Childhood Neglect Is Associated with Reduced Corpus

Callosum Area
Martin H. Teicher, Nathalie L. Dumont, Yutaka Ito, Catherine Vaituzis, Jay N. Giedd, and
Susan L. Andersen
Background: Childhood abuse has been associated with abnormalities in brain development, particularly corpus callosum (CC)
morphology. The impact of neglect has not been assessed, though it is the most prevalent form of childhood maltreatment.
Methods: Regional CC area was measured from magnetic resonance imaging scans in 26 boys and 25 girls admitted for psychiatric
evaluation (28 with abuse or neglect) and compared with CC area in 115 healthy control subjects. Data were analyzed by multivariate
analysis of covariance, with age and midsagittal area as covariates.
Results: Total CC area of the abused/neglected patients was 17% smaller than in control subjects (p ! .0001) and 11% smaller than
in psychiatric patients who had not been abused or neglected (contrast group; p ! .01). Control subjects and the contrast group did
not differ in total CC area. Neglect was the strongest experiential factor and was associated with a 15%–18% reduction in CC regions
3, 4, 5, and 7 (all p " .02). In contrast, sexual abuse seemed to be the strongest factor associated with reduced CC size in girls.
Conclusions: These data are consistent with animal research that demonstrated reduced CC size in nursery-reared compared with
semi-naturally reared primates. Early experience might also affect the development of the human CC.

Key Words: Abuse, neglect, corpus callosum, magnetic resonance dren and adolescents with PTSD and 61 matched control sub-
imaging, sex differences, children jects. The CC and its anterior and posterior midbody and
splenium were the regions and subregions that differed most

D
uring the last decade, there has been increasing interest
in the long-term adverse consequences of childhood
maltreatment. We and others have suggested that child-
hood physical or sexual abuse might be associated with dimin-
ished left hemisphere and left hippocampal development (Brem-
ner et al 1997; Ito et al 1998; Stein 1997), decreased right/left
cortical integration (Schiffer et al 1995; Teicher 1994), increased
incidence of electroencephalogram (EEG) abnormalities (Ito et al
1993), and diminished size of the corpus callosum (CC) (De
Bellis et al 1999; Teicher et al 1997). In contrast, the potential
neurobiological sequelae of early neglect as a distinct form of
abuse has been ignored.
In this study, we sought to ascertain whether there were
abnormalities in the regional anatomy of the CC in children with
a history of abuse or neglect. The CC is the major myelinated
fiber tract in the brain that connects the right and left hemi-
spheres. We hypothesized that the functional differences in the
lateralization of EEG coherence (Ito et al 1998; Teicher et al 1997)
and evoked potentials (Schiffer et al 1995) observed in children
and adults who suffered early abuse with or without neglect
might be associated with attenuated maturation of the CC
(Teicher et al 1997, 2002).
To date, studies on the morphology of the CC in maltreated
children have focused on the consequences of physical or sexual
abuse and have largely involved subjects who met criteria for
posttraumatic stress disorder (PTSD) (De Bellis et al 1999;
Teicher et al 1997). De Bellis et al (1999) conducted detailed
morphometric analyses of 43 structures or subregions from
magnetic resonance imaging (MRI) scans of 44 maltreated chil-
From the Department of Psychiatry (MHT, YI, SLA), Harvard Medical School,
Boston; Developmental Biopsychiatry Research Program (MHT, NLD,
SLA), McLean Hospital, Belmont, Massachusetts; and the Child Psychia-
try Branch (CV, JNG), National Institute of Mental Health, Bethesda, Mary-
land.
Address reprint requests to Martin H. Teicher, M.D., Ph.D., McLean Hospital,
Developmental Biopsychiatry Research Program, 115 Mill Street, Bel-
mont, MA 02478.
Received August 4, 2003; revised January 8, 2004; accepted March 17, 2004.
reliably between PTSD patients and control subjects, with the
greatest differences occurring in male subjects.
Posttraumatic stress disorder, however, is not a routine con-
sequence of childhood maltreatment. Only one third of physi-
cally or sexually abused children develop PTSD at any point in
their lifetime (Widom 1999). It is unclear from De Bellis et al
(1999) whether abnormalities in CC size relate to the effects of
maltreatment per se or emerge in the subset who go on to
develop PTSD. Previously, we had reported that physical or
sexual abuse was associated with abnormalities in the develop-
ment of the left hemisphere independent of whether or not
subjects met criteria for PTSD (Ito et al 1998).
The development of the CC is affected by a number of factors,
rendering it vulnerable to the environment or insult. Nerve fiber
connections passing through the CC are fully established before
birth, and experience-dependent pruning and elimination of
fibers through the CC is more than balanced by a rostral– caudal
pattern of myelination that increases callosal size and continues
into young adulthood (Giedd et al 1996). Very early experience
can dramatically influence the morphometry of the CC (Denen-
berg and Yutzey 1985). Corpus callosum size was reduced in
male primates that were isolated (nursery-reared) during early
development relative to semi-naturally reared primates (Sanchez
et al 1998). This suggests that the CC might be vulnerable to the
effects of neglect. The rostral– caudal myelination sequence also
suggests that different regions of the CC might have different
windows of vulnerability to early experience. Thus, the purpose
of the present study was to examine the relative contributions of
neglect, physical abuse, sexual abuse, PTSD, psychiatric illness,
and gender to the development of CC size.
Methods and Materials
Medical records were reviewed for 115 consecutive pediatric
patients admitted to McLean Hospital between June 1988 and
May 1989. Patients were eliminated for possible preexisting
neurologic abnormalities, including loss of consciousness, peri-
natal complications, in utero exposure to alcohol or drugs of
abuse, and physical illness that could affect growth or develop-

0006-3223/04/$30.00 BIOL PSYCHIATRY 2004;56:80 – 85

doi:10.1016/j.biopsych.2004.03.016 © 2004 Society of Biological Psychiatry
M.H. Teicher et al
ment (e.g., cancer or steroid-dependent asthma). A total of 51
subjects (26 male, 25 female) admitted for psychiatric evaluation
(28 with a history of abuse or neglect) were compared with 18
carefully screened healthy control subjects from McLean and 97
carefully screened healthy control subjects evaluated at the
National Institute of Mental Health (NIMH) (70 male, 45 female)
of comparable age. Informed consent was obtained from parents
or guardians of all subjects who directly participated in the study.
Children provided informed assent. The study was approved by
the McLean Hospital institutional review board (IRB). Collection
of normal control images at NIMH was approved by the IRB at
the National Institutes of Health (NIH).
Mean age (# SD) was 12.9 # 2.9 years for psychiatric subjects
and 11.9 # 3.5 years for control subjects. Inpatient records were
blindly reviewed by two independent raters (an M.D., Ph.D.
pediatric neurologist/psychiatrist and a Ph.D. clinical nurse spe-
cialist licensed to practice in an expanded role in child psychia-
try) using all clinical information and Department of Social
Service investigative reports to ascertain whether the children
had a history of physical abuse, sexual abuse, psychological
abuse (witnessing domestic violence, verbal abuse), or neglect.
Physical, sexual, and psychological abuse were scored as 0, 1, or
2 to indicate no abuse, probable abuse, or documented and
severe abuse, respectively. Neglect was scored as present or
absent. The two clinicians conduct their independent ratings
with the understanding that sexual abuse occurred when minors
were caused to engage in sexual acts through force, threat, or
coercion. Physical abuse was the nonaccidental injury of a minor
that required or should have received medical attention, or
frequent instances of harsh and severe corporal punishment.
Psychological abuse occurred from witnessing domestic violence
or frequent exposure to abusive verbal comments. Neglect was
the chronic failure of a parent or caretaker to provide a minor
with basic needs, such as food, clothing, shelter, medical care,
educational opportunities, protection, and supervision. There
was 88% agreement between raters for sexual abuse, 78%
agreement for physical abuse, 82% agreement for psychological
abuse, and 90% agreement for neglect. All disparate ratings were
presented to a third experienced clinical rater (M.D., Ph.D.
psychiatrist), blind to results, to resolve the discrepancy. These
ratings in each case agreed with one of the two previous ratings
and were used as the designated consensus classification.
On the basis of these assessments there were nine children
(three boys, six girls) with a documented history of sexual abuse,
with a mean age of onset and duration of abuse of 3.6 # 1.4 and
5.1 # 3.8 years, respectively. There were six children (two boys,
four girls) with a history of probable sexual abuse that com-
menced at approximately 8.0 # 3.9 years of age and continued
for 3.3 # 2.9 years. Eighteen children (nine boys, nine girls) had
a history of physical abuse that began at 4.3 # 2.5 years of age
and ended on average 5.6 # 2.9 years later. Twenty children (13
boys, 7 girls) had experienced neglect, with an average age of
onset of 4.5 # 3.4 years and duration of 5.0 # 3.1 years.
There were 28 subjects (13 boys, 15 girls; aged 12.9 # 3.2
years) in the abused/neglected group and 23 subjects in the
psychiatric contrast group (13 boys, 10 girls; aged 12.9 # 2.5
years). Half of the subjects in the abused/neglected group had a
discharge diagnosis of PTSD according to DSM-III-R criteria. As
expected, none of the subjects in the contrast group had a PTSD
diagnosis (Fisher exact test p ! .003). Seventy-one percent of the
abused/neglected group had serious mood disorders accompa-
nied by suicidal ideation or self-destructive behavior. This was
true for 52% of the subjects in the psychiatric contrast group
BIOL PSYCHIATRY 2004;56:80 – 85 81
(Fisher exact test p $ .2). Twenty-six percent of the contrast
group had a primary problem with a disruptive behavioral
disorder, and this was seen as a primary problem for 14% of the
children in the abused/neglected group (Fisher exact test p $ .4).
The mean duration of hospitalization was 90 # 55 days in the
abused/neglected group and 127 # 93 days in the contrast group
[F (1,49) ! 3.09, p ! .09].
Magnetic resonance imaging consisted of T1-weighted scans
obtained from GE 1.5 Tesla Signa scanners (General Electric
Medical Systems, Milwaukee, Wisconsin) at McLean or NIMH to
provide 1.5-mm-thick sagittal slices with three-dimensional
spoiled gradient recalled echo in the steady state (time to echo !
5 msec, repetition time ! 24 msec, flip angle ! 45°, acquisition
matrix ! 192 % 256, field of view ! 24 cm). Stored image files
were used for analysis whenever available. When no stored
image was available, archived films were scanned into a com-
puter workstation with OPTIMAS software (Media Cybernetics,
Silver Spring, Maryland; www.optimas.com). All images were
subsequently analyzed with NIH IMAGE software (NIH, Be-
thesda, Maryland). Anatomical measurements of the CC were
obtained from the midsagittal image, and an automated algo-
rithm divided the CC into seven regions, as defined by Witelson
(1989). Magnetic resonance image measures were performed by
two independent researchers blind to all clinical variables, with
interrater reliability of .83 across all regions. Corpus callosum
measurements from the inpatient subjects had been previously
reported (Teicher et al 1997), though these values had not been
compared with those of control subjects or examined for the
statistical effects of exposure to neglect or development of PTSD.
Differences between groups were calculated with multivariate
analysis of covariance (MANCOVA), with age, midsagittal area,
and scanner as covariates (SPSS software; SPSS, Chicago, Illi-
nois). Multivariate analysis of covariance is similar to analysis of
variance (ANOVA), except that the values are adjusted by the
covariates, and all CC regions were assessed in tandem to
ascertain whether there were group differences for the entire
data set. This technique provides a composite evaluation of the
entire set of measures, thus avoiding the problems that occur
when separate analyses are made of multiple measures that are
correlated and not fully independent. Group differences in each
region were also assessed by ANCOVA, with age and midsagittal
area and scanner as covariates. Stepwise regression analysis
(SPSS) was performed to ascertain the relative influence of a
history of neglect, sexual abuse, or physical abuse, PTSD, age,
and midsagittal area on regional CC size. This is a useful way to
address the analysis, because some children in the sample were
both abused and neglected, and it is difficult to form homoge-
neous and independent groups of adequate sample size. Step-
wise regression first identifies the independent variable that most
strongly correlates with the dependent variable. Other indepen-
dent variables are added to the regression if they significantly
enhance the fit. Several alternative models were compared with
both forward and backward regression. This technique identified
a parsimonious composite of independent variables, in which
each significantly contributed to the overall regression and
provided the best overall fit.
Results
There were no significant differences between the NIMH and
McLean scanners in the assessment of regional CC area in the
healthy control subjects (MANCOVA Wilks F ! 1.27, S ! 1, M !
2.5, n ! 51, p ! .27). Total adjusted CC area (SYSTAT, Evanston,
www.elsevier.com/locate/biopsych
82 BIOL PSYCHIATRY 2004;56:80 – 85 M.H. Teicher et al

Table 1. Comparison Between Abused/Neglected and Nonabused Psychiatric Inpatients Versus Healthy Control Subjects

Area (mm2) p Values

Abused/ Control Overall Group
Region Neglected (A) Contrast (B) Subjects (C) Difference A vs. C B vs. C A vs. B

1 (Rostrum) 18.220 17.014 16.229 .91 .65 .86 .75

2 (Genu) 118.914 132.892 129.476 .22 .27 .81 .15
3 (Rostral Body) 76.036 83.595 88.732 .10 .10 .38 .16
4 (Anterior Midbody) 57.711 67.676 77.847 .00001 .000001 .03 .02
5 (Posterior Midbody) 52.414 62.361 71.132 .00007 .000009 .05 .02
6 (Isthmus) 141.081 156.893 163.774 .10 .04 .62 .17
7 (Splenium) 59.026 70.618 81.426 .01 .003 .19 .08
Total 523.385 590.959 628.547 .005 .0001 .18 .01
n 28 23 115 MANCOVA p ! .002 p ! .00006 p ! .27 p ! .2
MANCOVA, multivariate analysis of covariance.

Illinois) for the healthy control subjects was 614 # 89 mm2 for associated with the changes in relative regional CC size. Neglect
the NIMH scanner and 627 # 95 mm2 for the McLean scanner, a was the most consistent experiential factor and was associated
nonsignificant 2% difference [F (1,110) ! .30, p $ .5]. Age (r ! with a 13% reduction in total adjusted CC area and a 15%–18%
.724, p ! .02) and midsagittal area (r ! .848, p ! .001) were reduction in the size of regions 3, 4, 5, and 7 (all p " .02). Neither
highly significant covariates; gender (r ! &.02, p $ .9) was not physical abuse, sexual abuse, nor a DSM-III-R PTSD discharge
a significant covariate or factor [F (1,109) ! .48, p ! .5] in this diagnosis contributed significantly to overall CC area once the
sample. effects of neglect were taken into account. Physical abuse was
Table 1 summarizes the differences in adjusted regional CC associated with a 12% reduction in region 6. Sexual abuse was
area in hospitalized children with a history of abuse or neglect, associated with a 17% reduction in region 4. The presence or
hospitalized children without any history of abuse or neglect but absence of a PTSD diagnosis did not significantly contribute to
with serious psychiatric illness (contrast group), and normal the regression analysis for any region; however, there might have
healthy control subjects. Overall, there were prominent differ- been too few subjects with a PTSD diagnosis in the sample to
ences between the three groups (MANCOVA Wilks F ! 2.56, S ! adequately evaluate this factor. Age exerted significant effects on
2, M ! 2, n ! 74.5, p ! .002). There was a robust difference total CC size and area of regions 1, 4, 5, 6, and 7. Midsagittal area
between children with a history of abuse or neglect and healthy
control subjects (MANCOVA Wilks F ! 5.50, S ! 1, M ! 2.5, n !
64, p ! .00006). There were no significant multivariate differ-
ences between the contrast group and the healthy control
subjects (Wilks F ! 5.50, S ! 1, M ! 2.5, n ! 61.5, p ! .274) or
between the contrast group and the abused/neglected group
(Wilks F ! 5.50, S ! 1, M ! 2.5, n ! 18.5, p ! .20). Compared
with the healthy control subjects, total adjusted CC area was
reduced by 16.7% in the subjects with a history of abuse or
neglect [F (1,136) ! 15.26, p ! .0001]. Total CC area was 11%
smaller in the abused/neglected group than in the contrast
control subjects [F (1,45) ! 6.74, p ! .01]. Total CC area did not
differ significantly between the contrast group and the healthy
control subjects [6.0% difference; F (1,131) ! 1.85, p ! .18].
Univariate F tests indicated that the most prominent differ-
ences between the abused/neglected group and the healthy
control subjects were in CC regions 4, 5, and 7, which were
reduced by 26%, 26%, and 28% respectively (Table 1). Group
differences in region 4 are illustrated in Figure 1. Region 4 of the
CC of abused/neglected individuals was substantially reduced in
size compared with normal control subjects [F (1,136) ! 26.22, p
" 10&6] and contrast control subjects [15% reduction, F (1,45) !
6.28, p ! .016]. Region 4 of the CC was also smaller in the
contrast group than in the healthy control group [13% reduction,
F (1,131) ! 5.04, p ! .026].
There were no significant differences attributable to gender in
this sample, nor were there significant gender % group interac-
tions (e.g., abused/neglected vs. healthy control subjects: group
% gender interaction Wilks F ! .61, S ! 1, M ! 2.5, n ! 64, p ! Figure 1. Scatter plot of the midsagittal area of the anterior midbody (re-
.743). gion 4) of the corpus callosum in pediatric subjects. Mean values for each
Table 2 summarizes the results of stepwise regression analy- group are indicated by a horizontal line. NIMH, National Institute of Mental
ses performed to ascertain which factors were most directly Health.

www.elsevier.com/locate/biopsych
M.H. Teicher et al BIOL PSYCHIATRY 2004;56:80 – 85 83

Table 2. Influence of Early Experience, Age, Gender, and Midsagittal Area on Regional Area of the Corpus Callosum

% Change in Area
Control
Region Mean (mm2) Neglect Physical Abuse Sexual Abuse PTSD Gendera Ageb Midsagittal areac

1 (Rostrum) 16.229 4.97d

2 (Genu) 129.476 .001e
f
3 (Rostral Body) 88.732 &14.90
4 (Anterior Midbody) 77.847 &17.93g &17.10h 1.09i
5 (Posterior Midbody) 71.132 &17.56j 2.33k
6 (Isthmus) 163.774 &11.94l &9.30i 1.793g
d
7 (Splenium) 81.426 &17.18 2.11m .001j
Total 628.547 &12.91j 1.17g .0003j
PTSD, posttraumatic stress disorder.
a
Percent reduction in girls.
b
Percent change per year.
c
Percent change per mm2.
d
p " .02.
e
p " .00002.
f
p " .005.
g
p " .001.
h
p " .0005.
i
p " .01.
j
p " .0002.
k
p " .000001.
l
p " .05.
m
p " .004.

contributed to the strength of the regression analysis for the total gender and PTSD history. Region 6 was the only segment of the
CC and area of regions 2 and 7. CC in the present study for which the area was directly influ-
Neglect and exposure to sexual abuse were associated with enced by gender. The De Bellis et al studies (De Bellis and
equivalent reductions in the area of region 4. Both factors might Keshavan 2003; De Bellis et al 1999) are quite important but are
be exerting comparable effects on all exposed subjects, or there
might be subsets more vulnerable to the effect of one adverse
experience or the other. Boys and girls were compared for
possible differential responses to neglect or sexual abuse on
regional CC area. Neglect was associated with a very strong effect
on CC area 4 in boys [F (1,91) ! 16.07, 'p2 ! .150, p ! .0001],
whereas sexual abuse was not [F (1,91) ! .43, 'p2 ! .005, p ! .5].
In contrast, sexual abuse was associated with a significant
reduction in CC area 4 in girls [F (1,65) ! 10.37, 'p2 ! .138, p !
.002], but neglect was not [F (1,65) ! .67, 'p2 ! .01, p ! .4].
Figure 2 illustrates the differential effect sizes (with 95% confi-
dence intervals) for the effects of neglect or sexual abuse on CC
areas in boys and girls. Neglect was associated with substantially
greater effect sizes than was sexual abuse in boys, for the CC as
a whole and particularly for regions 4, 5, and 7. Sexual abuse
tended to be associated with larger effect sizes than neglect in
girls, most noticeably in regions 3 and 4.

Discussion
In the present study, early traumatic experience rather than
psychiatric illness was associated with decreased CC size. When
factors that exerted the greatest effect on the size of the CC were
analyzed, neglect had the greatest influence in boys relative to
physical abuse, sexual abuse, or PTSD. Neglect in boys was
associated with a marked decrease in the size of regions 3, 4, 5,
and 7. Neglect was associated with far less effect in girls and was
limited to the most posterior aspects of the CC. In contrast, sexual
abuse exerted significant effects on regions 3– 6 of girls but only
exerted a trend-level effect on region 7 in boys.
De Bellis et al (1999) also found that the CC was more affected
by early maltreatment in male subjects than in female subjects. Figure 2. Comparative effect of sexual abuse and neglect on the regional
More recently, De Bellis and Keshaven (2003) reported that CC size of the corpus callosum (CC). Effect sizes (Cohen’s d) are shown # 95%
region 6 showed the strongest interactive influence between confidence intervals.

www.elsevier.com/locate/biopsych
84 BIOL PSYCHIATRY 2004;56:80 – 85
complicated by the fact that their comparisons were only be-
tween a PTSD sample with a high rate of comorbid psychiatric
illness (38 of 44 patients) and healthy normal control subjects.
Were CC differences between the abused/PTSD/psychiatrically
ill subjects and healthy control subjects a consequence of abuse,
PTSD, or severe psychiatric illness? Was reduced CC size a
pre-existing risk factor for the development of PTSD (Gilbertson
et al 2002)? The present study suggests that differences in CC size
might not be a consequence of psychiatric illness per se and
seems to suggest that PTSD as assessed was neither necessary
nor sufficient to account for these effects.
There are, however, a number of limitations in the present
study. First, this was a retrospective study, and information was
gleaned from the extensive inpatient records of the abused and
contrast groups rather than from research-specific structured
interviews. The subjects had been hospitalized for an average of
107 # 76 days, so they had been carefully assessed, evaluated,
and observed by well-trained clinicians; however, we had no
control over the consistency and rigor of the assessment from
clinician to clinician and the degree to which past histories of
abuse or neglect were documented. The second major limitation
was the use of a large group of healthy control subjects who
were imaged at the NIMH on a different scanner. Though the
scanners were identical models, they might provide nonidentical
results. No significant differences, however, were observed in CC
measures between healthy control subjects scanned at McLean
and healthy control subjects scanned at NIMH. Further, CC
differences reported between subjects with a history of abuse/
neglect and the combined NIMH/McLean control group were
also apparent if only the McLean control group was used. For
example, total CC area in the abused/neglected group imaged at
McLean was 19% smaller than total CC area observed in healthy
control subjects scanned at McLean [F (1,43) ! 17.56, p ! .0001].
Although scanner differences can be a confounding factor, the
large size and quality of the NIMH control group enhanced our
confidence in the observations. The third limitation was the
relatively small size of the abused/neglected group. Although the
size of this group was adequate to readily reveal differences from
control subjects, there was insufficient statistical power to detect
small potential differences between abused subjects with or
without PTSD. There was also insufficient statistical power to
ascertain whether there were windows of vulnerability in which
abuse or neglect at specific ages would be associated with
gender-dependent differences in regional CC area.
Research has shown that myelination in some regions of the
human brain occurs with a gender-dependent time course that
might provide windows of susceptibility to early stress. Aboitiz et
al (1996) have reported prominent gender differences in fiber
development of the CC in adults. He found that the number of
relatively abundant axons with diameters between 1 and 3 (m
correlated with age only in female subjects, whereas the number
of scarce fibers more than 3 (m in diameter correlated signifi-
cantly with age only in male subjects. Benes et al (1994) found
that myelin ratios in the hippocampus were equal in boys and
girls aged 0 to 5 years, but they abruptly diverge, with girls
having a 41% greater myelin ratio from ages 6 to 11 and a 33%
greater ratio from ages 12 to 19. Hence, hippocampal myelina-
tion occurs more rapidly in girls from age 6 years onward,
possibly enhancing their vulnerability to stress during the period
when childhood sexual abuse is likely to occur. In contrast,
neglect often occurs much earlier in life (U.S. Department of
Health and Human Services 1999), at a time when boys seem to
require more stimulation than girls for normal development. This
www.elsevier.com/locate/biopsych
M.H. Teicher et al
increased need for stimulation might render them more vulner-
able to the effect of neglect, and studies examining neglect
sequelae almost invariably report greater effects in male subjects
if they examine gender differences (e.g., Kaufman and Cicchetti
1989; Mak 1994; Rogeness et al 1986).
Results from studies with animals also suggest a gender-
dependent difference in the effects of the environment on CC
size. Berrebi et al (1988) showed that male and female rats
differed in the effects of early handling (now generally recog-
nized as a form of brief beneficial stimulation [Liu et al 2000]) on
the development of their CC. In adulthood (110 days), handled
male rats had significantly greater width of their CC than
nonhandled male control animals, with greatest differences
emerging in the middle portions. In contrast, the CC of handled
female animals was reduced in width relative to nonhandled
control animals. Juraska and Kopcik (1988) found that rearing in
a complex environment (postweaning) enhanced the size of the
CC of both male and female rats, though the magnitude of the
effect varied by gender. Moreover, ultrastructural studies showed
that the effects of early experience were mediated through
different mechanisms. Rearing in the complex environment
produced a significant increase in the number of myelinated
axons in the CC of female but not male rats (Juraska and Kopcik
1988). In contrast, rearing in the complex environment markedly
increased the diameter of myelinated axons in the CC of male rats
but had no effect on myelinated axon diameter in females. These
findings lend credence to the hypothesis that early experience
can affect the development of the human CC, and that boys and
girls might differ to a significant degree in the manner in which
the CC is affected.
These clinical studies indicate a significant association be-
tween neglect, abuse, and size of the CC. They do not, however,
prove that there is a cause– effect relationship. Animal studies
bridge this gap to a large degree because subjects are randomly
assigned to groups. This eliminates the alternative hypotheses
that the brain abnormalities increase the risk of being abused/
neglected or that the brain abnormalities are inherited and that
similar abnormalities in one’s relatives increases risk of abuse or
neglect. Hence, the findings by Sanchez et al (1998) that differ-
ential rearing experience affected the development of the CC and
the cognitive function of male rhesus monkeys helps support the
present hypotheses. Briefly, infant monkeys raised individually
in a nursery from 2 to 12 months were compared with age-
matched infants raised in a semi-naturalistic social environment.
Although overall brain volumes did not differ, the CC was
significantly decreased in the nursery group. Rearing differences
were not found in the hippocampus, cerebellum, or anterior
commissure, consistent with the clinical findings of De Bellis et al
(1999). Cognitive difficulties emerged in delayed nonmatching to
sample and object reversal learning, and degree of impairment
correlated with alterations in CC size.
Neglect might qualitatively differ from abuse, because it is
often associated with lack of appropriate stimulation and inter-
action. One possible implication of the above findings is that
male subjects might be more dependent than female subjects on
receipt of adequate early stimulation to foster full development
of their CC. The greater association between sexual abuse and
alterations in the CC of girls might suggest that the female CC has
a later period of vulnerability/sensitivity, given that sexual abuse
often occurs at a later age than neglect. Neglect is more prevalent
than child abuse (U.S. Department of Health and Human Services
1999) and might exert more dramatic effects on brain develop-
ment than previously hypothesized. To further understand the
M.H. Teicher et al
role of neglect on brain maturation, more attention needs to be
devoted to the biopsychosocial consequences of neglect as an
entity separate and distinct from other forms of maltreatment.
This research was funded by RO1 MH-53636 from the Na-
tional Institute of Mental Health (NIMH) to MHT, with additional
support from NIMH Grant RO1 MH-66222 and RO1DA-016934
from the National Institute of Drug Abuse. This work was
presented previously at the annual meeting of the Society for
Neuroscience, November 4 –9, 2000, New Orleans, Louisiana.
We thank Carol A. Glod R.N., C.S., Ph.D. for her evaluation of
the clinical records.
Aboitiz F, Rodriguez E, Olivares R, Zaidel E (1996): Age-related changes in
fibre composition of the human corpus callosum: Sex differences. Neu-
roreport 7:1761–1764.
Benes FM, Turtle M, Khan Y, Farol P (1994): Myelination of a key relay zone in
the hippocampal formation occurs in the human brain during child-
hood, adolescence, and adulthood. Arch Gen Psychiatry 51:477–484.
Berrebi AS, Fitch RH, Ralphe DL, Denenberg JO, Friedrich VL Jr, Denenberg
VH (1988): Corpus callosum: Region-specific effects of sex, early experi-
ence and age. Brain Res 438:216 –224.
Bremner JD, Randall P, Vermetten E, Staib L, Bronen RA, Mazure C, et al
(1997): Magnetic resonance imaging-based measurement of hippocam-
pal volume in posttraumatic stress disorder related to childhood physi-
cal and sexual abuse—a preliminary report. Biol Psychiatry 41:23–32.
De Bellis MD, Keshavan MS (2003): Sex differences in brain maturation in
maltreatment-related pediatric posttraumatic stress disorder. Neurosci
Biobehav Rev 27:103–117.
De Bellis MD, Keshavan MS, Clark DB, Casey BJ, Giedd JN, Boring AM, et al
(1999): Developmental traumatology. Part II: Brain development. Biol
Psychiatry 45:1271–1284.
Denenberg VH, Yutzey DA (1985): Hemispheric laterality, behavioral asym-
metry, and the effects of early experience in rats. In: Glick SD, editor.
Cerebral Lateralization in Nonhuman Species. Orlando, Florida: Academic
Press, 109 –133.
Giedd JN, Snell JW, Lange N, Rajapakse JC, Casey BJ, Kozuch PL, et al (1996):
Quantitative magnetic resonance imaging of human brain develop-
ment: Ages 4 –18. Cereb Cortex 6:551–560.
Gilbertson MW, Shenton ME, Ciszewski A, Kasai K, Lasko NB, Orr SP, Pitman
RK (2002): Smaller hippocampal volume predicts pathologic vulnerabil-
ity to psychological trauma. Nat Neurosci 5:1242–1247.
Ito Y, Teicher MH, Glod CA, Ackerman E (1998): Preliminary evidence for
aberrant cortical development in abused children: A quantitative EEG
study. J Neuropsychiatry Clin Neurosci 10:298 –307.
BIOL PSYCHIATRY 2004;56:80 – 85 85
Ito Y, Teicher MH, Glod CA, Harper D, Magnus E, Gelbard HA (1993): In-
creased prevalence of electrophysiological abnormalities in children
with psychological, physical, and sexual abuse. J Neuropsychiatry Clin
Neurosci 5:401–408.
Juraska JM, Kopcik JR (1988): Sex and environmental influences on the size
and ultrastructure of the rat corpus callosum. Brain Res 450:1–8.
Kaufman J, Cicchetti D (1989): Effects of maltreatment on school-age chil-
dren’s socioemotional development: Assessments in a day-camp set-
ting. Dev Psychol 25:516 –524.
Liu D, Caldji C, Sharma S, Plotsky PM, Meaney MJ (2000): Influence of neona-
tal rearing conditions on stress-induced adrenocorticotropin responses
and norepinephrine release in the hypothalamic paraventricular nu-
cleus. J Neuroendocrinol 12:5–12.
Mak AS (1994): Parental neglect and overprotection as risk factors in delin-
quency. Special issue: Research in psychophysiology. Aust J Psychol
46:107–111.
Rogeness GA, Hernandez JM, Macedo CA, Amrung SA, Hoppe SK (1986):
Near-zero plasma dopamine-B-hydroxylase and conduct disorder in
emotionally disturbed boys. J Am Acad Child Psychiatry 25:521–527.
Sanchez, MM Hearn EF, Do-D, Rilling JK, Herndon JG (1998): Differential
rearing affects corpus callosum size and cognitive function of rhesus
monkeys. Brain Res 812:38 – 49.
Schiffer F, Teicher MH, Papanicolaou AC (1995): Evoked potential evidence
for right brain activity during the recall of traumatic memories. J Neuro-
psychiatry Clin Neurosci 7:169 –175.
Stein MB (1997): Hippocampal volume in women victimized by childhood
sexual abuse. Psychol Med 27:951–959.
Teicher MH (1994): Early abuse, limbic system dysfunction, and borderline
personality disorder. In: Silk K, editor. Biological and Neurobehavioral
Studies of Borderline Personality Disorder. Washington, DC: American Psy-
chiatric Press, 177–207.
Teicher MH, Andersen SL, Polcari A, Anderson CM, Navalta CP (2002): Devel-
opmental neurobiology of childhood stress and trauma. Psychiatr Clin
North Am 25:397–426.
Teicher MH, Ito Y, Glod CA, Andersen SL, Dumont N, Ackerman E (1997):
Preliminary evidence for abnormal cortical development in physically
and sexually abused children using EEG coherence and MRI. Ann N Y
Acad Sci 821:160 –175.
U. S. Department of Health and Human Services, Administration for Children
and Families (1999): Child Maltreatment 1997: Reports from the States to
the National Child Abuse and Neglect Data System. Washington, DC:
U. S. Government Printing Office.
Widom CS (1999): Posttraumatic stress disorder in abused and neglected
children grown up. Am J Psychiatry 156:1223–1229.
Witelson SF (1989): Hand and sex differences in the isthmus and genu of the
human corpus callosum. A postmortem morphological study. Brain
112:799 –835.
www.elsevier.com/locate/biopsych