TL;DR: In a small fMRI pilot built around Superstorm Sandy exposure, prenatal stress did not reliably scar the limbic system on its own. It only showed up as reduced emotion-circuit activation in the children who, by ages 2 to 6, had not built strong everyday adaptive skills.
Key Findings
- The exposure effect was conditional, not universal: Sandy exposure on its own did not reach significance. The link to lower limbic activation only appeared when childhood adaptive skills were low.
- Cohen’s d ranged roughly −1.12 to −1.46 at low adaptive skills: In children with weak adaptive skills, prenatal exposure tracked with large reductions in left amygdala, bilateral hippocampus, bilateral insula, and bilateral rostral ACC activation.
- Sandy gave the design a dated exposure window: 11 children whose mothers were pregnant when the storm struck the New York region in October 2012, vs. 23 children who were not.
- Adaptive skills are modifiable: The buffering variable — communication, self-care, social flexibility — is exactly the kind of capacity early-intervention programs can build after birth.
- Activation differences clustered in emotion-processing circuitry: Amygdala, hippocampus, insula, and rostral ACC together — a coherent network, not a single isolated spot.
- 34 children is a pilot, not a verdict: The signal needs replication in larger longitudinal cohorts before any clinical translation.
Source: Developmental Neuroscience (2026) | Deingeniis et al.
Prenatal stress can sound like a locked-in biological sentence — something that happens before a child is born and writes itself into the brain whether anyone wants it to or not. This pilot study asks a different and more useful question. When a fetus is exposed to a major stressor, do any postnatal capacities change how that exposure later shows up in emotion circuitry?
Their answer points toward adaptive skills — the practical, everyday abilities children use to communicate needs, manage routines, and navigate social demands. These are not IQ. They are not academic achievement. They are the kind of capacities families, schools, and therapists can actively build long after the prenatal exposure is in the past.
Why Superstorm Sandy Made the Design Possible
Most prenatal-stress research depends on retrospective self-report — asking mothers years later whether pregnancy felt stressful — and inherits all the noise of memory and meaning-making that comes with it. Sandy gave researchers something rarer: a major stressor with a fixed date.
The storm struck the New York region in October 2012. Children whose mothers were pregnant then counted as exposed; children whose mothers were not, did not. The exposure was not randomized — it never can be — but the timing anchor was clean, and the participants came from a Stress in Pregnancy cohort already enrolled before landfall. The disaster did not need to be reconstructed years later from a survey.
The neuroimaging cohort was small: 11 prenatally exposed children, 23 unexposed, all scanned at about age 8.
What “Adaptive Skills” Actually Means Here
From ages 2 to 6, the children’s adaptive functioning was tracked using the Behavior Assessment System for Children, Second Edition. The skills in question are concrete: communicating wants, managing daily routines, adjusting behavior, handling social demands. They are the ground-level abilities that let a child act in the world.
That distinction matters. Adaptive skills are not innate fixed traits. They develop through interaction, modeling, repetition, and intervention. If they turn out to moderate how prenatal stress later expresses in the brain, the implication is not “we couldn’t have helped this” — it is the opposite.
The Face-Matching Task and the Network It Was Designed to Wake Up
At age 8, each child completed an emotional face-matching fMRI task. On face blocks they matched angry or fearful expressions; on control blocks they matched geometric shapes. The face-vs-shape contrast estimated brain activity specifically tied to emotional face processing rather than basic visual-motor demand.
The target regions were chosen deliberately. The amygdala handles emotional salience — especially threat detection. The hippocampus supplies memory and context. The ventral anterior insula tracks bodily state and salience. The rostral anterior cingulate cortex ties emotion to regulation. Together they form the limbic circuitry most often implicated in stress vulnerability.
A difference spread across this whole network is more interpretable than a hit on any single region.
The Conditional Effect Is the Real Finding
The headline reading of “Sandy exposure changed brain activation” turned out to be wrong, or at least incomplete. Looking at exposure alone, only a few subthreshold trends appeared — exposed children tended toward lower adaptive scores and lower right insula activation, but neither cleared the statistical bar.
The interaction was where the signal lived. Adaptive skills moderated the exposure-to-brain relationship in the left amygdala, both hippocampi, both ventral anterior insulae, and both sides of the rostral ACC. Among children with low adaptive skills, prenatal exposure tracked with markedly reduced activation. Among children with average or higher adaptive skills, the activation pattern resembled that of unexposed peers.
The effect sizes in those low-skill simple slopes were large — Cohen’s d values from roughly −1.12 to −1.46. The small sample limits how confidently those numbers can be quoted, but the direction was consistent across regions, which is what gives the pattern its weight.
Why “Lower Activation” Needs a Careful Read
The authors are appropriately cautious about interpreting the activation difference. More activation is not automatically healthier — it can mean effort, compensation, inefficiency, or just heightened arousal. Less activation is not automatically worse for the same reasons.
Still, in this set of regions, blunted response to fearful or angry faces in low-adaptive-skill exposed children fits a pattern that the developmental-stress literature has been circling for a while. The novelty of this study is not the activation difference itself; it is that the difference appeared only in the children who had not built strong everyday skills by middle childhood.
That moves the prenatal-stress conversation away from a fixed-damage frame. Postnatal functioning seems to shape how prenatal adversity later expresses in emotion circuitry — and the postnatal variable here is one that interventions already know how to target.
Why the Pilot Has to Stay a Pilot
34 children is small for fMRI, especially when the exposed cell holds 11 of them and the analysis covers multiple brain regions. The authors corrected for child sex, age at scan, parental age, maternal race, and socioeconomic class, and used false-discovery-rate correction on the interactions. Those choices help — but they cannot rescue a sample this small from the basic statistical risks of imaging research.
There is also a causal limit. Brain activation at one age cannot show whether adaptive skills produced a better trajectory or whether they marked some other family or child resource — engaged caregivers, language-rich environments, lower household stress — that did the actual work.
What this study earns is a hypothesis worth testing in larger longitudinal cohorts: that the way prenatal adversity expresses itself in the developing brain depends, at least partly, on capacities that can still be built after birth. If that holds up, early communication, daily-living, and social-flexibility work stops being a generic “good for kids” recommendation and becomes a specific intervention target for children with high prenatal-stress exposure.
Citation: Deingeniis et al. Adaptive Skills May Moderate the Association between Prenatal Stress Exposure and Limbic Brain Activation: A Developmental Functional Magnetic Resonance Imaging Study of Superstorm Sandy Exposure. Developmental Neuroscience. 2026. DOI: 10.1159/000551574
Study Design: Quasi-experimental developmental fMRI pilot using prenatal Superstorm Sandy exposure.
Sample Size: 34 children — 11 prenatally exposed to Superstorm Sandy, 23 unexposed.
Key Statistic: At low adaptive skills, prenatal exposure linked to reduced activation across left amygdala, bilateral hippocampus, bilateral insula, and bilateral rostral ACC, with simple-slope Cohen’s d roughly −1.12 to −1.46.
Caveat: Small pilot sample; replication needed in larger longitudinal cohorts.
