antidepressants in pregnancy: what the brain scans show, and what they don’t
picture the headline: “antidepressants in pregnancy linked to smaller brain regions in babies.” it is true, in the sense that a careful 2026 study found it. it is also true that in the same study the babies of untreated, severely depressed mothers looked the same, that a 2018 study found the opposite region larger, and that the biggest analysis of what actually happens to these children found the fathers’ prescriptions predicted the outcomes too. this page puts all of that on one table, because a scan without its context is a very effective way to frighten someone who is trying to stay well.
the 2026 study, as it was written
smaller hippocampi in the womb, and a control group that changes the reading
a team at children’s national hospital in washington scanned 182 pregnancies between 20 and 40 weeks: 62 women taking a serotonin reuptake inhibitor and 120 with no mental illness and no antidepressant.[1] the exposed fetuses had hippocampi about fifteen percent smaller, less cortical folding, and a smaller cortical surface. their placentas were larger. that is the headline, and it is a real finding in a careful study.
the same paper contains the sentence the headline drops. when the unexposed mothers were split by how depressed they were, the exposed fetuses were smaller than the babies of mothers with low or moderate scores but not smaller than the babies of severely depressed mothers who took nothing. and among the unexposed, severe depression alone meant smaller hippocampi. the authors list it plainly: they could not separate the drug from the condition it treats, they had no dose or timing data, nobody in the cohort had psychotherapy so the treated group likely skews sicker, and the cohort was mostly high-income and college-educated. their conclusion, verbatim: “the clinical significance and long-term neurodevelopmental consequences of these structural alterations remain unknown and are currently under study.” no drug company money; the authors declare no competing interests.
every scan study, on one table
six studies in ten years. read the “found” column against the “depression-only control” column: the studies that had one mostly found the drug-exposed babies different from the depressed-and-untreated babies too, except the largest and newest, where severe untreated depression looked the same.
| year | who | scanned | how many | found | depression-only control? |
|---|---|---|---|---|---|
| 2026 | Wu et al., Children’s National (DC)[1] | in the womb, 20–40 weeks | 62 exposed vs 120 | hippocampi about 15% smaller; less cortical folding and surface area; larger placenta | yes: exposed fetuses were no different from unexposed fetuses of severely depressed mothers |
| 2018 | Lugo-Candelas et al., Columbia[2] | about 3 weeks old | 16 exposed vs 21 untreated depression vs 61 | right amygdala and insula LARGER (medium to large effects); more amygdala–insula connectivity | yes: differed from both controls and untreated depression |
| 2016 | Jha et al., UNC[3] | newborn | 27 exposed vs 54 matched; 41 depression-only vs 82 | widespread white-matter microstructure differences; NO volume differences | yes: depression-only newborns showed no differences |
| 2019 | Rotem-Kohavi et al.[4] | 6 days old | 20 exposed vs 16 depression vs 17 | higher connectivity in an auditory network | yes: differed from both other groups |
| 2021 | Campbell et al.[6] | 7 days old | 24 exposed vs 48 | localized corpus-callosum microstructure differences, some sex-specific | no depression-only group |
| 2017 | Podrebarac et al.[5] | very preterm, scanned twice | 14 exposed of 177 | microstructural and metabolic differences | no depression-only group |
what they agree on: something is measurably different in exposed babies. what they do not agree on: where, or in which direction. a 2018 study found the amygdala and insula larger;[2] the 2026 study found the hippocampus smaller.[1] a 2016 study found no volume differences at all, only white-matter microstructure.[3] none is bigger than 62 exposed babies. none has followed its babies far enough to say what the differences mean.
why the fathers matter
scans measure structure. what parents want to know is outcomes: does the child end up with adhd, autism, a learning problem. the largest answer is a 2026 systematic review and meta-analysis in the lancet psychiatry: 37 studies, about 650,000 exposed and 25 million unexposed pregnancies.[7] on its face it found what the headlines say: modestly higher rates of adhd and autism after prenatal antidepressant exposure.
then it did the test that decides these questions. it looked at fathers who took antidepressants during the pregnancy. a father’s pill cannot reach the fetus, so if the children of medicated fathers show the same excess, the cause is not the drug; it is whatever runs in families that leads a parent to be on the drug. the children of medicated fathers showed the same excess, for both adhd and autism, at about the same size. there was no dose–response for autism, and the associations were just as strong for exposure before conception. once the authors accounted for maternal mental illness, family and genetic factors, the associations “were attenuated or became non-significant,” and with confounding by indication minimised, “no significant associations were found for specific SSRIs or SNRIs.” they rate their own evidence “low to very low” certainty. a scan cannot do that test. this analysis did.
what to take from this
- the 2026 fetal study is real, careful, small, and says of itself that the meaning of its finding is unknown. quote it with that sentence attached or not at all.
- six scan studies disagree on region and direction; the only consistent result is “different,” which is also what severe untreated depression produced in the newest one.
- the fathers’ result is the single most useful fact on this page for someone deciding: the outcome studies carry the signature of confounding, and the specific SSRIs show no significant risk once it is handled.
- the regulators’ own record is the same shape: a warning, a number, a walk-back.[8] untreated depression has its own page and its own numbers.[9]
- never stop on your own. every source here says it.
questions people ask
Do antidepressants in pregnancy change a baby’s brain?
Six imaging studies since 2016 each found measurable differences in exposed babies: smaller hippocampi and less cortical folding in a 2026 fetal MRI study, a larger amygdala and insula in a 2018 newborn study, white-matter and connectivity differences in others. The studies are small (16 to 62 exposed babies each), they do not agree on which regions or in which direction, and every one of them says the meaning of the differences is unknown. The 2026 authors’ own words: “The clinical significance and long-term neurodevelopmental consequences of these structural alterations remain unknown and are currently under study.”
Is it the drug or the depression?
The studies cannot fully separate them, and the best of them say so. In the 2026 fetal study, babies of severely depressed mothers who took nothing had hippocampi the same size as the exposed babies. In the 2016 study, the depression-only newborns showed no differences while the SSRI newborns did. The strongest evidence on the question is not a scan: in the 2026 Lancet Psychiatry meta-analysis of 25 million pregnancies, fathers’ antidepressant use during pregnancy predicted childhood ADHD and autism about as well as mothers’ use did. A father’s prescription cannot reach the fetus, so that pattern points at the family and the illness, not the drug. Once confounding was accounted for, no specific SSRI showed a significant association.
What is the actual size of the difference in the 2026 study?
Left hippocampus 0.45 versus 0.53 cubic centimetres, right 0.48 versus 0.56, in fetuses scanned between 20 and 40 weeks. About fifteen percent smaller on average, with the exposed mothers also heavier and more likely to report alcohol use in pregnancy, which the authors adjusted for and tested. Whether a difference of that size in the womb means anything for a child is the question the same laboratory says it is studying now.
Should I stop my antidepressant because of this?
Not from a page, and not from a scan. The regulators’ notices, the 2026 fetal study and the 2026 meta-analysis all point the same way: talk to your prescriber, weigh the risk of untreated depression, and do not stop on your own. The untreated side has its own page here, with numbers.
sources
- Wu Y, Andescavage N, Wisner KL, Kapse K, Murnick J, Ngwa JS, Limperopoulos C. Prenatal serotonin reuptake inhibitor exposure and maternal depression symptoms are associated with altered fetal brain and placental development. Neuropsychopharmacology. Published 9 May 2026. DOI 10.1038/s41386-026-02436-9. PMID 42106476. Open access (CC-BY). Fetal MRI at 20–40 weeks; 62 SRI-exposed (59 SSRI, 3 SNRI) vs 120 controls without mental illness or antidepressant exposure. Hippocampal volume left 0.45 vs 0.53 cm³, right 0.48 vs 0.56 cm³ (adjusted p = 0.0004); gyrification index 1.24 vs 1.28; curvedness 0.21 vs 0.24 mm⁻¹; cortical surface area 143.35 vs 151.36 cm². Exposed fetuses were not different from unexposed fetuses of mothers with high depression scores; among unexposed, high maternal scores alone meant smaller hippocampi. Placental volume and diffusion increased with exposure. Verbatim conclusion: “The clinical significance and long-term neurodevelopmental consequences of these structural alterations remain unknown and are currently under study.” Funded by the A. James & Alice B. Clark Foundation; authors declare no competing interests. Limitations stated by the authors: no dose, timing or duration data; wide gestational-age span; depression itself may alter fetal activity; no participant had psychotherapy, so the treated group may be more severely ill; cohort mostly high-income and college-educated. https://doi.org/10.1038/s41386-026-02436-9
- Lugo-Candelas C, Cha J, Hong S, et al. Associations between brain structure and connectivity in infants and exposure to selective serotonin reuptake inhibitors during pregnancy. JAMA Pediatrics. 2018;172(6):525-533. DOI 10.1001/jamapediatrics.2017.5227. PMID 29630692. 98 infants scanned at about 3 weeks: 16 SSRI-exposed, 21 exposed to untreated maternal depression, 61 controls. Greater gray-matter volume in the right amygdala (Cohen d 0.65) and right insula (d 0.86) and increased amygdala–insula connectivity (d 0.99) in the SSRI group versus both other groups. https://pubmed.ncbi.nlm.nih.gov/29630692/
- Jha SC, Meltzer-Brody S, Steiner RJ, et al. Antenatal depression, treatment with selective serotonin reuptake inhibitors, and neonatal brain structure: a propensity-matched cohort study. Psychiatry Research: Neuroimaging. 2016;253:43-53. DOI 10.1016/j.pscychresns.2016.05.004. PMID 27254086. 27 SSRI-exposed neonates matched to 54 controls; 41 depression-history, no-SSRI neonates matched to 82 controls. Widespread white-matter microstructure differences in the SSRI group; no differences in the depression-only group; no significant differences in global or regional tissue volumes. https://pubmed.ncbi.nlm.nih.gov/27254086/
- Rotem-Kohavi N, Williams LJ, Virji-Babul N, et al. Alterations in resting-state networks following in utero selective serotonin reuptake inhibitor exposure in the neonatal brain. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging. 2019;4(1):39-49. DOI 10.1016/j.bpsc.2018.08.004. PMID 30292808. 6-day-old neonates: 20 SSRI-exposed, 16 depression-exposed without SSRI, 17 controls. Higher connectivity in an auditory resting-state network in the SSRI group versus both other groups. https://pubmed.ncbi.nlm.nih.gov/30292808/
- Podrebarac SK, Duerden EG, Chau V, et al. Antenatal exposure to antidepressants is associated with altered brain development in very preterm-born neonates. Neuroscience. 2017;342:252-262. DOI 10.1016/j.neuroscience.2016.11.025. PMID 27890829. 177 very preterm neonates, 14 SSRI-exposed; microstructural and metabolic differences on serial MRI. https://pubmed.ncbi.nlm.nih.gov/27890829/
- Campbell KSJ, Williams LJ, Bjornson BH, et al. Prenatal antidepressant exposure and sex differences in neonatal corpus callosum microstructure. Developmental Psychobiology. 2021;63(6):e22125. DOI 10.1002/dev.22125. PMID 33942888. Diffusion imaging at day 7: 24 SSRI-exposed vs 48 unexposed term neonates; localized corpus-callosum differences, some sex-specific. https://pubmed.ncbi.nlm.nih.gov/33942888/
- Chan JKN, Zhong X, Lam H, et al. Maternal and paternal antidepressant use before and during pregnancy and offspring risk of neurodevelopmental disorders: a systematic review and meta-analysis. Lancet Psychiatry. 2026;13:472-484. DOI 10.1016/S2215-0366(26)00089-1. PMID 42134364. 37 studies, 648,626 exposed and 24,967,806 unexposed pregnancies. Prenatal antidepressant use: neurodevelopmental disorders RR 1.13 (1.08–1.18); ADHD 1.35 (1.24–1.47); autism 1.69 (1.24–2.30); no dose–response for autism; similar associations for pre-conception exposure. Paternal use during pregnancy, the negative control: ADHD 1.46 (1.38–1.56), autism 1.28 (1.16–1.40). Verbatim: associations “were attenuated or became non-significant in sensitivity analyses accounting for confounding factors”; when confounding by indication was minimised “no significant associations were found for specific SSRIs or SNRIs”; certainty of evidence “low to very low.” https://pubmed.ncbi.nlm.nih.gov/42134364/
- The regulators’ notices in order, with the 2011 FDA walk-back: veisund.com/topics/who-warned-about-antidepressants-in-pregnancy. https://www.veisund.com/topics/who-warned-about-antidepressants-in-pregnancy
- Untreated depression in pregnancy, both directions, with the numbers: veisund.com/topics/depression-and-anxiety-in-pregnancy. https://www.veisund.com/topics/depression-and-anxiety-in-pregnancy
related: who warned about antidepressants in pregnancy · depression and anxiety in pregnancy
adam