If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Childhood maltreatment affects approximately 25% of the world’s population. Importantly, the children of mothers who have been maltreated are at increased risk of behavioral problems. Thus, one important priority is to identify child neurobiological processes associated with maternal childhood maltreatment (MCM) that might contribute to such intergenerational transmission. This study assessed the impact of MCM on infant gray and white matter volumes and infant amygdala and hippocampal volumes during the first 2 years of life.
Fifty-seven mothers with 4-month-old infants were assessed for MCM, using both the brief Adverse Childhood Experiences screening questionnaire and the more detailed Maltreatment and Abuse Chronology of Exposure scale. A total of 58% had experienced childhood maltreatment. Between 4 and 24 months (age in months: mean = 12.28, SD = 5.99), under natural sleep, infants completed a magnetic resonance imaging scan using a 3T Siemens scanner. Total brain volume, gray matter volume, white matter volume, and amygdala and hippocampal volumes were extracted via automated segmentation.
MCM on the Adverse Childhood Experiences and Maltreatment and Abuse Chronology of Exposure scales were associated with lower infant total brain volume and gray matter volume, with no moderation by infant age. However, infant age moderated the association between MCM and right amygdala volume, such that MCM was associated with lower volume at older ages.
MCM is associated with alterations in infant brain volumes, calling for further identification of the prenatal and postnatal mechanisms contributing to such intergenerational transmission. Furthermore, the brief Adverse Childhood Experiences questionnaire predicted these alterations, suggesting the potential utility of early screening for infant risk.
). However, little work has examined whether maternal CM (MCM) affects brain development of the affected mother’s offspring. To date, two studies have examined whether MCM is associated with offspring brain volume and function. Most relevant to this study, in a sample of newborns, Moog et al. (
) found that MCM was associated with lower infant intracranial volume and reduced gray matter volume (GMV) but not with differences in white matter volume (WMV), cerebrospinal fluid volume, amygdala volume, or hippocampal volume. Thus, MCM may have an effect on infant brain volume that is evident close to birth, prior to the impact of postnatal risks. In addition, Hendrix et al. (
) found that maternal childhood emotional neglect was associated with stronger infant functional connectivity between the amygdala and the dorsal anterior cingulate cortex and ventromedial prefrontal cortex, connectivity that has previously been associated with fear conditioning. This study extends the work by Moog et al. (
), other studies have demonstrated that additional psychosocial risk factors are associated with child GMV and WMV. Lower socioeconomic status (SES) has been linked to lower GMV but not with lower WMV (
). Although most studies have focused on older children, a few studies have examined associations between adversity during the first 2 years of life (institutional rearing, maltreatment, maternal depression) and amygdala and hippocampal volumes later in childhood (
) used the Childhood Trauma Questionnaire to assess MCM in relation to infant brain volume. However, detailed assessments such as the Childhood Trauma Questionnaire can be time intensive. Given the public health priority for early identification of infants with altered brain development, in this study, maternal CM experiences were assessed using both the brief Adverse Childhood Experiences (ACE) questionnaire (
). Use of both measures allowed an evaluation of whether a brief screening measure is sensitive enough to identify associations between MCM and infant brain volume.
Study aims were as follows: 1) assess associations between MCM and indices of brain volume during the first 2 years of life, including total brain volume (TBV), GMV, and WMV (for cerebrospinal fluid volume findings, see the Supplement); 2) examine associations between MCM and infant amygdala and hippocampal volumes; 3) evaluate whether these associations were moderated by infant age at scan during the first 2 years of life; and 4) assess how the brief ACE assessment compares to the more detailed MACE in relation to aims 1, 2, and 3.
) regarding MCM and limbic volumes in newborns, we hypothesized that associations between MCM and amygdala and hippocampal volumes would be moderated by age, with effects emerging later in the first year. Given previous conflicting findings, specific hypotheses were not offered regarding whether MCM would be associated with increased or decreased amygdala and hippocampal volumes.
Methods and Materials
Fifty-seven mother-infant dyads were included in the magnetic resonance imaging (MRI) analyses. Dyads were drawn from a cohort of 181 families enrolled in the Harvard MIND (Mother-Infant Neurobiological Development) study. Mothers in the MIND study were recruited through prenatal classes, community flyers, and local birth records. Participants were screened and stratified such that at least 50% of mothers had experienced one or more forms of CM (physical, sexual, emotional abuse; witnessed domestic violence; physical, emotional neglect). Exclusion criteria were 1) English not a primary language spoken at home, 2) maternal age >44 years at infant birth, 3) infant <36-week gestation or <2500 g at birth, and 4) infant congenital defect or disorder. Dyads participated in behavioral assessments at infant ages 4 and 15 months.
All MIND study participants were offered participation in the infant MRI assessment. Of these, 17.12% (n = 31) declined and 17.12% (n = 31) could not be scheduled or withdrew from the MRI prior to scanning. Of the consenting and available participants, 52.10% (n = 62) attempted to complete the MRI but were unsuccessful (see the Supplement). This resulted in 57 infants successfully completing the MRI. Infants with MRIs were similar to infants without MRIs on demographic characteristics and key study variables (see the Supplement). The study was approved by the institutional review board (Partners Healthcare IRB Protocol #: 2014P002522).
Infants participated in MRIs between 4 and 25 months of age (age in days: mean = 357.93, SD = 185.98; 49% male). Table 1 shows sociodemographic characteristics of the MRI sample. On study entry, the first half of participants were offered MRI scans after the 15-month assessment, while the second half of participants were offered scans after the 4-month assessment. Fifteen-month scans were offered to earlier participants owing to the long lead time needed from recruitment to the post–15-month scan. Thus, as seen in Figure 1, ages at scans clustered after the 4-month assessment and after the 15-month assessment. As shown in Table 2, age at scan was not correlated with sociodemographic variables. However, there was a moderate correlation between older ages at scan and lower ACE (r = −0.35, p < .01) and MACE (r = −0.28, p < .05) severity scores, reflecting the difficulty of obtaining successful scans from at-risk toddlers. Thus, age at scan was covaried in all analyses that included the ACE or MACE.
Table 1Sociodemographic Characteristics and Descriptive Statistics
n = 57, except for regions with outliers that were removed, resulting in total brain volume n = 56; RH/LH amygdala volumes n = 56; RH/LH hippocampal volumes n = 55. Infant age is measured in days; gestational age at birth is measured in weeks; infant sex (0 = male; 1 = female); maternal education is measured in years of education.
EPDS, Edinburgh Postnatal Depression Scale; LH, left hemisphere; MRI, magnetic resonance imaging; RH, right hemisphere.
) during an initial screening call. The ACE screens for the presence of emotional, physical, and sexual abuse; emotional and physical neglect; and witnessing domestic violence (DV) prior to age 18. Although DV was originally conceptualized as a household dysfunction on the ACE scale (
). Thus, for consistency across MCM measures, DV was included in the ACE severity score. The ACE severity score ranged from 0 to 6, reflecting how many of the six maltreatment types were experienced. The ACE has validity in relation to maternal health and infant development (
). The MACE assesses the severity of seven types of parental maltreatment, including verbal abuse, nonverbal emotional abuse, physical abuse, sexual abuse, emotional neglect, physical neglect, and witnessing parental DV, as well as three less often assessed forms of nonparental maltreatment including witnessing violence to siblings, peer emotional abuse, and peer physical bullying. The overall severity score used here summed the severity of the seven types of parental maltreatment and thus had a possible range from 0 to 70. Note that these types parallel the types of CM assessed on the ACE questionnaire. MACE scores correlate highly with other measures of CM and have high test-retest reliability (
Sociodemographic characteristics were assessed by maternal interview, including infant sex, infant gestational age at birth (in weeks), family income, and maternal education. Infant non-White race and infant Hispanic ethnicity were also reported. However, only 3 infants were identified as Hispanic, and most non-White participants were of mixed race (Table 1), so classification by race had unclear meaning. Therefore, race and ethnicity were not considered further.
Maternal depressive symptoms have previously been associated with infant brain volume (
All infant MRIs were performed on a 3T Siemens Skyra scanner (Siemens Healthineers) with a 64-channel head coil. Infants were scanned during natural sleep, and no sedation was used. The T1-weighted acquisition used an advanced version of the magnetization prepared rapid acquisition gradient-echo sequence, where fast, low-resolution volumetric navigators were played for each repetition period and were used for prospective motion correction (
). The specific imaging parameters of the magnetization prepared rapid acquisition gradient-echo acquisition included voxel size = 1 × 1 × 1 mm3, repetition time = 2500 to 2540 ms, echo time = 1.65 to 2.37 ms, inversion time = 1450 to 1470 ms, field of view = 192 × 192 mm2 and between 144 and 173 slices, enough to cover the entire brain of the infant. After performing visual quality control using the Freeview software (surfer.nmr.mgh.harvard.edu), the T1-weighted volumes were manually aligned along the anterior commissure–posterior commissure plane and underwent N4 bias correction (
). Quality control of the segmentations was visually performed using the FSLView software (https://fsl.fmrib.ox.ac.uk/). These segmentations enabled the extraction of TBV, GMV, WMV, cerebrospinal fluid volume (see the Supplement), and right and left hemisphere amygdala and hippocampal volumes (Figure 2).
Statistical analyses were conducted using IBM SPSS Statistics (version 26) and Mplus (version 8; Muthén & Muthén). Linear regressions were conducted using maximum likelihood estimation with robust standard errors and full information maximum likelihood to account for missing data. Outliers were identified using boxplots in SPSS. Extreme outliers (e.g., third quartile + 3 × interquartile range) for brain volumes were removed and estimated using full information maximum likelihood to retain participants. Regression analyses were used to assess MCM associations with infant brain volumes, controlling for relevant covariates. Models were run assessing MCM and age main effects and interactions.
Given that Mplus will not run models when the variances of variables are too large, all brain volumes were divided by 1000, and age (in days) was divided by 100 to reduce variance of the age by MCM interactions.
Analyses were conducted using both the MACE and ACE scores to assess whether the ACE demonstrated similar strength in capturing the associations between MCM and infant brain volumes. Our rationale for not correcting for multiple comparisons is included in the Supplement.
Descriptives and Covariates
There was one outlier for TBV, one for amygdala volume, and two for hippocampal volume (3 participants total). Outliers were removed and estimated using full information maximum likelihood in regression analyses. Table 1 displays descriptive statistics for all study variables, excluding outliers.
Family income, maternal education, maternal depressive symptoms, infant sex, gestational age at birth, and age at scan were assessed as potential covariates in relation to infant brain volumes of interest (Table 2). Infant age at MRI was positively correlated with GMV (Figure 1) and TBV (Figure S1). Infant sex was not a significant covariate in relation to infant brain volumes when all ages were combined. However, infant female sex was associated with smaller GMV (r = −0.46, p <.05) and hippocampal volumes (left hemisphere r = −0.48, p < .05; right hemisphere r = −0.44, p < .05) in infants older than 12 months. Given that sex is routinely covaried in brain volume studies (
), we included it as a covariate in all analyses. In addition, higher maternal depressive symptoms were significantly correlated with larger infant left hemisphere hippocampal volume (r = 0.28, p < .05). No other potential covariates were associated with infant brain volumes (Table 2). Most of the sample (68.4%) fell in the EPDS category of low probability of depression, which might account for the nonsignificant associations between the EPDS and other brain volumes. As is common in MRI analyses, we controlled for infant age at MRI and infant sex across all models. Analyses of amygdala and hippocampal volumes also controlled for total GMV. Lastly, given the significant bivariate association between depression and left hemisphere hippocampal volume, depression was included as a covariate in all left hemisphere hippocampal analyses.
MCM and Overall Infant Brain Volumes
Greater MACE severity was associated with lower TBV (β = −0.179, SE = 0.067, 95% CI −0.310 to −0.048) and GMV (β = −0.194, SE = 0.093, 95% CI −0.377 to −0.011). MACE severity was not associated with WMV (β = −0.120, SE = 0.147, 95% CI −0.410 to 0.169). The interaction of MACE severity with age was not significant for TBV (β = −0.127, SE = 0.136, 95% CI −0.394 to 0.140) or GMV (β = −0.086, SE = 0.194, 95% CI −0.466 to 0.294). However, the interaction trended toward significance for WMV (β = −0.518, SE = 0.264, p = .050, 95% CI −1.036 to −0.001).
ACE severity was associated with lower TBV (β = −0.197, SE = 0.069, 95% CI −0.332 to −0.062) and GMV (β = −0.229, SE = 0.064, 95% CI −0.388 to −0.070) but not WMV (β = −0.100, SE = 0.149, 95% CI −0.391 to 0.191). There were no significant interactions between ACE severity and age at scan on TBV (β = −0.150, SE = 0.136, 95% CI −0.417 to 0.117), GMV (β = −0.112, SE = 0.168, 95% CI −0.441 to 0.216), or WMV (β = −0.432, SE = 0.269, 95% CI −0.960 to 0.096). See Figure 3 for distribution of infant GMV by ACE and MACE severity.
MCM and Infant Amygdala and Hippocampal Volumes
In contrast to GMV, the association between MACE severity and right amygdala volume was moderated by age (β = −0.611, SE = 0.184, 95% CI −0.973 to −0.250).
In models excluding the significant interaction terms, the main effects of MACE severity (β = 0.074, SE = 0.191, CI −0.601 to 0.297) and ACE severity (β = 0.122, SE = 0.198, CI −0.068 to 0.128) were not significant.
Greater MACE severity was associated with decreased right amygdala volume as age increased. Simple slope analyses indicated that the negative association between MACE severity and right hemisphere amygdala became significant at older infant ages (β = −0.419, SE = 0.213, 95% CI −0.837 to −0.002) (Figure 4) (
Results for ACE severity were similar, with age moderating the association between ACE MCM severity and right amygdala volume, such that greater ACE severity was associated with more pronounced decreased volume for older infants (β = −0.554, SE = 0.176, 95% CI −0.899 to −0.209).
There were no main effects of MACE or ACE severity on left amygdala volume, and there were no significant interactions between either MACE or ACE severity and age (MACE main effects: β = 0.023, SE = 0.179, 95% CI −0.328 to 0.373; MACE × age interaction: β = −0.291, SE = 0.201, 95% CI −0.685 to 0.102; ACE main effects: β = 0.047, SE = 0.195, 95% CI −0.334 to 0.428; ACE × age interaction: β = −0.247, SE = 0.212, 95% CI −0.662 to 0.167).
Neither the main effect of MACE severity nor the interaction between MACE and infant age was significant for right hippocampal volume (MACE: β = 0.112, SE = 0.177, 95% CI −0.235 to 0.459; MACE × age interaction: β = −0.206, SE = 0.179, 95% CI −0.556 to 0.145). Results were similar for the ACE. Neither the main effect of ACE severity (β = 0.092, SE = 0.198, 95% CI −0.296 to 0.479) nor the interaction between infant age and ACE severity was significant for the right hippocampal volume (β = −0.305, SE = 0.193, 95% CI −0.683 to 0.074).
Similarly, neither MACE nor ACE severity was associated with left hippocampal volume, nor were the interactions between MACE/ACE severity and age significant (MACE main effect: β = 0.010, SE = 0.158, 95% CI −0.301 to 0.320; MACE × age interaction: β = 0.059, SE = 0.238, 95% CI −0.409 to 0.526; ACE main effect: β = −0.018, SE = 0.150, 95% CI −0.312 to 0.275; ACE × age interaction: β = 0.046, SE = 0.222, 95% CI −0.389 to 0.480).
This study provides evidence that maternal history of CM is linked to infant brain volumes beyond the immediate postnatal period. Findings indicated that during the first 2 years of life, greater severity of MCM was associated with 1) lower infant TBV and GMV, 2) marginally lower WMV, and 3) lower infant right amygdala volume. These results emerged with similar magnitude when using either the ACE or MACE instruments.
To our knowledge, this is the first report to link MCM to lower infant TBV and GMV beyond the neonatal period. One previous study (
) found that MCM was associated with lower TBV and GMV in newborns. Similarly, our finding that MCM was associated with lower TBV seemed to be driven primarily by GMV, not WMV, effects. Notably, the relation between MCM and infant GMV was not moderated by infant age, suggesting that this association is consistent during the first 2 years of development.
Studies of typical trajectories of brain development find that overall brain volume undergoes rapid development in the latter part of gestation and early postnatal period (
). After controlling for infant age at MRI, the current findings suggest that infants of mothers with more severe MCM exhibit less GMV increase during the first 2 years of life.
The association between MCM severity and infant GMV during the first 2 years is consistent with both prenatal and postnatal effects on GMV. The majority of individual variation in GMV has occurred by 1 year of age (
). The moderate to high SES of this sample may contribute to the lack of association between SES and infant brain volume. Taken together, these prenatal and postnatal factors should be explored in future work to assess whether they act as mediating or moderating mechanisms linking MCM and infant brain volumes.
These data also revealed a trend suggesting that at older infant ages, greater MCM is associated with marginally decreased WMV. Previous research has not found direct links between MCM (
) and WMV in infants, possibly because of the long developmental course of myelination. Future research, with larger samples, is needed to confirm a possible association between MCM and WMV at older ages.
The current data exhibited an increasing negative association between MCM severity and right amygdala volume during the first 2 years of life, which became significant at approximately 18 months of age (Figure 4). The study of newborns by Moog et al. (
) did not find a significant association between MCM severity and amygdala volume shortly after birth. Thus, associations between MCM and the amygdala may not be evident at birth but may become detectable during the period of rapid amygdala growth that occurs over the early years of life.
Given the more pronounced association between MCM and the amygdala in older infants, postnatal environmental factors associated with MCM may be particularly relevant to development of the right amygdala. Among 9- to 15-year-old children, reduced amygdala volume was related to early-life stress, including abuse, neglect, and sociodemographic risk (
). In addition, in early adolescence, fewer maternal positive behaviors were associated with smaller right amygdala volumes followed by accelerated growth in the right amygdala from early to mid adolescence (
). Future work should examine both prenatal and postnatal factors that might mediate the link between MCM and right amygdala volume.
It remains unclear why MCM findings were lateralized to the right amygdala. However, the right-sided effect found here converges with previous research in adults showing that CM and posttraumatic stress disorder are more consistently associated with right-sided differences in adult amygdala volume (
). Continued work is needed to identify environmental or epigenetic factors that differentially affect the right versus left amygdala.
Finally, the association of MCM and lower GMV using the more complex MACE assessment was similar when using the six yes/no maltreatment questions from the brief ACE questionnaire. These ACE results have important public health implications because the ACE scale is more easily administered in clinical settings. Replications of these findings would support the screening of pregnant women and new mothers for CM histories as part of routine care to provide early support for maltreated mothers and their infants.
Study limitations are also important to note. First, this study included a relatively small sample of infants assessed between 4 and 24 month of age. Because the study was cross-sectional, no conclusions can be drawn regarding changes in brain volume over time. Future work should examine longitudinal trajectories of infant brain growth related to MCM. In addition, owing to the small sample, we were not able to explore differential correlates of maternal childhood abuse versus neglect. Future work with larger samples should assess effects related to the type and timing of MCM on infant brain volumes. This sample also demonstrated relatively high income and education, which might explain why SES was not associated with infant brain volumes. Measures of MCM also rely on self-report and may be subject to biased recall, although similar measures have shown good validity in relation to child protective service reports (
). Finally, there was a modest correlation between age at scan and lower severity of MCM in this study. Age at scan was controlled in all analyses, however, and the tendency for older infants to be at lower risk would be conservative to the finding that MCM was related to lower amygdala volume at older ages. Nevertheless, further work in larger, more diverse samples is needed to corroborate and generalize study findings.
MCM is associated with decreased infant TBV and GMV during the first 2 years of life and with decreased volume in the right amygdala among infants 18 months and older. Findings using the comprehensive MACE instrument (
). Thus, using the ACE questionnaire in pregnancy may allow for early identification and intervention for maltreated mothers and their infants at risk for altered neurobiological development. Further research is needed to identify the contributing prenatal and postnatal biological and behavioral mechanisms that might account for the influence of MCM on infant TBV, GMV, and right amygdala volume so that appropriate interventions may be developed.
Acknowledgments and Disclosures
This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (Grant No. R01HD079484 [to KL-R, PEG, MBE, and MHT]).
We would like to thank the families whose interest and participation made this work possible and also express our appreciation for the hard work and dedication of the study staff.
The authors report no biomedical financial interests or potential conflicts of interest.