al. submitted), elsewhere.
Reaction to the storm
The mothers in our sample spent, on average, 14.9
ðSD ¼ 8:9Þ days without electricity and 4.4 ðSD ¼
8:4Þ days without the use of their telephone.
Approximately 65% of the mothers spent at least
one night away from their home for a period, on
average, of 9.2 ðSD ¼ 11:2Þ nights away from their
home, changing locations, on average, 1.3 ðSD ¼ 1:2Þ
times. Of the mothers who remained in their homes
during the ice storm, 37.7% had guests staying with
them: On average, 1.4 ðSD ¼ 2:4Þ guests stayed for an
average of 30 days ðSD ¼ 5:9Þ: Approximately half of
the mothers reported damage to their homes (46.6%)
and loss of personal income (44.7%) directly related
to the ice storm. Six percent of the mothers were
physically injured as a result of the ice storm and
37.7% reported being worried about the personal
safety of loved ones.
Based on the items included in the four 8-point
scales of our STORM32 objective stress measure, the
mothers in our sample scored, on average, 3.0 points
ðSD ¼ 2:6Þ on our Scope dimension, 3.0 points ðSD ¼
1:9Þ on our Loss dimension, 3.1 points ðSD ¼ 1:7Þ on
our Change dimension, and 1.5 points ðSD ¼ 1:4Þ on
our Threat dimension. Overall, the mothers in our
sample scored, on average, 10.4 points ðSD ¼ 4:9Þ on
our total STORM32 scale.
In terms of subjective stress reaction, the mothers in
our sample scored, on average, 11.9 points ðSD ¼
12:5Þ on our French-version of the IES-R scale. Using
the cut-off of 22 points on this scale, 16.6% of the
mothers exhibited levels of subjective stress reaction
that placed them within the clinical range for potential
post-traumatic stress disorder.
Intellectual, language and play abilities at 2 years
To examine interactions between trimester of
exposure and severity of ice storm stress, we split
STORM32 into three groups (low, moderate and high
stress), and contrasted the low stress group with the
combined moderate-high stress group. We found that
moderate-high objective prenatal maternal stress was
associated with poorer intellectual and language
functioning at the age of two years (Laplante et al.
2004). These findings were reinforced by the results of
our analyses of the children’s play behaviours
(Laplante et al. submitted). For intellectual abilities,
children whose mothers were exposed to the ice storm
during their 1st or 2nd trimester of pregnancy and
who experienced moderate or high objective stress had
significantly lower Bayley MDI scores. For the 1st
trimester exposed children, the moderate-high stress
group had an average score 14 points lower (i.e. 0.95
of 1 SD) than their counterparts in the low stress
group (see Figure 1). For the 2nd trimester exposed
group, children from the moderate-high stress group
scored an average of 19.5 points lower than their
counterparts in the low stress group, a 1.3 SD
difference. These effects remained even after control-
ling for age at testing, birth weight, subjective stress
scores (IES-R), and socioeconomic status, with an
adjusted overall difference of 9.5 points between stress
groups across all three trimesters. There were no
effects of stress severity on the MDI scores for children
exposed in the third trimester of pregnancy.
Using maternal reports of their children’s word
comprehension and word use at age 2 years, we found
Figure 1. Toddlers’ mean (^ standard error) Bayley mental development idex (MDI) scores at 2 years of age, as a function of objective
prenatal maternal stress levels (low, high) and trimester of exposure (1st, 2nd, or 3rd).
S. King & D. P. Laplante40
similar effects of objective stress severity, but without
the differential effect of trimester of exposure
(Figure 2). Children whose mothers were exposed to
moderate-high objective stress spoke, on average, 20.2
fewer words ðp , 0:001Þ which represents a 30%
reduction in productive vocabulary. There was also a
significant effect of objective ice storm exposure on
receptive vocabulary, again with no differential effect
by timing of exposure. At age 2 years, children of
mothers from the moderate-high objective stress
group understood, but did not yet speak, an average
of 10.5 fewer words ðp , 0:001Þ than their low stress
counterparts, representing 11% fewer words.
Because a two year-old’s performance on structured
tasks may reflect temperamental or behavioural
characteristics rather than purely cognitive skills, we
included a non-structured free-play task that was
videotaped and rated blind to prenatal stress level. Toy
play during early childhood evolves in a predictable
manner and provides excellent discriminative and
predictive validity as a measure of cognitive function.
The youngest toddlers will engage in “stereotypical”
play which involves banging, waving or mouthing toys.
Older toddlers will engage in “relational” play by
touching two toys together in a non-functional
manner. A more mature level of toy play involves
using objects according to their intended function.
Thus, in “functional” toy play, a child may roll a car on
the floor while making car noises, or pour imaginary
tea from a pot into a teacup. The results of this study
support the findings from the Bayley Scales with
moderate-high stress children engaging in significantly
more stereotypical play and less functional play than
their low stress counterparts, and with the effects
limited to children exposed during the 1st and 2nd
trimesters (p , 0:01; Figure 3). For 2nd trimester
children, the level of objective maternal stress
(STORM32 score) accounted for 53.8% of the
variance in the level of the children’s functional toy
play, while the mother’s subjective stress (IES-R)
explained an additional 13.6% of the variance, even
after controlling for the children’s birth weights, the
number of obstetric complications experienced by the
mothers and the mothers’ level of anxiety during
pregnancy ðp , 0:005Þ: Maternal objective and sub-
jective PNMS was not related to the functional play
levels of children exposed during the 1st or 3rd
trimester of pregnancy.
We have conducted additional cognitive assess-
ments of the children at age 5
1
2
years. Preliminary
examination of the data suggests that the effects of
objective prenatal maternal stress on cognitive and
language development persist into middle childhood
(unpublished data). A repeat assessment is planned
for age 7
1
2
during 2005 – 2006.
Discussion
The findings from Project Ice Storm strongly suggest that
a major stressful event, independent of maternal
personality factors, can have a negative impact on
cognitive and language development of the unborn
child. Other results not reported here also demonstrate
significant noxious effects of the severity and/or timing
of prenatal maternal stress on perinatal outcomes,
infant temperament, behavioural and emotional
functioning, and even physical development of
children exposed in utero. We have uncovered several
effects that are related to the timing of the stressor,
with a particular emphasis on the 2nd trimester.
Figure 2. Toddlers’ mean (^ standard error) MacArthur communicative development index (MCDI) scores for productive and receptive
language abilities at 2 years of age, as a function objective prenatal maternal stress levels (low, high) and trimester of exposure (1st, 2nd, or 3rd).
Effects of prenatal maternal stress 41
Furthermore, while there have been a handful of results
implicating the mother’s subjective reaction to the
stressor, the majority of findings implicate the degree of
her objective exposure to the event.
The mechanism by which prenatal maternal stress
affects the pregnancy and the subsequent develop-
ment of the children, however, remains unknown for
human populations, although maternal glucocorti-
coids are often cited as a likely teratogen. There is
much support for the role of maternal glucocorticoids
in the PNMS effect in animals (Weinstock 2001,
Schneider et al. 1992). In humans, indirect evidence
for effects of maternal hormones is available.
Trautman et al. (1995) conducted a study of children
whose mothers had received dexamethasone (DEX)
(a synthetic glucocorticoid) during pregnancy for
congenital adrenal hyperplasia (CAH). They found
that these children were significantly more withdrawn
and emotional and significantly less social at one year
of age, and had significantly higher scores on a scale of
internalising problems at ages 2–3 years than a group
of comparison children whose CAH mothers went
untreated with DEX. These data suggest that it is
maternal stress hormones that are responsible for the
effects seen in children.
Although we have shown, in Project Ice Storm, that
objective PNMS in early gestation predicts poorer
cognitive functioning, we have not so far been able to
show that these findings are the direct effects of the
stress exposure either on maternal cortisol levels or on
disruptions in foetal brain development. Although we
obtained salivary cortisol samples from the mothers in
our study at the time of our first contact, those samples
were taken five-to-six months after the storm, when
half of the mothers had given birth, rendering our
maternal cortisol data difficult to interpret.
There are also animal studies showing that PNMS is
associated with alterations in stress reactivity in the
offspring which may be involved in a variety of
cognitive and behavioural outcomes. In rhesus
monkeys, for example, prenatal exposure to either
glucocorticoids or to stress is associated with greater
increases in cortisol following a stressor (Uno et al.
1994). These offspring also show a slower recovery
following stress (Uno et al. 1994). Greater stress
reactivity is also found in prenatally stressed rats
(Koehl et al. 1999, Takahashi and Kalin 1991,
Weinstock et al. 1992), whose stress-induced cortico-
sterone secretion is also prolonged, that is, their
recovery from the stressor is slower (Fride et al. 1986,
Maccari et al. 1995). We are currently analyzing data
related to the children’s hormonal stress responsivity
which may support an effect of PNMS on the
development of the foetal HPA axis. A subset of
children provided saliva samples before and after their
pre-kindergarten inoculation session. Their cortisol
data will be analyzed to determine effects of PNMS,
and to determine moderating effects of the trimester of
exposure and the child’s sex.
There are other potential mechanisms for the effect
of PNMS that Project Ice Storm is unable to investigate.
Although a stress- or disaster-related reduction in
supply of nutrients to the mother and foetus may be
involved in the process, we failed to include questions
about eating patterns in our initial ice storm
questionnaire. Maternal thyroid function may have
compromised as a result of the cold, but this, too, went
unmeasured. There was a great deal of water damage
to homes as a result of ice build-up on roofs which
fostered the production of moulds in many homes and
could, potentially, have had an effect on the pregnant
women. Yet this type of damage, as well, was not
Figure 3. Percentage of functional and stereotypical play exhibited by the toddlers as a function of level of maternal objective stress
(high or low) and trimester of exposure.
S. King & D. P. Laplante42
addressed in our survey. Finally, a compromised gas
exchange across the placenta may have been involved
in the stress effect seen in our sample, but these data
are not available in the majority of hospital records.
For a random subsample of cases, however, placental
weight at birth is available from hospital charts, and
we are currently investigating associations between
stress levels and placental weight on one hand, and
between placental weight and developmental out-
comes on the other.
Although we are lacking data on the immediate
biological processes that may have mediated between
stress exposure or stress reaction in the mother and
developmental processes in the foetus, we are in a
position to determine whether the effects of PNMS
exposure on cognitive outcomes are due to perma-
nent insults to brain development by obtaining
structural magnetic-resonance images (MRI) of the
children, a project currently in development.
Research suggests that either stress or administration
of stress hormones to pregnant females results in
severe degenerative changes in neurons within the
hippocampus (Uno et al. 1990), resulting in an
overall reduction in hippocampal volume, even
though the total brain volume remains unchanged
(Uno et al. 1994) in these animals. In addition,
magnetic resonance imaging studies indicate that this
hippocampal volume loss is maintained 2-years
postnatally (Uno et al. 1994). Using an acoustic
startle protocol with rhesus monkeys, Coe and
colleagues (Coe et al. 2003, Coe et al. 2002) report
similar findings to those obtained by Uno. Both early
(days 50 – 92) and late (days 105 –147) stress to
pregnant non-human primate females result in a
reduction in the size (Coe et al. 2003) and alterations
to the shape (Coe et al. 2002) of the hippocampus.
To date, no study has used a prospective design to
examine the potential impact of prenatal maternal
stress on the structural development of the hippo-
campus in humans. Thus, our MRI scans of ice storm
children will target hippocampal volumes as well as
other regions.
Weather researchers agree that the visible effects of
global warming will be more frequent and more severe
extreme weather events. Moreover, man-made
threats, such as terrorist attacks on civilian popu-
lations, are becoming more prevalent. Such events will
lead to greater numbers of pregnant women facing
stressful events that are outside of their control and
that will result in foetuses experiencing increasing
levels of prenatal maternal stress. Therefore, research
is required to better understand the process by which
prenatal maternal stress influences prenatal develop-
ment and postnatal outcomes.
We propose that a first step in understanding these
mechanisms is to separate the stressful event into its
components, such as objective levels of exposure to
hardship, and subjective emotional reactions to that
hardship. Evaluation of objective stress involves
examining the specific events faced by the pregnant
women during the crisis, such as duration of power
loss, loss of income and amount of change to the
physical environment. The second component, sub-
jective reaction to the objective stress, involves
assessing the pregnant women’s psychological
response to these specific events. Borrowing from the
literature on post-traumatic stress disorder, these
subjective responses will include varying degrees of
hyperarousal, avoidance of the traumatic event, and
intrusive thoughts about the event. Separation of the
objective and subjective aspects of a person’s stressful
experienceisimportantinthatitwillenable
researchers to make recommendations on how to
prevent, or limit, the potential negative effects of
prenatal stress on the developing foetus. For example,
if it is found that particular aspects of the objective
exposure (such as amount of change, or physical
exertion) carry the greatest weight in the teratogenic
effects, then public health and safety personnel may
need to ensure that pregnant women exposed to a
natural disaster are sent immediately to very safe areas
rather than being moved several times. If, however, it is
found that the greatest harm to the foetus results from
the mother’s physiological hyperarousal following a
disaster, then psychological (e.g. biofeedback) or
pharmacological interventions may be warranted to
limit damage to the foetus. Thus, projects such as
Project Ice Storm will assist in, first, identifying the most
vulnerable subgroups of pregnant women and,
secondly, lead to recommendations targeted at limit-
ing damage.
Acknowledgements
This study was funded by grants from the Stairs
Memorial Fund, the Quebec Mental Health Research
Network, the Canadian Psychiatric Research Foun-
dation, and the Canadian Institutes of Health
Research, to the first author, and by a grant from the
Schizophrenia Axis of the Fonds de la recherche
en sante
´
du Que
´
bec’s Network on Mental Health
Research to the second author. The first author is also
a recipient of a salary award from the Quebec Health
Research Fund (FRSQ).
References
Adams RE, Bromet EJ, Panina N, Golovakha E, Goldgaber D,
Gluzman S. 2002. Stress and well-being in mothers of young
children 11 years after the Chernobyl nucelar power plant
accident. Psychol Med 32:143–156.
Andreason NC. 1999. A unitary model of Schizophrenia: Bleuler’s
“fragmented phrene” as schizencephaly. Arch Gen Psychiatry
56:781–787.
Bayley N. 1993. Bayley scales of infant development. 2nd ed.,
San Antonio: Psychological Corporation.
Effects of prenatal maternal stress 43
Bowman. 2004. Sexually dimorphic effects of prenatal stress on
cognition, hormonal responses, and central neurotransmitters.
Endocrinology 145:3778– 3787.
Bromet E, Dew MA. 1995. Review of psychiatric epidemiologic
research on disasters. Epidemiol Rev 17(1):113–119.
Brouwers EPM, van Baar EL, Pop VJM. 2001. Maternal anxiety
during pregnancy and subsequent infant development. Infant
Behav Dev 24:95– 106.
Brunet A, St-Hilaire A, King S. 2003. Validation of a French version
of the impact of event scale—revised. Can J Psychiatry 48:
55–60.
Clarke AS, Schneider ML. 1993. Prenatal stress has long-term
effects on behavioral responses to stress in juvenile rhesus
monkeys. Dev Psychobiol 26:293–304.
Coe CL, Lulbach GR, Schneider ML. 2002. Prenatal stress alters
the size and shape of the corpus callosum in the infant monkey.
Dev Psychobiol 41:178 –185.
Coe CL, Kramer M, Cze
´
h B, Gould E, Reeves AJ, Kirschbaum C,
et al. 2003. Prenatal stress diminishes neurogenesis in the
dentate gyrus of juvenile rhesus monkeys. Biol Psychiatry
54:1025–1034.
Costa JF, McCrae RR. 1992. Revised NEO personality inventory
(NEO-PI) and NEO five-factor inventory (NEO-FFI) pro-
fessional manual. Odessa, FL: Psychological Assessment
Resources.
Cox JL, Holden JM, Sagovsky R. 1987. Detection of postnatal
depression: Development of the 10-item Edinburgh postnatal
depression scale. Br J Psychiatr 150:782–786.
Crandon AJ. 1979a. Maternal anxiety and neonatal wellbeing.
J Psychosom Res 23:113–115.
Crandon AJ. 1979b. Maternal anxiety and obstetric complications.
J Psychosom Res 23:109–111.
Fenson L, Dale P, Reznick J. 1993. MacArthur communicative
development inventories: User’s guide and technical manual.
San Diego: Singular Publishing Group.
Field T, Sandberg D, Quetel TA, Garcia R, Rosario M. 1985.
Effects of ultrasound feedback on pregnancy anxiety, fetal
activity, and neonatal outcome. Obstet Gynecol 66:525– 528.
Fride E, Dan Y, Feldon J, Halevy G, Weinstock M. 1986. Efects of
prenatal stress on vulnerability to stress in prepubertal and adult
rats. Physiol Behav 37:681–687.
Gitau R, Fisk NM, Teixeira JMA, Cameron A, Glover V. 2001. Fetal
hypothalamic– pituitary – adrenal stress responses to invasive
procedures are independent of maternal responses. J Clin
Endocrinol Metab 86:104–109.
Glynn LM, Wadhwa PD, Dunkel Schetter C, Chicz-Demet A,
Sandman CA. 2001. When stress happens matters: Effects of
earthquake timing on stress responsivity in pregnancy. Am J
Obstet Gynecol 184:637 – 642.
Goldberg DP. 1972. The detection of psychiatric illness by
questionnaire: A technique for the identification and assessment
of non-psychiatric illness. London: Oxford University Press.
Grimm EG. 1961. Psychological tension in pregnancy. Psychosom
Med 23:520– 527.
Gue. 2004. Sex differences in learning deficits induced by prenatal
stress in juvenile rats. Behav Brain Res 150:149–157.
Hollingshead AB. 1973. Four-factor index of social status.
New Haven: Yale University Press.
Huizink AC, Mulder EJH, Buitelaar JK. 2004. Prenatal stress and
risk for psychopathology: Specific effects or induction of general
susceptibility? Psychol Bull 130:115–142.
Huttenen MO, Niskanen P. 1978. Prenatal loss of father and
psychiatric disorders. Arch Gen Psychiatry 35:429–431.
Kammerer M, Adams D, von Castelberg B, Glovert V. 2002.
Pregnant women become insensitive to cold stress. BMC
Pregnancy and Childbirth 2:8.
Koehl M, Darnaudery M, Dulluc J, Van Reeth O, Le Mooal M,
Maccari S. 1999. Prenatal stress alters circadian activity of
hypothalamic– pituitary – adrenal axis and hippocampal
corticosteroid receptors in adult rats of both genders. J
Neurobiol 40:302– 315.
Laplante DP, Barr RG, Brunet A, Galbaud du Fort G, Meaney M,
Saucier J-F, et al. 2004. Stress during pregnancy affects
intellectual and linguistic functioning in human toddlers. Pediatr
Res 56:400– 410.
Laplante DP, Zelazo PR, Brunet A, King S. (submitted). Functional
play at 2 years of age: Effects of prenatal maternal stress. J Am
Acad Child Adolesc Psychiatry.
Lazarus RS. 1991. Emotion and adaptation. New York: Oxford
University Press.
Lechat MF. 1979. Disasters and public health. Bull World Health
Organ 57:11– 17.
Lou HC. 1993. Prenatal stressful events of human life affect fetal
brain development. Neuropediatrics 24(Abstract):180.
Lou HC, Hansen D, Nordentoft M, Pryds O, Jensen F, Nim J, et al.
1994. Prenatal stressors of human life affect brain development.
Dev Med Child Neurol 36:826 –832.
Maccari S, Piazza PV, Kabbaj M, Barbazanges A, Simon H, Le Moal
M. 1995. Adoption reverses the long-term impairment in
glucocorticoid feedback induced by prenatal stress. J Neurosci
15:110–116.
McFarlane AC. 1988. Relationship between psychiatric impairment
and a natural disaster: The role of distress. Psychol Med
18(1):129–139.
McIntosh DE, Mulkins RS, Dean RS. 1995. Utilization of maternal
perinatal risk indicators in the differential diagnosis of ADHD
and UADD children. Int J Neurosci 81:35–46.
Mednick SA, Machon RA, Huttunen MO, Bonett D. 1988. Adult
schizophrenia following prenatal exposure to an influenza
epidemic. Arch Gen Psychiatry 45:189–192.
Mulder EJH, Robles de Medina PG, Huizink AC, Van den Bergh
BRH, Buitelaar JK, Visser GHA. 2002. Prenatal maternal stress:
Effects on pregnancy and the (unborn) child. Early Hum Dev
70:3–14.
Nishio H, Kasuga S, Ushijima M, Harada Y. 2001. Prenatal stress
and postnatal development of neonatal rats—sex-dependent
effects on emotional behavior and learning ability of neonatal
rats. Int J Dev Neurosci 19(1):37–45.
O’Connor TG, Heron J, Golding J, Beveridge M, Glover V.
2002. Maternal antenatal anxiety and children’s behavioural/
emotional problems at 4 years. Br J Psychiatry 180:
502 – 508.
O’Connor TG, Heron J, Golding J, Glover V, Team AS. 2003.
Maternal antenatal anxiety and behavioural/emtional problems
in children: A test of a programming hypothesis. J Child Psychol
Psychiatry 44:1025–1036.
Paarlberg KM, Vingerhoets AJ, Passchier J, Dekker GA, Van Geijn
HP. 1995. Psychosocial factors and pregnancy outcome: A review
with emphasis on methodological issues. J Psychosom Res
39(5):563–595.
Paarlberg KM, Vingerhoets AJ, Passchier J, Dekker GA, Heinen
AG, van Geijn HP. 1999. Psychosocial predictors of low
birthweight: A prospective study. Br J Obstet Gynaecol
106(8):834–841.
Palinkas LA, Petterson JS, Russell J, Downs MA. 1993. Community
patterns of psychiatric disorders after the Exxon Valdez oil spill.
Am J Psychiatry 150:1517– 1523.
Rowe DC. 1994. The limits of family influence: Genes, experience,
and behavior. New York: Guilford.
Rubonis AV, Bickman L. 1991. Psychological impairment in the
wake of disaster: The disaster-psychopathology relationship.
Psychol Bull 109:384–399.
Rutter M, Silberg J. 2002. Gene-environment interplay in relation to
emotional and behavioral disturbance. Annu Rev Psychol
53:463–490.
Sarason IG, Johnson JH, Siegel JM. 1978. Assessing the impact of
life changes: Development of the life experience survey. J Consult
Clin Psychol 46(5):932–946.
S. King & D. P. Laplante44
Schneider ML. 1992a. Delayed object permanence development in
prenatally stressed rhesus monkey infants (Macaca mulatta).
Occup Ther J Res 12(2):96–110.
Schneider ML. 1992b. Prenatal stress exposure alters postnatal
behavioral expression under conditions of novelty challenge in
rhesus monkeys. Dev Psychobiol 25:529–540.
Schneider ML, Coe CL. 1993. Repeated social stress during
pregnancy impairs neuromotor development in the primate
infant. J Dev Behav Pediatr 14(2):81– 87.
Schneider ML, Coe CL, Lubach GR. 1992. Endocrine activation
mimics the adverse effects of prenatal stress on the neuromotor
development of the infant primate. Dev Psychobiol 15:427–439.
Schneider ML, Roughton EC, Koehler AJ, Lubach GR. 1999.
Growth and development following prenatal stress exposure in
primates: An examination of ontogenetic vulnerability. Child
Dev 70:263– 274.
Stott DH. 1973. Follow-up study from birth of the effects of
prenatal stresses. Dev Med Child Neurol 15:770– 787.
Takahashi LK, Kalin NH. 1991. Early developmental and
temporal characteristics of stress-induced secretion of pituitary–
adrenal hormones in prenatally stressed rat pups. Brain Res
558:75–78.
Trautman PD, Meyer-Bahlburg HFL, Postelnek J, New MI. 1995.
Effects of early prenatal dexamethasone on the cognitive and
behavioral development of young children: Results of a pilot
study. Psychoneuroendocrinology 20(4):439 – 449.
Uno H, Tarara R, Else J, Sulemen M, Sapolsky RM. 1989.
Hippocampal damage associated with prolonged and fatal stress
in primates. J Neurosci 9:1705 –1711.
Uno H, Lohmiller L, Thieme C, Kemnitz JW, Engle MJ, Roecker
EB, et al. 1990. Brain damage induced by prenatal exposure to
dexamethasone in fetal rhesus macaques, 1. Hippocampus.
Dev Brain Res 53:157– 167.
Uno H, Eisele S, Sakai A, Shelton S, Baker E, DeJesus O, et al.
1994. Neurotoxicity of glucocorticoids in the primate brain.
Horm Behav 28:336–348.
Van Os J, Selton J. 1998. Prenatal exposure to maternal stress and
subsequent schizophrenia. Br J Psychiatry 172:324–326.
Van den Bergh BRH, Marcoen A. 2004. High antenatal maternal
anxiety is related to ADHD symptoms, externalizing problems,
and anxiety in 8- and 9-year olds. Child Dev 75(4):1085–1097.
Watson JB, Mednick SA, Huttunen M, Wang X. 1999. Prenatal
teratogens and the development of adult mental illness. Dev
Psychopathol 11(3):457–466.
Weinberger DR. 1995. Schizophrenia as a neurodevelopmental
disorder. In: Hirsch SR, Weinberger DR, editors. Schizophrenia.
Oxford: Blackwell Science.
Weinstock M. 2001. Alterations induced by gestational stress in
brain morphology and behaviour of the offspring. Prog
Neurobiol 65:427– 451.
Weinstock M, Matlina A, Maor GI, Rosen H, McEwan BS. 1992.
Prenatal stress selectively alters the reactivity of the hypothala-
mic–pituitary– adrenal system in the female rat. Brain Res
595:195–200.
Weiss DS, Marmar CR. 1997. The impact of event scale—revised.
New York: Guilford.
Weller A, Glaubman H, Yehuda S, Caspy T, Ben-Uria Y. 1988.
Acute and repeated gestational stress affect offspring learning
and activity in rats. Physiol Behav 43(2):139 –143.
Zelazo PR, Kearsley RB. 1980. The emergence of functional plat in
infants: Evidence for a major cognitive transition. J Appl Dev
Psychol 1:95– 117.
Effects of prenatal maternal stress 45
