Introduction

Climate change is increasing ambient temperature and causing natural disasters (e.g., hurricanes, tropical storms, wildfires, and droughts) with heightened frequency and severity. In 2012, Superstorm Sandy (SS), a category 3 hurricane at its peak in the Caribbean, made landfall in the New York City Metro area as a post-tropical cyclone. The storm caused extensive damage and ranked as one of the deadliest and most destructive storms in U.S. history [1]. Climate evidence suggests several warming-related factors (e.g., above-average sea surface temperatures, sea ice loss in the Arctic, and increased atmospheric moisture) may have intensified the storm’s impact [2]. Moreover, almost all the warmest years on record have occurred since 2000 and climate models project the next five years will be the hottest 5-year period on record [3,4].

While exposure to natural disasters and extreme ambient heat poses physical and mental health risks to the general population, certain groups, including pregnant individuals, especially those living in urban and under-resourced areas, are more vulnerable to its effects [5,6]. Pregnancy induces various physiological changes to the body, which makes thermoregulation more challenging and increases the risk of hyperthermia [7]. It is crucial to recognize that the negative consequences of exposure to prenatal stress are not limited to pregnant individuals. Exposure to stress during pregnancy increases the risk of poorer neurodevelopmental outcomes among child offspring [8,9]. This may initially present as disrupted emotion regulation mechanisms, such as difficulty with managing emotions and coping with stressful postnatal experiences [10]. As children grow, these early disruptions to emotion regulation may increase vulnerability to neurodevelopmental challenges and alterations in structural brain volume and functional activity. These effects are shaped by multiple factors, including the timing of exposure, genetic predispositions, and postnatal experiences [10].

Numerous studies have found associations between non-climate related stressors, such as maternal anxiety and/or depression, and greater externalizing (e.g., hyperactive, and aggressive) and internalizing (e.g., anxious and depressive) behavioral symptoms among child and adolescent offspring [11–16]. These associations between perceived stress and child behavioral symptoms have been replicated with measurable climate-related stressors, such as exposure to natural disasters (e.g., the 1998 Quebec Ice Storm and SS), via quasi-experiments among human populations [17,18].

As introduced above, exposure to extreme heat significantly threatens human physical health. It is important to recognize that various sociodemographic, infrastructural, and environmental factors strongly influence an individual’s susceptibility to adverse consequences from heat. For example, evidence suggests that those who represent a lower socioeconomic status (SES) and/or minority race are far more likely to experience negative consequences from heat exposure [19]. Moreover, heat morbidity is worsened in urban areas, like New York City, due to the urban “heat island” where man-made surfaces (e.g., concrete and dark roofs) absorb most of the sunlight and re-radiate it as heat while less vegetation is present to cool the air by evapotranspiration [20]. While studies examining the impacts of prenatal exposure to extreme heat on child mental health outcomes are still in their early stages, extreme heat exposure in-utero has been strongly linked to adverse obstetric outcomes (e.g., preterm birth and low birthweight) via reduced blood flow to the placenta, dehydration, and increased inflammation [7,21]. Of note, adverse obstetric outcomes among pregnant parents are a risk factor for various suboptimal developmental outcomes in children (e.g., impaired cognition, reduced physical growth, chronic health issues, and neuropsychiatric disorders) [22].

Although the link between climate and non-climate-related prenatal stress and increased child problem behaviors has gained considerable support, less attention has been paid toward understanding their consequences for child brain development. Optimal fetal brain development is critical as the brain grows most rapidly during the fetal period, reaching more than a third of the volume of the adult brain by the time of birth [23]. Neural phenotyping using magnetic resonance imaging (MRI) has been employed to help understand the biophysiological impacts of prenatal stress on child brain development. Non-climate-related (e.g., prenatal anxiety, depression, and normative stress) and climate-related (e.g., Quebec Ice Storm and Superstorm Sandy) stressors have been associated with discernible alterations in the gray matter volume and thickness of various cortical (e.g., prefrontal cortex) and subcortical (e.g., amygdala) brain regions among child offspring [24–29]. However, these findings remain limited and elusive, as studies contain mixed results with both smaller and larger brain volumes being reported.

The basal ganglia, a group of subcortical nuclei comprised of the striatum (caudate-putamen), pallidum (globus pallidus and ventral pallidum), and nucleus accumbens (NAcc) [30,31], is a promising region for examining optimal child neurobehavioral development. While initially thought to facilitate voluntary motor regulation and functioning [31,32], it has gained increasing attention in recent years for its ability to influence various non-motor functions, including executive functioning, working memory, language, reward valuation, and emotion recognition [32,33]. The non-motor role of the basal ganglia is evidenced by neuroimaging studies which have found basal ganglia subregions to form extensive circuits with limbic structures during affective tasks [32,33]. Moreover, functional alterations in the basal ganglia have been detected in various non-motor-related psychiatric disorders, such as anxiety, depression, and obsessive-compulsive disorder (OCD) [34].

Preliminary evidence has found strong associations between non-climate-related prenatal stress and structural alterations in basal ganglia brain volume. Prenatal mental health problems have been associated with decreased volumes in the basal ganglia, particularly the putamen among child offspring [35]. In contrast, increased volumes in the caudate, putamen, and pallidum during infancy have been attributed to maternal antenatal depression [36,37]. Our own recent work using a machine learning algorithm on a smaller sample of the stress in pregnancy (SIP) cohort found in-utero SS exposure to be highly associated with a larger right caudate [29]. It is important to conduct additional research to clarify these associations as they remain varied with both smaller and larger brain volumes reported.

Despite its potential importance, no study to date has investigated the individual and joint impact of two prominent and interconnected climate-related prenatal stressors, such as SS and ambient heat, on the brain volume of the basal ganglia among school-aged children. The study investigates the independent associations of both SS and extreme heat (main effects) and the interaction of the two (joint effect) on the brain volume of the core regions of the basal ganglia (i.e., caudate, putamen, pallidum, and nucleus accumbens). First, we investigate whether exposure to SS and extreme heat will individually lead to discernible differences in basal ganglia structures. We next explore whether exposure to extreme heat in-utero will amplify the risk prenatal exposure to SS poses on basal ganglia brain development as measured by changes in brain volume.

Methods

Results

Demographics

Except for child age at MRI assessment, no demographic differences were detected between the SS exposed and unexposed children. The age difference occurred due to more children being born prior to the storm than after. Full demographic differences are summarized in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and Environmental Characteristics of Participants by Superstorm Sandy Exposure Status.

https://doi.org/10.1371/journal.pone.0324150.t001

Individual effect of Superstorm Sandy (SS) and extreme heat exposure on child brain volume

Main effects of in-utero SS and extreme heat exposure on child brain volume at age 8 were examined. In-utero SS exposure was associated with a larger left (+6.35%, adj-p = .001) and right (+6.22%, adj-p < .001) putamen, larger left (+1.89%, adj-p = .04) and right (+1.71%, adj-p = .007) pallidum, and a larger right caudate (+3.92%, adj-p = .08). Apart from the marginal association between extreme heat and a larger left pallidum (p = .09), no associations were found with extreme heat exposure. Full numerical results are summarized in Table 2 (SS) and Table 3 (heat). Supplementary figures S1 Fig (SS) and S2 Fig (heat) provide visual representations of these main effects.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. SS Exposure on Basal Ganglia Structural Volumes.

https://doi.org/10.1371/journal.pone.0324150.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Extreme Heat Exposure on Basal Ganglia Structural Volumes.

https://doi.org/10.1371/journal.pone.0324150.t003

Interaction effects of Superstorm Sandy (SS) and extreme heat exposure on child brain volume

Extreme heat exposure moderated the association between SS exposure and the volume of the left NAcc (adj-p = .05) and left pallidum (adj-p = .14) at age 8. When heat exposure was present, SS exposure was associated with a significantly smaller left NAcc and a larger left pallidum. Full numerical results are summarized in Table 4 and visually represented in supplementary figure S3 Fig.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Interaction Effects of SS and Extreme Heat Exposure on Basal Ganglia Structural Volumes.

https://doi.org/10.1371/journal.pone.0324150.t004

Discussion

There are two major findings. First, SS exposure during pregnancy, but not heat, was associated with larger gray matter volume in the putamen and pallidum bilaterally and the right caudate in middle childhood (approximately age 8) compared to those who were SS unexposed. Second, prenatal exposure to extreme heat amplified the adverse impact SS exposure poses on basal ganglia gray matter volume. Our significant main effects of SS exposure on brain volume are consistent with prior research that demonstrates the influence of prenatal stress exposure on child brain volume. However, this is one of the first studies to demonstrate the effect of prenatal stress in response to a natural disaster on basil ganglia gray matter volumes. Importantly, while we did not find main effects of extreme heat on basal ganglia gray matter volumes, our study was the first to uncover that SS exposure, when coupled with extreme heat, was linked to both discernibly larger (e.g., pallidum) and smaller (e.g., NAcc) basal ganglia gray matter volumes compared to unexposed peers. Our findings provide initial evidence for the importance of evaluating co-occurring risks related to climate change on child brain morphology while also expanding existing knowledge in the field.

Limitations

The current study has limitations. First, the sample size is small (n = 34), which prevented the exploration of potential sex or trimester-specific risks. Exploring sex-based comparisons is especially important as it plays a crucial role in child developmental processes and morphological characteristics of brain structures. Second, our analysis was limited to structural brain indices. Additional investigation using functional imaging analysis would enhance our understanding by identifying potential subcortical activation patterns in response to prenatal stress. Moreover, considering the interdependent relationship between the brain and behavior, future work should explore whether clinical behavioral indices influence basal ganglia volume and vice-versa. Third, we could not assess brain volume longitudinally, nor could we determine whether the observed brain volume differences exceed typical developmental expectations due to lack of normative comparison data. We are currently conducting a longitudinal follow-up study based on these preliminary findings. Our imaging protocol models the Adolescent Brain Cognitive Development (ABCD) study, a large-scale longitudinal study which tracks the brain development of several thousand 9–10 year olds through adolescence [57]. Our ongoing research will allow us to track brain volume development throughout childhood in relation to these early exposures while also establishing better normative benchmarks for comparison. Fourth, it is important to recognize the difficulty of measuring heat exposure, as there is no standard definition of extreme heat episodes [58]. For example, global studies have associated adverse pregnancy outcomes with exposure to temperatures between 84.2°F (29°C) to 107.6°F (42°C) [59–63]. Despite this, we chose 95°F (35°C) as the threshold temperature because in New York City, specifically, heat-related morbidity (e.g., exhaustion, sun stroke, or heat-related fatigue) is possible at 95°F (35°C) even with lower relative humidity (%) levels [64]. Moreover, the World Health and Meteorological Organizations (WHO; WMO) state the use of electric fans at 95°F (35°C) may worsen heat-related stress [65]. Nonetheless, our findings may not be fully generalizable to populations residing in non-urban environments or regions with different climatic conditions. Fifth, as commented earlier, it is challenging to ascertain the full extent of individual-level heat exposure. Considering not everyone spends the same time indoors and outdoors, has access to inside air conditioning, proper ventilation, and greenspace, it is challenging to precisely discern how long each day, and to what extent (whether indoors or outdoors), the participants were exposed to the threshold temperature. Despite accounting for some of these influences in our current analyses by adjusting for the New York City neighborhood-level HVI, our results should be interpreted with caution. To elucidate this association, future work should consider measuring more precise measures of heat exposure (e.g., outdoor vs indoor occupation and individual-level access to air conditioning, proper ventilation, and cool public spaces) that may influence the degree to which heat exposure impacts child outcomes [19]. Moreover, as mentioned above, potential delayed effects of prolonged heat exposure should be considered since short-term exposures may be insufficient to induce significant changes in brain volumes as evidenced by our study results. Nonetheless, we found a marginally significant main effect between heat exposure and left pallidum volume. However, this finding was not corrected for multiple comparisons, necessitating cautious interpretation. Furthermore, trimester-specific risks should be considered to identify potential critical developmental windows. Additionally, we found a statistically significant age difference between the SS-exposed and non-exposed groups with the SS-exposed group having a lower mean age. Despite adjusting for biological age at MRI assessment in all our analyses, this baseline difference could nonetheless potentially influence the volume of the basal ganglia. Finally, some potential sources of bias remain. This includes selection bias, as participants undergoing neuroimaging may not be fully representative of the general population, measurement error inherent in neuroimaging techniques, and residual confounding from unmeasured factors. However, we believe these biases are more likely to attenuate rather than inflate the observed associations.

Strengths

This study also has strengths. First, the study’s quasi-experimental design allowed our measures of prenatal stress exposure (i.e., SS and extreme heat) to be objectively defined and naturally assigned in a human population. Second, despite the small sample size, our findings are robust to multiple testing adjustment, intrafamilial correlation (n = 4 siblings), and various pertinent confounders connected to the pregnant individual (e.g., pregnant individual’s age at childbirth, SES, and normative stress class), their offspring (e.g., age at MRI assessment and sex assigned at birth), and their environment (e.g., season of conception and neighborhood heat vulnerability).

Conclusion

Our findings warrant additional precautions regarding the health dangers of exposure to extreme climate events and natural disasters, specifically for pregnant individuals. Exposure to SS during pregnancy impacted child brain development evidenced by discernibly larger basal ganglia gray matter volume. Extreme heat amplified this risk via increased and decreased volume in various basal ganglia subregions of interest. The world is in a state of climate emergency. Extreme weather events and natural disasters are projected to increase in frequency and magnitude [66]. In addition to promoting initiatives to combat climate change, it is imperative to alert pregnant individuals to the ongoing danger of exposure to extreme climate events in order to protect healthy long-term brain development in young school-aged children. Our results warrant future work to investigate the link between alterations in basal ganglia volume and behavioral disruption, with a potential role of basal ganglia structural variations to either reveal behavioral disruptions or serve as a risk factor for impairment.

Supporting information

Acknowledgments

The authors would like to acknowledge the Magnetic Resonance Imaging Facility of the CUNY Advanced Science Research Center for instrument use and technical assistance. We would like to acknowledge Jessica Buthmann and Wai Man Wong for their assistance in recruiting and scanning our study participants and Barbara Kinsella-Kammerer and Kathryn Sie for research coordination.