Content uploaded by Ana C. Hanlon-Dearman
Author content
All content in this area was uploaded by Ana C. Hanlon-Dearman on Mar 04, 2016
Content may be subject to copyright.
RESEARCH ARTICLE
Relationships between Head Circumference,
Brain Volume and Cognition in Children with
Prenatal Alcohol Exposure
Sarah Treit
1
*, Dongming Zhou
2
, Albert E. Chudley
3
, Gail Andrew
4,5
, Carmen Rasmussen
4
,
Sarah M. Nikkel
6
, Dawa Samdup
7
, Ana Hanlon-Dearman
8
, Christine Loock
9
,
Christian Beaulieu
1,2
1Neuroscience and Mental Health Institute, University of Alberta, Edmonton, Alberta, Canada,
2Department of Biomedical Engineering, University of Alberta, Edmonton, Alberta, Canada, 3Departments
of Pediatrics and Child Health and Biochemistry and Medical Genetics, University of Manitoba, Winnipeg,
Manitoba, Canada, 4Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada, 5FASD
Diagnostic Clinic, Glenrose Rehabilitation Hospital, Edmonton, Alberta, Canada, 6Department of Pediatrics,
University of Ottawa, Ottawa, Ontario, Canada, 7Department of Pediatrics, Queens University, Kingston,
Ontario, Canada, 8Department of Pediatrics and Child Health, University of Manitoba, Winnipeg, Manitoba,
Canada, 9Department of Pediatrics, University of British Columbia and Sunny Hill Health Centre for
Children, Vancouver, British Columbia, Canada
*treit@ualberta.ca
Abstract
Head circumference is used together with other measures as a proxy for central nervous
system damage in the diagnosis of fetal alcohol spectrum disorders, yet the relationship
between head circumference and brain volume has not been investigated in this population.
The objective of this study is to characterize the relationship between head circumference,
brain volume and cognitive performance in a large sample of children with prenatal alcohol
exposure (n = 144) and healthy controls (n = 145), aged 5–19 years. All participants under-
went magnetic resonance imaging to yield brain volumes and head circumference, normal-
ized to control for age and sex. Mean head circumference, brain volume, and cognitive
scores were significantly reduced in the prenatal alcohol exposure group relative to con-
trols, albeit with considerable overlap between groups. Males with prenatal alcohol expo-
sure had reductions in all three measures, whereas females with prenatal alcohol exposure
had reduced brain volumes and cognitive scores, but no difference in head circumference
relative to controls. Microcephaly (defined here as head circumference !3rd percentile)
occurred more often in prenatal alcohol exposed participants than controls, but 90% of the
exposed sample had head circumferences above this clinical cutoff indicating that head cir-
cumference is not a sensitive marker of prenatal alcohol exposure. Normalized head cir-
cumference and brain volume were positively correlated in both groups, and subjects with
very low head circumference typically had below-average brain volumes. Conversely, over
half of the subjects with very low brain volumes had normal head circumferences, which
may stem from differential effects of alcohol on the skeletal and nervous systems. There
were no significant correlations between head circumference and any cognitive score.
These findings confirm group-level reductions in head circumference and increased rates of
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 1 / 15
OPEN ACCESS
Citation: Treit S, Zhou D, Chudley AE, Andrew G,
Rasmussen C, Nikkel SM, et al. (2016) Relationships
between Head Circumference, Brain Volume and
Cognition in Children with Prenatal Alcohol Exposure.
PLoS ONE 11(2): e0150370. doi:10.1371/journal.
pone.0150370
Editor: Jaroslaw Harezlak, Indiana University,
UNITED STATES
Received: August 11, 2015
Accepted: February 12, 2016
Published: February 29, 2016
Copyright: © 2016 Treit et al. This is an open access
article distributed under the terms of the Creative
Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: Raw data is presented
in plots in the paper. The actual MRI images cannot
be publicly shared as per ethics requirements from
the University of Alberta ethics committee and
restrictions for vulnerable populations. Processed
data (i.e., head circumference and/or brain volume
values) can be made available upon request; please
contact treit@ualberta.ca for details.
Funding: This study was funded by the Canadian
Institutes for Health Research (CIHR), the Networks
of Centres for Excellence (via NeuroDevNet), and the
Women and Children’s Health Research Institute
microcephaly in children with prenatal alcohol exposure, but raise concerns about the pre-
dictive value of this metric at an individual-subject level.
Introduction
Head circumference (HC) trajectories are commonly used as a gross measure of neurological
development in infancy and early childhood, providing a rapid and cost-effective means to
screen for abnormalities such as hydrocephalus or delayed development [1]. Although not typi-
cally used in routine care beyond 3 years of age, HC is shown to positively correlate with brain
volume in groups of healthy children and adolescents. However, this relationship is noted to
weaken after the age of 7 years, likely given that total brain volume expansion peaks early in
childhood while skull thickness and non-neural tissue growth continue throughout adoles-
cence [2,3]. HC is often smaller in children with mental retardation [4] and has been correlated
with intelligence in healthy populations [3], though relationships between HC and cognitive
ability appear to be less consistent [5–7]. Nonetheless, large deviations from the norm at any
age can indicate micro- or macrocephaly associated with a multitude of genetic disorders, peri-
natal brain injuries, and teratogenic exposures [8,9] including prenatal alcohol exposure
(PAE).
Neurotoxicity from prenatal alcohol exposure is sometimes observable at birth, as evidenced
by reduced birth weight [10], lower Apgar scores [11] and increased rates of microcephaly (low
HC) in infants with PAE [12]. HC reductions persist throughout childhood and adolescence,
and one study has shown correlations between HC and performance IQ in 8–16 year olds with
PAE [13]. Several studies of PAE suggest that the degree of HC reduction relates to timing
[12], amount [14,15] and pattern [16,17] of alcohol exposure in utero, though HC reductions
are typically modest (e.g. -1.3 to -3.9% in children of heavy drinkers compared to abstainers
[14,16,18]) and are not present in every sample [19].
Nonetheless, HC is used (among other measures) in the diagnosis of fetal alcohol spectrum
disorders (FASD) as “evidence of deficient brain growth or abnormal morphogenesis”[20],
“structural evidence of CNS damage”[21] or “structural CNS dysfunction”[22]. However, the
relationship between HC and brain volume has not been reported in children with PAE and
may differ from observations in typically developing children or children with microcephaly of
other etiologies. Further investigation of this relationship is needed to better characterize the
clinical significance of microcephaly in children with PAE and to inform the use of HC in
FASD diagnostic guidelines.
Methods
Participants
This study was approved by the Health Research Ethics Boards at the University of British
Columbia, University of Alberta, University of Manitoba and Queens University. Participants
were 144 individuals with confirmed PAE (5–19 years, mean 12.5 ± 3.3 years; 76 males) and
145 controls (5–19 years, mean 11.9 ± 3.4 years; 69 males). This subject pool includes partici-
pants from previous [23–25] and current FASD studies conducted at the University of Alberta,
as well as from a new multi-site MRI study of brain development [26]. PAE participants were
recruited through various multi-disciplinary FASD diagnostic clinics across Canada, had con-
firmed prenatal alcohol exposure and were assessed according to the Canadian Guidelines for
the Diagnosis of FASD [22] and the 4-Digit Code [21]. Of the 144 participants in the PAE
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 2 / 15
(WCHRI). Salary support was provided by Alberta
Innovates Health Solutions to co-authors ST & CB.
All participants or parents/legal guardians provided
written informed consent, and participants were
screened for contraindications to MRI.
Competing Interests: The authors have declared
that no competing interests exist.
group, 33 (23%) had a dysmorphic diagnosis of fetal alcohol syndrome (FAS) or partial fetal
alcohol syndrome (pFAS), 79 (55%) were diagnosed with static encephalopathy: alcohol
exposed (SE:AE), neurobehavioural disorder: alcohol exposed (NBD:AE), alcohol related neu-
rodevelopmental disorder (ARND), or FASD that was not further specified, and 32 (22%) had
confirmed pre-natal alcohol exposure but did not meet criteria for formal diagnosis or were
deferred for re-evaluation. Controls were recruited through advertising and had no self-
reported history of neurological, psychiatric, or developmental disorders. All participants or
parents/legal guardians provided written informed consent, and participants were screened for
contraindications to MRI.
Image Acquisition and Analysis
Head circumference and brain volumes were calculated from magnetic resonance imaging
(MRI) scans (T1-weighted 3D-MPRAGE, 1x1x1 mm
3
) collected on 4 scanners across Canada:
3T Philips Intera at University of British Columbia (12 PAE and 16 controls), 1.5T Siemens
Sonata at University of Alberta (100 PAE and 106 controls), 3T Siemens Trio at each of Uni-
versity of Manitoba (9 PAE and 8 controls) and Queens University (23 PAE and 15 controls).
Total brain volume (excluding brainstem, cerebellum and cerebrospinal fluid) and the volume
of the frontal, temporal, parietal and occipital lobes were calculated with Freesurfer v5.1, aver-
aging left and right hemispheres. HC was manually traced by the same user (ST) on an axial
oblique slice aligned with the most prominent parts of the occiput and forehead in OsiriX
v5.8.5.
Inter-site Reliability
Given that data were collected on multiple MRI scanners, reliability of HC and brain volume
measurements between scanners was assessed with a ‘travelling phantom’study. In short, 8
adult subjects were flown to all four sites to be scanned twice each (8 scans per person, 64 scans
total, mean of 102 days from first to last scan), using the identical scanning protocol as this
study. Imaging data was analyzed with identical methods as described here, and reliability was
evaluated by computing the within and between-subject coefficient of variation (CV) and
Intraclass Correlation Coefficients (ICC; absolute agreement) for brain volume and head
circumference.
Cognitive Testing
Cognitive testing was performed by a trained research assistant on the same day as each sub-
jects’MRI scan. The test battery included: Woodcock Johnson (WJ) Quantitative Concepts;
Woodcock Reading Mastery Test-Revised (WRMT-R) Word ID; Working Memory Test Bat-
tery-Children (WMTB-C) Digit and Block recall; NEPSY-II Animal Sorting, Auditory Atten-
tion, Inhibition, and Memory for Names; Behavior Rating Inventory of Executive Function
(BRIEF) Parent form and the Wide Range Intelligence Test (WRIT) General IQ (Table 1). For
a small subset of PAE subjects (n = 12), Wechsler Intelligence Scale for Children (WISC) full
scale IQ was instead collected via chart review.
Normalization of Cognitive Scores, HC and Brain Volume
Raw cognitive scores were converted to standard, scaled or t-scores, according to the proce-
dures outlined by each test manual. Raw HC values were converted to standard deviations
(SDs) and percentiles based on a large population based sample [1], in order to control for age
and sex. Given that there are no normative standards for brain volume, raw brain volumes
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 3 / 15
were converted to Z scores based on the control group mean and standard deviation, calculated
separately in males and females.
Group Differences, Change with Age, and Correlations Between HC,
Brain Volume and Cognition
Group differences (PAE vs control) in normed cognitive scores, raw HC and raw brain volume
were determined with independent sample t-tests, while change with age was tested with Pear-
son’s correlations (alpha set to p<0.05); in these cases males and females were analyzed sepa-
rately. Relationships between normalized HC (SD), brain volume Z scores, and cognitive test
Table 1. Subject Characteristics and Cognitive Test Scores.
Control PAE PAE with HC !3
rd
percentile
a
Sample size 145 144 14
Age (years) 11.9 ±3.4 12.5 ±3.3 13.9 ±3.3
Number of Males 69 (48%) 76 (53%) 8 (57%)
Ethnicity
Caucasian 126 (87%) 38 (26%) 4 (33%)
Aboriginal 4 (3%) 72 (50%) 7 (50%)
Other/Unknown 15 (10%) 34 (24%) 3 (25%)
Wide Range Intelligence Test/Weschler
Intelligence Scale
c
General IQ 112 ±12 (n = 66) 88 ±17** (n = 50) —
e
Woodcock Johnson
b
Quantitative Concepts 18A&B 106 ±15 (n = 141) 82 ±19** (n = 122) 89 ±18 (n = 11)
Woodcock Reading Mastery Test
b
Word ID 106 ±13 (n = 140) 90 ±15** (n = 110) 93 ±13 (n = 10)
BRIEF (parent form)
c
Behavioural Regulation Index 48 ±8 (n = 131) 72 ±12** (n = 116) 76 ±13 (n = 11)
Metacognitive Index 51 ±12 (n = 130) 68 ±10** (n = 116) 70 ±12 (n = 11)
Global Executive Composite 49 ±10 (n = 130) 73 ±10** (n = 116) 75 ±11 (n = 11)
Working Memory Test Battery
b
Digit 99 ±16 (n = 125) 85 ±13** (n = 108) 88 ±9 (n = 7)
Block 100 ±16 (n = 124) 86 ±16** (106) 88 ±15 (n = 8)
NEPSY-II
d
Animal Sorting 9.4 ±3.8 (n = 129) 7.0 ±3.4** (n = 86) 8.3 ±5.6 (n = 7)
Auditory Attention 10.3 ±3.1 (n = 135) 7.1 ±4.1** (n = 103) 7.9 ±5.1 (n = 9)
Response Set 10.1 ±3.5 (n = 129) 8.9 ±3.9 (n = 101) 10.7 ±2.8 (n = 9)
Inhibition-Naming 9.3 ±3.5 (n = 135) 6.6 ±4.0** (n = 99) 8.4 ±5.0 (n = 8)
Inhibition-Inhibition 9.7 ±3.8 (n = 135) 6.1 ±3.6** (n = 98) 6.4 ±4.2 (n = 8)
Inhibition-Switching 10.2 ±3.8 (n = 129) 6.4 ±4.1** (n = 96) 6.4 ±6.5 (n = 8)
Memory for Names 9.3 ±2.9 (n = 135) 6.1 ±3.6** (n = 105) 6.7 ±4.5 (n = 9)
**p<0.001 on independent sample t-tests (PAE versus Controls)
a
Relative to population norms reported in Rollins et al Journal of Pediatrics, 2010 [1]. Significance versus controls not tested for HC!3
rd
percentile due to
small sample size (n = 3)
b
Standard scores, mean = 100, SD = 15, higher score indicates better performance
c
T scores, mean = 50, SD = 10, higher score indicates worse performance
d
Scaled scores, mean = 10, SD = 3, higher score indicates better performance
e
Mean ±SD not reported given n = 3 with IQ scores in this category
doi:10.1371/journal.pone.0150370.t001
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 4 / 15
scores were explored with Pearson’s correlations, assessed separately for PAE and controls;
note that the standard deviation and Z values have already factored sex into account.
In addition, correlations between normed HC and brain volume Z scores were then
repeated including only subjects with HC at least 1 SD below the population norm (below the
~15
th
percentile) to further test this relationship outside of the normal range. Although a diag-
nosis of microcephaly is not given unless HC is !10
th
percentile [20] or !3
rd
percentile
[21,22], this more liberal cutoff was chosen to increase power in a subset of PAE participants
on the low end of the HC spectrum (n = 34), given that the clinical cutoffs of the 10
th
or 3
rd
per-
centile would limit the sample to 22 and 14 participants, respectively. Instead descriptive statis-
tics are presented for participants below these clinical cut-offs. Specifically, the proportion of
subjects with HC !3
rd
percentile, !10
th
percentile, and 11
th
-99
th
percentile that have brain
volumes that are !3
rd
percentile, !10
th
percentile, and 11
th
-99
th
percentile are reported sepa-
rately, as well as proportions of subjects with brain volume HC !3
rd
percentile, !10
th
percen-
tile, and 11
th
-99
th
percentile who have head circumferences that are !3
rd
percentile, !10
th
percentile, and 11
th
-99
th
percentile.
Results
Inter-site Reliability
The travelling phantom study (of 8 adult subjects scanned twice each at each of the 4 sites)
indicated excellent absolute agreement between scanners for both brain volume and head cir-
cumference, as reflected by ICCs of 0.994 and 0.995, respectively (Fig 1). Mean within-subject
variability (across scanners) was 1.5% for brain volume and 0.4% for head circumference; for
both measures this was roughly 5 times lower than the between-subject variability of 7.7% and
2.5%, respectively. These results indicate little effect of scanner on these measurements, and
suggest that data can be combined across sites with limited risk of introducing site-related bias.
Group Differences and Changes with Age
Raw HC increased with age in both groups and sexes (male controls R = 0.53 p<0.001; female
controls R = 0.51, p<0.001; male PAE R = 0.45, p<0.001; female PAE R = 0.44, p<0.001- Fig
2A and 2B), but raw brain volume did not change with age (Fig 2D and 2E). Raw HC was
2.2% lower in males with PAE versus male controls, but females with PAE did not differ from
female controls (t = -3.81, p<0.001 and t = -1.58, p = 0.117, respectively—Fig 2C). Conversely,
raw brain volume was lower in both males (-8.8%) and females (-5.1%) with PAE relative to
controls (t = -6.47, p<0.001 and t = -3.32, p = 0.001, respectively—Fig 2F), though greater
overlap between sexes can be seen in the PAE than control group (Fig 2D and 2E). In the sub-
set of participants with IQ scores, IQ did not change with age (as expected for a standard score)
and sex differences were not significant in either group (Fig 2G–2I), though the control group
(IQ~112) significantly outperformed the PAE group (IQ~88) (t = -8.73, p<0.001- Table 1,Fig
2I). Likewise, the control group scored better than the PAE group on all other cognitive tests
(t = -2.37–19.19, p<0.001–0.019—Table 1).
Normed HC, brain volume and IQ scores had right-shifted distributions toward higher val-
ues in controls relative to the lower values in the PAE group (Fig 3A–3C). Notably, only 10%
of controls had a HC more than 1 SD below the population norm, compared to 24% of the
PAE group. The greatest reductions of HC, IQ and brain volume were found in participants
with dysmorphic features indicative of FAS/pFAS (data not shown) in keeping with previous
literature [27]; however, subgroup analysis was not further explored given that IQ and HC are
used in the sub-classification of FASD. Despite group differences, substantial overlap between
the PAE and control groups is evident for all three metrics (Figs 2and 3).
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 5 / 15
Head Circumference Correlations
Normed HC standard deviation and brain volume Z scores were positively correlated in both
the control (R = 0.66, p<0.001) and PAE groups (R = 0.60, p<0.001), indicating that on the
whole, subjects with larger HC for their age/sex have larger brain volumes (Fig 4A and 4B).
However, this correlation did not hold among PAE participants with HC more than 1 SD
below the norm (n = 34), or in all subjects (PAE and Control combined) with HC more than 1
SD below the norm (n = 49). Of note, similar R value correlations were found when analysis
was repeated with partial correlations controlling for age (data not shown). IQ did not signifi-
cantly correlate with normed HC (controls R = 0.22, p = 0.080; PAE R = 0.24 p = 0.093- Fig 5A
and 5B), total brain volume Z scores (controls R = 0.09, p = 0.479; PAE R = 0.21, p = 0.146), or
any lobe volume Z scores (controls R = -0.01–0.16, p = 0.957–0.202; PAE R = 0.19–0.26,
p = 0.190–0.072). IQ was only available in 50 PAE and 66 control participants; however, this
subsample had a similar age, sex and diagnostic sub-group distribution as the total sample (Fig
2G and 2H), and demonstrated tight correlations between brain volume Z scores and normed
HC (Fig 5C and 5D), indicating that these negative findings (no relationships between HC or
brain volume and IQ) are unlikely to stem from sample bias. Likewise, no other cognitive
scores correlated with normed HC or total brain volume Z scores in either group, despite larger
sample sizes of n~90–140 in each group.
PAE Participants with Head Circumference Below the 3
rd
Percentile
Only a small minority of PAE participants (14/144, 10%) met the clinical definition of micro-
cephaly with HC !3
rd
percentile. Of these 14 subjects, 11 (80%) had brain volume !3
rd
per-
centile (Fig 6A). Similar proportions are observed for participants with HC !10
th
percentile.
Conversely, of the PAE participants with brain volume !3
rd
percentile (n = 28), only ~35%
had HC !3
rd
percentile and more than half had HC percentiles in the normal range (Fig 6B).
However, given that HC and brain volume were normed on different scales (population based
Fig 1. Inter-site reliability of (A) brain volume and (B) head circumference measurements from 8 adult subjects, scanned twice each at all four sites. For both
brain volume and head circumference measurements, within-subject variability was much lower than between subject variability (as reflected by ICCs of
0.994 and 0.995, respectively) suggesting that data can be combined across scanners for these measures without obvious bias.
doi:10.1371/journal.pone.0150370.g001
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 6 / 15
Fig 2. Head circumference increased with age for both males (blue) and females (red) in both groups (A, B), and HC was lower in PAE than controls for
males but not females (C). Brain volume did not change with age in either group (D,E), and was consistently reduced in PAE relative to controls for both
males and females (F). IQ standard scores did not change with age in either group (G, H), and were again lower in the PAE group (I). Sex differences within
groups were larger in the control group for both head circumference (C) and brain volume (F). ns = non-significant.
doi:10.1371/journal.pone.0150370.g002
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 7 / 15
sample versus control group, respectively) direct comparison may be confounded by group dif-
ferences between our controls and the population norm. Nonetheless, only 50% subject overlap
was found between the 14 subjects with HC !3
rd
percentiles and the 14 subjects with the low-
est brain volume percentiles in the PAE group, again suggesting a disconnect between these
metrics. Cognitive test scores of the 14 PAE participants who had HC !3
rd
percentile were not
different than the whole PAE group (Table 1), suggesting that these participants are not more
cognitively impaired than those PAE subjects with HC in the normal range.
Fig 3. (A) Head circumference (HC) standard deviation distributions showing a shift towards the number of
participants with higher normed HC in controls and lower normed HC in PAE subjects, albeit with substantial
overlap between groups. (B) Z score distribution for total brain volume is left- shifted in the PAE group (grey
curve) relative to controls (black curve). PAE distribution curves for brain lobe volumes show similar leftward
shifts towards negative Z scores. (C) Likewise, IQ score profile is right-shifted in the control group compared
to PAE (C), peaking above the population norm of 100 in controls, and below in the PAE group.
doi:10.1371/journal.pone.0150370.g003
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 8 / 15
Discussion
This study provides evidence to support previous assertions of reduced head circumference in
children and adolescents with prenatal alcohol exposure [14–18] by demonstrating: (i) a
29-point gap in median HC percentile, (ii) a 2-fold increase in the number of subjects with HC
more than 1 SD below the population mean, and (iii) a nearly 5-fold increase in the number of
subjects with HC !3
rd
percentile in our PAE group compared to controls. This large sample
of 144 PAE participants spans a wide range from 5 to 19 years of age, suggesting that HC defi-
cits persist into young adulthood, in agreement with longitudinal studies of PAE [28]. Likewise,
significant reductions of brain volume, IQ and cognitive performance are demonstrated in the
PAE group, in keeping with previous literature in this population [29].
However, these results also highlight several important limitations of HC measurement in
children with PAE. Despite group differences, the substantial spread in HC and its overlap
between groups suggests that this metric does not discriminate individuals with PAE from
healthy controls at a single-subject level. Moreover, although microcephaly (HC !3
rd
percen-
tile) occurred at a higher frequency in the PAE than control group, 90% of the PAE group had
HC values above this clinical cutoff. Rates of microcephaly have previously been found to be
low even in very large samples, e.g. 64 of 973 (~6.5%) PAE participants had HC<10
th
percen-
tile [12], again suggesting that microcephaly is not a sensitive marker of PAE.
In both the controls and PAE groups, HC is shown to positively correlate with brain vol-
ume; this relationship is not surprising—a child with a very low HC would be expected to have a
smaller brain than a child on the high end of the HC range. However, increased variability in this
relationship at the lower end of the HC range suggests that HC is a poor predictor of brain vol-
ume among groups of children with roughly similar normed HC values. Furthermore, lack of
correlations between HC and cognitive measures suggests that HC does not predict functional
impairment. As such, further investigation may be needed to determine if HC deficits are indeed
a reflection of central nervous system impairment rather than overall growth deficiency.
Among the 14 PAE children with clinically significant microcephaly (HC!3
rd
percentile),
there is no greater impairment in cognition (Table 1) and three participants have normal brain
Fig 4. Brain volume Z scores and normed head circumference (HC) standard deviations are shown to positively correlate in both the control (A) and prenatal
alcohol exposure (PAE) groups (B).
doi:10.1371/journal.pone.0150370.g004
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 9 / 15
volumes (defined as above the 10
th
percentile). Nonetheless, most (11/14) have brain volumes
below 10
th
percentile, as expected given that HC (i.e. the skull) poses a physical limitation on
brain volume. However, it is important to note that the reverse relationship is not observed:
about 60% of the subjects with very small brain volumes are shown to have normal HC values
(again defined as above the 10
th
percentile).
The lack of consistent correlation between brain volume and HC at low normed HC values
may reflect greater variability in the ratio of brain tissue to cerebrospinal fluid in children with
PAE, which could explain cases where brain volume is very small but HC is in the normal
range. Conversely, reduced bone volume in the skull has been observed with micro computed
tomography in a mouse model of FASD [30], and may account for cases where HC was
reduced with intact brain volume. The skeletal and nervous systems are each uniquely affected
by the toxic effects of alcohol in utero [31–33], likely adding variability to the relationship
between brain volume and HC in children who were exposed to alcohol in varying amounts,
frequencies and time-points throughout pregnancy. A recent study in autism found that non-
neural tissue volume (skull, meninges, cerebrospinal fluid, etc.) was more highly correlated to
Fig 5. Age-standardized IQ versus normed head circumference standard deviation, showing non-significant
relationships in both the control (A) and PAE groups (B), despite significant correlations between brain
volume Z score and normed HC SD in this subset of participants (C,D). Likewise, there does not appearto be
any systematic pattern/grouping of IQ in controls (C) or PAE (D) again demonstrating that those with the
smallest normed HC and brain volume did not show consistently lower IQ scores.
doi:10.1371/journal.pone.0150370.g005
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 10 / 15
HC than total brain volume, and that brain volume was only significantly correlated with HC
in the control group (n = 26), but not the autism group (n = 34) [34], suggesting a disconnect
in this relationship associated with another common neurodevelopmental disorder.
Beyond in utero sensitivity, the distinct post-natal trajectories of nervous and skeletal sys-
tem development may also impact the relationship between HC and brain volume. HC
increased with age at similar rates in both PAE and control groups, while brain volume did not
change significantly in either group from 5–19 years, fitting with previous literature demon-
strating that brain volume reaches ~90% of adult maximum at around age 6 years, while skull
thickness continues to increase linearly with age into adolescence [35]. A recent study of
healthy controls demonstrated that the brain-scalp distance increases with age in healthy
Fig 6. When only including the 14 PAE participants below the clinical cutoff for microcephaly
(HC !3
rd
percentile, A–column 1), it is notable that 10 (~70%) of the subjects have brain volumes
below the 3
rd
percentile. Similarly, ~80% of the subjects with HC !10
th
percentile have brain volumes
under the 10
th
percentile (A–column 2). Conversely, among the PAE participants with total brain volume !3
rd
(n = 14, B–column 1) or 10
th
percentile (n = 22, B—column 2), 55–65% of subjects have HC in the ‘normal’
range from the 11
th
-99th percentiles, suggesting a disconnect between small brain volumes and head
circumference. Note that the category of !10
th
percentile on the x-axis of A and B includes subjects who
are !3
rd
percentile (to match cutoffs used in various diagnostic guidelines), while colour divisions within each
bar are non-overlapping.
doi:10.1371/journal.pone.0150370.g006
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 11 / 15
children, driven primarily by increases in cerebrospinal fluid and cranial thickness [36], pro-
viding further evidence that HC reflects the composite of multiple systems that each develop at
different rates with age. Nonetheless, when assessed separately, the developmental trajectories
of HC and brain volume appear to be similar between the PAE and control groups, albeit in
cross-sectional cohorts. Longitudinal samples may be better positioned to tease apart the rela-
tionship between these trajectories in PAE.
In addition to group differences, sex effects were observed for both brain volume and HC in
the control group, as expected [37], but were less prominent between males and females in the
PAE group (Fig 2). Both males and females with PAE showed significant brain volume reduc-
tions, albeit with a greater difference in males, in keeping with previous findings of more sub-
stantial brain volume reductions in males with FASD [24,38]. Head circumference was also
reduced in males with PAE, but was not significantly lower in females with PAE, raising the
question of the value of HC measurement particularly for females. Age-by-sex interactions
have been observed in PAE studies of longitudinal cortical volume development [39], though
the mechanisms underlying sex effects in PAE are unclear. Greater overlap in both HC and
brain volume is observed between males and females in the PAE group across the whole age
range, with no apparent divergence between sexes during adolescence. However, the effects of
puberty were not tested but cannot be ruled out. Nonetheless, it remains possible that prenatal
alcohol exposure has sex-specific impacts on early nervous system development, as observed in
some animal models of PAE [40–42], and as suggested in developmental programming models
of disease [43].
Several limitations of this study must be acknowledged. First, converting brain volumes to Z
scores based on our control group (n = 145) may be less generalizable and more sensitive to
sample bias than head circumference norms that were normalized here on a much larger pub-
lished normative sample (n = 537) [1]. However, normative standards do not exist for brain
volume, and thus the control sample was used here. Secondly, although weight was routinely
collected prior to MRI acquisition, height was not consistently collected thus precluding exami-
nation of the effects of stature and/or growth deficiency on reduced head circumference in this
population. Thirdly, our groups were imbalanced with respect to ethnicity, with 50% of our
PAE group but only 3% of our control group self-identifying as aboriginal. Post-hoc analysis
comparing brain volume Z scores and head circumference standard deviations between the
aboriginal and non-aboriginal participants within the PAE group revealed no significant differ-
ences, suggesting that ethnicity does not have a strong effect on either metric. Moreover, group
differences (PAE versus Control) and brain volume-head circumference correlations were re-
tested after excluding aboriginal participants, which again yielded very similar results as those
reported in our main study with ethnicity categories combined. Specifically, brain volume and
head circumference were reduced (t = -7.4, p<0.001; t = -4.9, p<0.001, respectively) in non-
aboriginal PAE participants (n = 72) compared to controls (n = 145). Likewise, brain volume Z
score—head circumference SD correlations remained significant among non-aboriginal PAE
participants (R = 0.611, p<0.001; n = 72), and again did not hold when including only those
with head circumference below -1 SD from the population norm (R = 0.336, p = 0.187, n = 17).
Lastly, although the lack of correlation between head circumference and brain volume among
PAE participants with head circumference more than 1 SD below the population norm is
intriguing, it is important to keep in mind that this reduced sample size (n = 34) limits power
relative to the larger sample (n = 144) across the entire head circumference spectrum. Nonethe-
less, lack of correlation in this subset underscores the large inter-subject variability in this rela-
tionship, and provides further caution against its application to single-subject data for
detecting central nervous system dysfunction (brain volumes or cognitive performance) based
on head circumference.
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 12 / 15
Here we confirm previous reports of reduced head circumference, brain volume, and cogni-
tive function in a large cohort of children with PAE relative to age and sex matched controls.
Positive correlations are demonstrated between HC and brain volume, but the relationship
weakens outside of the normal range in PAE, which may reflect the complex interplay between
skeletal and neural development, each differentially affected by prenatal alcohol exposure. Fur-
ther, although microcephaly is clearly more common in the PAE population, our findings sug-
gest that it is only present in a small subset of children with PAE and does not co-occur with
greater cognitive impairments.
Acknowledgments
This study was funded by the Canadian Institutes for Health Research (CIHR), the Networks
of Centres for Excellence (via NeuroDevNet), and the Women and Children’s Health Research
Institute (WCHRI). Salary support was provided by Alberta Innovates Health Solutions to co-
authors ST & CB.
Author Contributions
Conceived and designed the experiments: ST AEC CB. Performed the experiments: ST. Ana-
lyzed the data: ST DZ. Wrote the paper: ST CB. Subject recruitment: AEC GA CR SM DS AH-
D CL. Clinical interpretation: AEC GA CR SM DS AH-D CL.
References
1. Rollins JD, Collins JS, Holden KR. United States Head Circumference Growth Reference Charts: Birth
to 21 Years. Journal of Pediatrics. 2010; 156(6):907–913. doi: 10.1016/j.jpeds.2010.01.009 PMID:
20304425
2. Bartholomeusz HH, Courchesne E, Karns CM. Relationship between head circumference and brain
volume in healthy normal toddlers, children, and adults. Neuropediatrics. 2002; 33(5):239–41. PMID:
12536365
3. Ivanovic DM, Leiva BP, Perez HT, Olivares MG, Diaz NS, Urrutia MSC, et al. Head size and intelli-
gence, learning, nutritional status and brain development Head, IQ, learning, nutrition and brain. Neu-
ropsychologia. 2004; 42(8):1118–31. PMID: 15093150
4. O'Connell EJ, Feldt RH, Stickler GB. Head Circumference Mental Retardation and Growth Failure.
Pediatrics. 1965; 36(1):62–66.
5. Wright CM, Emond A. Head Growth and Neurocognitive Outcomes. Pediatrics. 2015; 135(6):E1393–
E8. doi: 10.1542/peds.2014-3172 PMID: 25986019
6. Gale CR, O'Callaghan FJ, Godfrey KM, Law CM, Martyn CN. Critical periods of brain growth and cogni-
tive function in children. Brain. 2004; 127:321–9. PMID: 14645144
7. Gale CR, O'Callaghan FJ, Bredow M, Martyn CN, Avon Longitudinal Study Parents C. The influence of
head growth in fetal life, infancy, and childhood on intelligence at the ages of 4 and 8 years. Pediatrics.
2006; 118(4):1486–92. PMID: 17015539
8. Von der Hagen M, Pivarcsi M, Liebe J, Von Bernuth H, Didonato N, Hennermann JB, et al. Diagnostic
approach to microcephaly in childhood: a two-center study and review of the literature. Developmental
Medicine and Child Neurology. 2014; 56(8):732–41. doi: 10.1111/dmcn.12425 PMID: 24617602
9. Williams CA, Dagli A, Battaglia A. Genetic disorders associated with macrocephaly. American Journal
of Medical Genetics Part A. 2008; 146A(15):2023–37. doi: 10.1002/ajmg.a.32434 PMID: 18629877
10. Little RE. Moderate Alcohol use During Pregnancy and Decreased Infant Birth-Weight. American Jour-
nal of Public Health. 1977; 67(12):1154–6. PMID: 596498
11. Bagheri MM, Burd L, Martsolf JT, Klug MG. Fetal alcohol syndrome: maternal and neonatal characteris-
tics. Journal of Perinatal Medicine. 1998; 26(4):263–9. PMID: 9846300
12. Feldman HS, Jones KL, Lindsay S, Slymen D, Klonoff-Cohen H, Kao K, et al. Prenatal Alcohol Expo-
sure Patterns and Alcohol-Related Birth Defects and Growth Deficiencies: A Prospective Study. Alco-
holism. 2012; 36(4):670–6. doi: 10.1111/j.1530-0277.2011.01664.x PMID: 22250768
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 13 / 15
13. Ervalahti N, Korkman M, Fagerlund A, Autti-Ramo I, Loimu L, Hoyme HE. Relationship between dys-
morphic features and general cognitive function in children with fetal alcohol spectrum disorders. Amer-
ican Journal of Medical Genetics Part A. 2007; 143A(24):2916–23. PMID: 18000897
14. Russell M, Czarnecki DM, Cowan R, McPherson E, Mudar PJ. Measures of Maternal Alcohol-Use as
Predictors of Development in Early-Childhood. Alcoholism. 1991; 15(6):991–1000. PMID: 1724115
15. Smith IE, Coles CD, Lancaster J, Fernhoff PM, Falek A. The Effect of Volume and Duration of Prenatal
Ethanol Exposure on Neonatal Physical and Behavioral-Development. Neurobehavioral Toxicology
and Teratology. 1986; 8(4):375–81. PMID: 3762847
16. Day NL, Leech SL, Richardson GA, Cornelius MD, Robles N, Larkby C. Prenatal alcohol exposure pre-
dicts continued deficits in offspring size at 14 years of age. Alcoholism. 2002; 26(10):1584–91. PMID:
12394293
17. Sampson PD, Bookstein FL, Barr HM, Streissguth AP. Prenatal Alcohol Exposure, Birth-Weight, and
Measures of Child Size from Birth to Age 14 Years. American Journal of Public Health. 1994; 84
(9):1421–8. PMID: 8092365
18. Coles CD, Brown RT, Smith IE, Platzman KA, Erickson S, Falek A. Effects of Prenatal Alcohol Expo-
sure at School Age .1. Physical and Cognitive-Development. Neurotoxicology and Teratology. 1991;
13(4):357–67. PMID: 1921915
19. O'Callaghan FV, O'Callaghan M, Najman JM, Williams GM, Bor W. Maternal alcohol consumption dur-
ing pregnancy and physical outcomes up to 5 years of age: a longitudinal study. Early Human Develop-
ment. 2003; 71(2):137–48. PMID: 12663151
20. Hoyme HE, May PA, Kalberg WO, Kodituwakku P, Gossage JP, Trujillo PM, et al. A practical clinical
approach to diagnosis of fetal alcohol spectrum disorders: Clarification of the 1996 Institute of Medicine
Criteria. Pediatrics. 2005; 115(1):39–47. PMID: 15629980
21. Astley S. Diagnostic Guide for Fetal Alcohol Spectrum Disorders: The 4-Digit Diagnostic Code. Fetal
Alcohol Syndrome Diagnostic and Prevention Network, University of Washington, Seattle USA., 2004.
22. Chudley AE, Conry J, Cook LL, Loock C, Rosales T, LeBlanc N. Fetal alcohol spectrum disorder: Cana-
dian guidelines for diagnosis. Can Med Assoc J. 2005; 172(5):S1–S21.
23. Lebel C, Rasmussen C, Wyper K, Walker L, Andrew G, Yager J, et al. Brain diffusion abnormalities in
children with fetal alcohol spectrum disorder. Alcoholism. 2008; 32(10):1732–40. doi: 10.1111/j.1530-
0277.2008.00750.x PMID: 18671811
24. Nardelli A, Lebel C, Rasmussen C, Andrew G, Beaulieu C. Extensive Deep Gray Matter Volume
Reductions in Children and Adolescents with Fetal Alcohol Spectrum Disorders. Alcoholism. 2011; 35
(8):1404–17. doi: 10.1111/j.1530-0277.2011.01476.x PMID: 21575012
25. Zhou D, Lebel C, Lepage C, Rasmussen C, Evans A, Wyper K, et al. Developmental cortical thinning in
fetal alcohol spectrum disorders. Neuroimage. 2011; 58(1):16–25. doi: 10.1016/j.neuroimage.2011.06.
026 PMID: 21704711
26. Reynolds JN, Weinberg J, Clarren S, Beaulieu C, Rasmussen C, Kobor M, et al. Fetal Alcohol Spec-
trum Disorders: Gene-Environment Interactions, Predictive Biomarkers, and the Relationship Between
Structural Alterations in the Brain and Functional Outcomes. Seminars in Pediatric Neurology. 2011; 18
(1):49–55. doi: 10.1016/j.spen.2011.02.006 PMID: 21575841
27. Astley SJ, Aylward EH, Olson HC, Kerns K, Brooks A, Coggins TE, et al. Magnetic Resonance Imaging
Outcomes From a Comprehensive Magnetic Resonance Study of Children With Fetal Alcohol Spec-
trum Disorders. Alcoholism. 2009; 33(10):1671–89. doi: 10.1111/j.1530-0277.2009.01004.x PMID:
19572986
28. Carter RC, Jacobson JL, Sokol RJ, Avison MJ, Jacobson SW. Fetal Alcohol-Related Growth Restric-
tion from Birth through Young Adulthood and Moderating Effects of Maternal Prepregnancy Weight.
Alcoholism. 2013; 37(3):452–62. doi: 10.1111/j.1530-0277.2012.01940.x PMID: 23013325
29. Lebel C, Roussotte F, Sowell ER. Imaging the Impact of Prenatal Alcohol Exposure on the Structure of
the Developing Human Brain. Neuropsychol Rev. 2011; 21(2):102–18. doi: 10.1007/s11065-011-9163-
0PMID: 21369875
30. Shen L, Ai H, Liang Y, Ren X, Anthony CB, Goodlett CR, et al. Effect of prenatal alcohol exposure on
bony craniofacial development: A mouse MicroCT study. Alcohol. 2013; 47(5):405–15. doi: 10.1016/j.
alcohol.2013.04.005 PMID: 23809873
31. Hernandez-Guerrero JC, Ledesma-Montes C, Loyola-Rodriguez JP. Effects of maternal ethanol intake
on second alcoholic generation murine skull and mandibular size. Arch Med Res. 1998; 29(4):297–
302. PMID: 9887546
32. Goodlett CR, Horn KH, Zhou FC. Alcohol teratogenesis: Mechanisms of damage and strategies for
intervention. Experimental Biology and Medicine. 2005; 230(6):394–406. PMID: 15956769
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 14 / 15
33. Dunty WC, Chen SY, Zucker RM, Dehart DB, Sulik KK. Selective vulnerability of embryonic cell popula-
tions to ethanol-induced apoptosis: Implications for alcohol-related birth defects and neurodevelopmen-
tal disorder. Alcoholism. 2001; 25(10):1523–35. PMID: 11696674
34. Tate DF, Bigler ED, McMahon W, Lainhart J. The relative contributions of brain, cerebrospinal fluid-
filled structures and non-neural tissue volumes to occipital-frontal head circumference in subjects with
autism. Neuropediatrics. 2007; 38(1):18–24. PMID: 17607599
35. Giedd JN. Structural magnetic resonance imaging of the adolescent brain. In: Dahl RE, Spear LP, edi-
tors. Adolescent Brain Development: Vulnerabilities and Opportunities. Annals of the New York Acad-
emy of Sciences. 10212004. p. 77–85.
36. Beauchamp MS, Beurlot MR, Fava E, Nath AR, Parikh NA, Saad ZS, et al. The Developmental Trajec-
tory of Brain-Scalp Distance from Birth through Childhood: Implications for Functional Neuroimaging.
PLoS One. 2011; 6(9).
37. Giedd JN, Raznahan A, Mills KL, Lenroot RK. Review: magnetic resonance imaging of male/female dif-
ferences in human adolescent brain anatomy. Biology of Sex Differences. 2012; 3.
38. Chen XC, Coles CD, Lynch ME, Hu XP. Understanding specific effects of prenatal alcohol exposure on
brain structure in young adults. Human Brain Mapping. 2012; 33(7):1663–76. doi: 10.1002/hbm.21313
PMID: 21692145
39. Lebel C, Mattson SN, Riley EP, Jones KL, Adnams CM, May PA, et al. A Longitudinal Study of the
Long-Term Consequences of Drinking during Pregnancy: Heavy In Utero Alcohol Exposure Disrupts
the Normal Processes of Brain Development. Journal of Neuroscience. 2012; 32(44):15243–51. doi:
10.1523/JNEUROSCI.1161-12.2012 PMID: 23115162
40. Sickmann HM, Patten AR, Morch K, Sawchuk S, Zhang C, Parton R, et al. Prenatal Ethanol Exposure
has Sex-Specific Effects on Hippocampal Long-Term Potentiation. Hippocampus. 2014; 24(1):54–64.
doi: 10.1002/hipo.22203 PMID: 23996604
41. Uban KA, Comeau WL, Ellis LA, Galea LAM, Weinberg J. Basal regulation of HPA and dopamine sys-
tems is altered differentially in males and females by prenatal alcohol exposure and chronic variable
stress. Psychoneuroendocrinology. 2013; 38(10):1953–66. doi: 10.1016/j.psyneuen.2013.02.017
PMID: 23579081
42. Weinberg J, Sliwowska JH, Lan N, Hellemans KGC. Prenatal alcohol exposure: Foetal programming,
the hypothalamic-pituitary-adrenal axis and sex differences in outcome. Journal of Neuroendocrinol-
ogy. 2008; 20(4):470–88. doi: 10.1111/j.1365-2826.2008.01669.x PMID: 18266938
43. Aiken CE, Ozanne SE. Sex differences in developmental programming models. Reproduction. 2013;
145(1):R1–R13. doi: 10.1530/REP-11-0489 PMID: 23081892
Head Circumference-Brain Volume Correlations in PAE
PLOS ONE | DOI:10.1371/journal.pone.0150370 February 29, 2016 15 / 15
