Babies born to mothers who experience domestic violence during pregnancy have altered brain development and changes in brain structure. In females, maternal exposure to IPV was associated with a smaller amygdala, a brain area associated with social and emotional development.
In males, the caudate nucleus size was increased. This brain area is associated with multiple functions including memory, learning, reward, and movement. The findings may explain why children of mothers who experience domestic abuse are more likely to suffer from mental health problems later in life.
Original Research: Open access.
“Antenatal maternal intimate partner violence exposure is associated with sex-specific alterations in brain structure among young infants: Evidence from a South African birth cohort” by Lucy V. Hiscox et al. Developmental Cognitive Neuroscience
This prospective study of South African infants (aged 2–6 weeks) was a novel examination of subcortical brain volumes and white matter microstructure whose mothers had been exposed to intimate partner violence (IPV) during pregnancy. We found tentative evidence for sex-specific effects of maternal antenatal IPV exposure on young infant subcortical volumes: IPV predicted a smaller amygdala among females but not males, and a larger caudate nucleus in males but not females.
Structure and Function
The amygdala is an almond-shaped structure that lies in the temporal lobe, lying just beneath the uncus. The amygdala is diverse and complex in structure and comprises approximately 13 nuclei. They further subdivide into extensive internuclear and intranuclear connections. These nuclei functionally sort into five major groups: basolateral nuclei, cortical-like nuclei, central nuclei, other amygdaloid nuclei, and extended amygdala.
Amygdala is one of the components of the limbic system, which is responsible for the control of emotions and behavior besides memory formation. Anatomically, the amygdala lies at the anterior border of the hippocampal formation and the anterior aspect of the lateral ventricle’s inferior horn where it merges with the peri-amygdaloid cortex, which forms part of the surface of the uncus.
Amygdala manages the processing of information between prefrontal-temporal association cortices and the hypothalamus. Amygdala has neural circuits to carry out its different functions with two major output pathways; the Dorsal route via stria terminalis that projects to the septal area and hypothalamus, and the ventral route via the ventral amygdalofugal pathway which terminates in the septal area, hypothalamus, and the medial dorsal thalamic nucleus.
The amygdala also has connections with the basal ganglia circuit via its projections to the ventral pallidum and ventral striatum; these projections are relayed back to the cortex via the dorsomedial nucleus of the thalamus. The basolateral circuit includes the amygdala (especially the basolateral amygdala), the orbitofrontal and anterior temporal cortex, and in the thalamus, the magnocellular division of the dorsomedial nucleus (frontothalamic pathway), which serves as a relay back to the orbitofrontal cortex. The circuit has been proposed as a substrate for the human ability to infer the intentions of others from their language, gaze, and gestures (Theory of mind and social cognition), and helps with social interactions.
The amygdala also functions in regulating anxiety, aggression, fear conditioning, emotional memory, and social cognition. Electrical stimulation of the amygdala evokes fear and anxiety responses in humans while lesions block certain types of unconditioned fear. For example, rats with lesions in the amygdala show reduced freezing in response to cats, or cat hair, attenuated analgesia, heart rate responses to loud noise, and have reduced taste neophobia.
However, amygdala lesions do not affect other measures of fear such as an open arm avoidance in an elevated plus-maze in rats or analgesia to shock. The amygdala is also necessary for learning by fear, amygdala lesions disrupt the acquisition of both active avoidance (escape from fear), and passive avoidance of conditioned responses, but do not affect retention.
The amygdala processes not only emotions of fear and aversive stimuli, but it is also involved in conditioning using stimuli of appetite such as food, sex, and drugs. As for its role in memory, the activation of the amygdala has a modulatory effect on the acquisition and consolidation of memories that evoke an emotional response.
Some parts of the amygdala have even more specific functions. The basolateral nucleus (BLA) is a cortical-like structure in the dorsal amygdala, and it regulates behavioral and physiological responses to stress. The central amygdala (CeA) plays a crucial role in physiological responses to stressors, such as fearful stimuli, stressful stimuli, and some drug-related stimuli. Meanwhile, the extended amygdala, named the bed nucleus of the stria terminalis (BNST), is involved in anxiety and stress.
reference link : https://www.ncbi.nlm.nih.gov/books/NBK537102/
No main effects of IPV or IPV-by-sex interactions were observed for the hippocampus, or any other subcortical region. In a subsample (n = 70) with diffusion imaging data available, we found robust evidence that infants exposed to IPV in utero possessed higher mean diffusivity in the uncinate fasciculus and corpus callosum and lower fractional anisotropy in the corticospinal tract.
Additional analyses suggest that these microstructural alterations also vary as a function of infant sex: diffusivity alterations were apparent in males but not females exposed to IPV. All the observed effects remained significant after excluding mother-infant dyads with pregnancy complications. Taken together, our results suggest that volumetric and microstructural brain alterations are observed in IPV-exposed infants shortly after birth, implying that they occur in the intrauterine environment.
Our observations of sex-specific effects of IPV exposure on neonatal amygdala volume builds on previous findings that have reported sex differences in amygdala volume in children and adolescents whose mother’s were exposed to antenatal psychological distress. However, the direction of effect was contrary to our initial prediction; amygdala volume was lower in IPV-exposed females in comparison to their unexposed counterparts, which contrasts with previous findings of enlarged amygdala volumes reported in older female children and adolescents (Acosta et al., 2019, Buss et al., 2012, Jones et al., 2019, Wen et al., 2017).
Interestingly, the pattern emerging from recent studies of neonatal amygdala structural connectivity in relation to higher maternal cortisol is one of enhanced maturation and greater connections in female neonates, and potentially reduced amygdala development in males (Graham et al., 2019, Stoye et al., 2020).
Taken together, these results suggest a possible delay in amygdala development in very young infant girls exposed to IPV and possible ‘catch- up’ growth thereafter, with brain growth patterns implicated in a vast range of psychiatric and developmental conditions (Brouwer et al., 2022). Longitudinal studies of brain development in this cohort will be able to shed light on this hypothesis.
In terms of other subcortical structures, we found no evidence that antenatal IPV exposure was related to variations to hippocampal volume in neonates, in support of previous literature (Buss et al., 2012, Favaro et al., 2015, Marečková et al., 2018), and there were also no effects on the thalamus, putamen, or pallidum. However, the caudate nucleus was found to be 5% larger among males, but not females, exposed to IPV.
To our knowledge, this is the first report of caudate nucleus volume being associated with prenatal stress exposure. Alongside a key role in executive functioning, sensory integration, and socio-emotional processing (Choi et al., 2022, Lanciego et al., 2012), caudate structure and function may have important neurodevelopmental implications.
For example, larger caudate volumes have been related to impaired problem solving and increased impulsivity in children with autism (Voelbel et al., 2006) and have recently been identified as a marker of severe neurodevelopmental delay within the first 2 years of life (Hüls et al., 2022). Given that our findings did not survive strict correction for multiple comparison testing suggests future studies may select the amygdala and caudate as a priori regions of interest in examining the neurological consequences of psychological distress during pregnancy.
Our examination of white matter microstructural integrity found that antenatal maternal IPV exposure was associated with higher MD in the uncinate fasciculus and corpus callosum, and lower FA in the corticospinal tract. Strikingly, only IPV-exposed boys, but not girls, showed changes in diffusivity in these tracts, with very large effect sizes (Cohen’s d = −0.90 and 0.97).
Lower FA/higher MD during development generally points to reduced microstructural integrity, including disrupted glial proliferation and maturation (Yoshida et al., 2013), as well as increased brain water content and decreases in axon density (Lautarescu et al., 2020, Wimberger et al., 1995).
However, the direction of effect can also depend on developmental stage (Tamnes et al., 2018). The uncinate fasciculus is part of the temporo-amygdala-orbitofrontal network, and diffusivity alterations within this tract have previously been associated with disorders that are more prevalent in males, such as conduct problems (Passamonti et al., 2012, Sarkar et al., 2013) and autism spectrum conditions (Catani et al., 2016), suggesting that these early microstructural alterations may have adverse developmental consequences.
The present study also supports previous literature which documents an association between prenatal psychological distress and changes to the diffusion of two major white matter pathways: the corpus callosum and corticospinal tract (Borchers et al., 2021, Zwicker et al., 2013, McCreary et al., 2016). Overall, our findings provide a picture of disrupted microstructural white matter integrity in tracts that are central to socioemotional functioning in association with IPV exposure during pregnancy, specifically in male infants.
Exposure to IPV is a particularly severe stressor, and intense or prolonged stress during pregnancy increases fetal exposure to stress biomarkers (e.g., cortisol and pro-inflammatory cytokines) which can alter the development of the nervous system (Goldstein et al., 2021). There are plausible biological mechanisms that can help explain how an infant’s biological sex may modify the effect of IPV on brain structure and connectivity, with sex chromosomes in the placenta likely to produce sex-specific transplacental signals to the developing brain (Bale, 2016, Carpenter et al., 2017).
Male and female fetuses also have different patterns of glucocorticoid expression during development, which may indicate different windows of vulnerability to cortisol exposure (Owen and Matthews, 2003). Further examination of these sex-specific effects may be key to understanding the sex bias in neurodevelopment disorders.
The present study has several strengths including its prospective design and the fact that the mothers and infants were well-characterized from a sociodemographic perspective. Statistical adjustment for plausible confounders that may impact fetal brain development is essential to isolate the specific impact of IPV exposure, with influential prenatal factors often not reported comprehensively in infant neuroimaging studies (Pulli et al., 2019).
These include intrauterine alcohol exposure (Archibald et al., 2001, Donald et al., 2015), HIV status (Wedderburn et al., 2022), smoking (Ekblad et al., 2015, El Marroun et al., 2014), and maternal depression (Barnett et al., 2021, Groenewold et al., 2022). The high rates of IPV exposure in this sample (44% of the sample) provide approximately equal group sizes meaning the cohort is well-suited for studying the impact of antenatal IPV. Crucially, scans were conducted at a median of 21 days of life, limiting exposure to postnatal environmental factors that may also contribute to neurodevelopment.
The study also had several limitations. First, the administered IPV questionnaire assessed the mother’s experiences of IPV over the preceding 12 months and is therefore not strictly specific to the gestational period. Therefore, we cannot rule out that some women may have been exposed to IPV prior to conception and not during pregnancy. Similarly, it is also possible that some women were exposed to IPV after the questionnaire was administered at the study visit (i.e., between 28 and 32 weeks and birth).
Second, the categorization of two groups based on IPV exposure does not capture significant heterogeneity within the IPV-exposed group; for example, some mothers may experience severe and chronic IPV whereas others may have only just reached the threshold for inclusion.
A minority of women (n = 11) reported exposure to all types of IPV (emotional, physical, and sexual) within the previous 12 months, whereas the majority of women were subjected to emotional abuse only. Nonetheless, our sample size limited the ability to examine the impact of exposure severity.
Third, we were unable to pinpoint the exact timing of IPV during pregnancy, with several studies highlighting the importance of considering the developmental window of vulnerability and its unique effects on brain outcomes (Acosta et al., 2019, Mareckova et al., 2022). Fourth, a greater proportion of IPV-exposed women drank alcohol during pregnancy (27%) compared to the control group (11%).
While we have adjusted for prenatal alcohol exposure in all models, we acknowledge that this is only a first-order linear approximation of alcohol effects on brain outcomes that may not fully capture the full extent of the association. Finally, only a sub-sample of neonates had diffusion imaging data available due to movement and technical artefacts (49% of the cohort with structural MRI), which minimized statistical power for detecting IPV-by-sex interactions on white matter microstructure.
Moreover, while TBSS is regarded as a standard approach for group comparisons of diffusion properties (Smith et al., 2006), concerns have been raised with regards to anatomical inaccuracies in skeleton projections (Bach et al., 2014). However, we sought to minimize this issue by using a custom template as a target in the registration step (Keihaninejad et al., 2012) to enhance registration quality and anatomical specificity (Bach et al., 2014).