Opioid use and misuse is an ongoing public health crisis. Opioid use disorder (OUD) in pregnant women in the U.S. has increased considerably over the last few decades (1, 2). Opioid related morbidity and mortality were initially driven by the abuse of prescription opioids in the 1990s (3) followed by marked rise in heroin use around 2010 (4), and later by increasing use of synthetic opioids such as illicitly manufactured fentanyl starting in 2013 (5). Multiple opioids may be misused simultaneously with or without other non-opioid substances (6). According to the Centers for Disease Control and Prevention 2019 data, 7% of women reported using prescription opioids during pregnancy, of whom more than one in five reported opioid misuse (1). The number of opioid involved deaths has doubled in pregnant women from 2007 to 2016 (7). In pregnant women with untreated OUD, fluctuating opioid levels in the maternal bloodstream and periods of withdrawal lead to fetal distress, negatively impacting fetal development and increasing the odds of still birth, poor fetal growth and preterm birth (8–12). Infants with prenatal opioid exposure (POE) are at a high risk for developing withdrawal symptoms after birth–neonatal opioid withdrawal syndrome (NOWS)–which occurs in 55–94% of these infants, and results in prolonged and expensive hospital stays (12–17). Children with a history of NOWS are also at risk for adverse long-term developmental outcomes that may persist to school age, including poor cognitive scores, impaired executive functioning and worse fine motor scores (18–21).

There are a small handful of neuroanatomical studies published so far that have evaluated POE in the fetus, neonate, and infant. These studies have suggested alterations in regional brain volumes (22, 23), microstructure (24–26), and impaired brain functional connectivity (24, 27, 28). However, these aforementioned studies did not specifically assess for the potential effects of maternal comorbidities. Pregnant women with OUD have a high prevalence of coexisting comorbidities including psychological illnesses such as depression, stress and anxiety (29–33), polysubstance use (34), and smoking (35). A few neuroimaging studies have attempted to assesses infant brain development in the presence of these maternal high risk factors individually, with or without polysubstance exposure, but there needs to be a more comprehensive understanding the role of these common concomitant maternal comorbidities in the setting of POE, specifically, their potential influence fetal and infant brain development (35–37).

The human brain grows rapidly in the fetal period and during infancy. Identifying alterations in normal brain developmental trajectory is beneficial in understanding brain function in developmental disorders (38–40). Graph-theoretical network modeling has previously been used to understand normal brain development and has revealed highly efficient topological architectures in the early life (41). Early brain structural connectome network topologies in the infant are shown to have predictive ability for later life cognition and behavior (41–44). Here, we applied brain functional network modeling techniques to understand global brain network properties in infants with POE. The aim of our study was to identify differences in global brain network connectivity in infants with prenatal opioid and drug exposure compared to healthy control infants, using resting-state functional MRI performed at <3 months completed postmenstrual age. We also assessed for associations of maternal comorbidities such as smoking, polysubstance use and maternal psychological factors on infant brain global network connectivity.

This study comparing prenatal opioid exposed infants with healthy infants without prenatal opioid exposure demonstrates significant alterations in global brain network connectivity in infants with POE, with some of these networks also showing alterations in graph theory metrics. Most importantly, our study identified that some of these global brain network connectivity alterations were associated with maternal psychological factors, a common co-morbidity in pregnant women with opioid use disorder. This suggests a role of the interplay between maternal opioid use and psychological disorders on the developing infant brain.

Brain resting state functional networks exist in infants and have recently been shown to be altered in the presence of prenatal exposure to substances (27, 28, 57). However, prenatal substance exposure has also shown to be associated with alterations in multiple other brain regions, both intra-network and inter-network, outside of these canonical resting state functional networks (57). Our study focused on a network-free model of global brain functional connectivity, rather than predefined networks. In our study, the identified CAP networks associated with POE spanned all lobes of the brain including the frontal, temporal, occipital, parietal and insula, suggesting a potential global influence of prenatal substance exposure on the developing brain. Many of the brain regions that comprised the CAP networks have been described as components of several networks including salience, executive function, rewards, dorsal and ventral attention, somatosensory, language, and visual networks. Two of the CAP networks in our study showed higher intra-network connectivity in infants with POE compared to control infants, while two of the networks showed lower intra-network connectivity. Similar findings of network hyper- and hypo-connectivity have been described in infants with opioid and other prenatal substance exposure (58). Functional MRI studies in adults with drug addiction reveal intra-network hyperconnectivity in the reward and salience networks and hypoconnectivity in the executive function network, the medial frontal regions, anterior cingulate gyrus and temporal regions (59, 60). We postulate that the CAP network alterations in infants with POE observed in our study, are a combination of hyperconnectivity and hypoconnectivity within and between these known resting state functional networks.

There have been a few recent small cohort studies showing alterations in brain resting state functional connectivity in infants with prenatal opioid exposure (27, 28). To date, ours is the largest cohort study focused on infants with POE. While Merhar et al. also studied global brain networks in infants with POE, and their results suggest correlations of brain connectivity with Finnegan score, we did not find any such correlation in our study, which is likely due to the differences in methodology, as we used the CAP method for brain network assessment after controlling for gestational age and sex (27). Additionally, we also identified alterations in infant brain resting state functional network connectivity associated with maternal psychological factors. Specifically, presence of maternal psychological factors showed intra-network hypoconnectivity in infants with POE in CAP C5 and C6, suggesting that maternal psychological factors should be considered when assessing for opioid related effects on the developing brain.

Imaging studies have also identified alterations in infant brain development in the setting of antenatal maternal depression, with these brain functional alterations shown to be associated with short term infant behavioral outcomes. Alterations in amygdala functional connections to several regions including the prefrontal cortex are identified in infants born to mothers with prenatal depression compared to control healthy infants (36, 37). Infant brain functional network hub distribution was also shown to be altered in those exposed to prenatal maternal depression, and these brain network alterations were predictive of 6 month infant temperament variance, highlighting the clinical relevance of these altered neonatal brain functional organization (61). Maternal depression and anxiety have also been associated with altered white matter microstructure in infants, especially decreased neuritic density and increased mean, radial and axial diffusivity in specific brain regions (62). SSRIs used to treat depression are also shown to be associated with the relationship between infant brain functional connectivity and early infant behavior (57); however, none of the women in our study were on SSRIs during pregnancy. We did not identify any significant effects of polysubstance use on altered infant brain functional connectivity in our cohort of prenatal opioid exposed infants, which may partly be explained by the differential “signatures” of altered infant brain functional connectivity of different prenatal substance exposures (57, 58).

Since the fetal period and infancy are a period of rapid and dynamic brain development, there are likely to be variable effects of POE on the maturing brain during this period, and hence variable effects on long term cognition and behavioral outcomes, which may also be modified by the presence or absence of opioid maintenance therapy, exposure to other substances or nicotine, maternal psychological factors and various demographic factors such as socioeconomic status, maternal education, and employment. Understanding this complex interplay of prenatal and postnatal factors in the setting of POE and maternal mental health on brain development and their effects on long term outcomes is key to developing optimal intervention strategies in this population. Visual and visual-motor problems are suggested to be more common in children with POE (63), but there is no convincing evidence for long-term cognitive impairment. While a few studies indicate increased risks for educational delay, lower IQ, and poor language performance in children with POE, other studies show no significant differences in cognitive outcomes (63–67). There is, however, greater evidence to suggest an increased risk of poor long-term behavioral outcomes in children with POE, including difficulties with executive functioning (67), attention deficit hyperactivity disorder (ADHD) (68, 69) and autism spectrum disorders (ASD) (70, 71). Children with POE and a history of NOWS are shown to be more likely to have poor neurodevelopmental consequences including lower IQ, poor educational testing performance, lower attention, meet disability criteria, and require additional classroom therapies and services when compared to children with POE who did not develop NOWS, or healthy controls (18, 71–74).

Our study has a few limitations. Opioid use in pregnant women is a complex issue, often associated with several other socioenvironmental comorbidities, as well as maternal confounders including polysubstance use, smoking and psychological factors, which may independently influence infant brain development. Although we were able to show the primary differences between the opioid-exposed and healthy infant brain connectivity, this study was not powered to study all maternal socioenvironmental factors such as maternal education level. We showed that maternal psychological factors correlated significantly with infant brain global brain network connectivity. However, many mothers of the POE group also had polysubstance use and smoking while none of the controls reported substance use or smoking. Although we did not find a significant correlation of smoking and polysubstance use with the altered infant brain networks in our study, in the absence of a control group with these covariates, it remains a challenge to assign the cause of the findings solely to opioid use. Moreover, the long-term clinical significance of the observed changes is currently unknown and could be evaluated by large scale longitudinal studies, such as the NIH HEALthy BCD study, which should also address the effects other confounding maternal and environmental factors.

In conclusion, our study showed significant alterations in global network level resting state functional connectivity in infants with POE, compared to healthy infants without prenatal opioid exposure, with altered graph network metrics within these regions. Additionally, these alterations in global brain network connectivity were significantly associated with the presence of maternal psychological factors. Since alterations in global brain connectivity may play a role in developmental outcomes, our study suggests the need for large prospective longitudinal studies to understand the prognostic ability of these early functional brain network alterations, and the benefits of early intervention on childhood long term development.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Indiana University Institutional Review Board. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

RR, YZ, and JY: conceptualization. RR, YZ, DH, GS, JY, BH, NS, RV, and SS: methodology and review of final manuscript. RR, YZ, NS, RV, JY, and BH: formal analysis and investigation. RR, RV, and SS: writing—original draft preparation. RR, YZ, DH, JY, BH, RV, and SS: writing—review and editing. RR and SS: funding acquisition. RR, YZ, JY, and SS: resources. RR, JY, and SS: supervision. All authors contributed to the article and approved the submitted version.

RR was supported by the American Roentgen Ray Scholarship Award 2018 and Radiological Society of North America Seed Grant 2018. SS and RR were supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under Award, R01HD096800 (PI: SS).

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.