Autism spectrum disorder is diagnosed far more frequently in boys than in girls, with a male-to-female ratio of approximately 3:1. Although females with autism often engage in camouflaging (or masking)—strategies that make autism-related traits less noticeable to adapt to a neurotypical society1—the biological mechanisms underlying sex differences in autism risk remain unclear. In particular, it is not well understood how these differences arise during pregnancy or early caregiving after childbirth. This project was motivated by a long-standing question: if maternal perinatal depression increases the risk of atypical neurodevelopment in children, does this risk differ between male and female offspring?
New Insights from the Current Cohort
This cohort study is based on the Tohoku Medical Megabank Birth and Three-Generation Cohort in Japan, initiated in 2011. Assessments were conducted longitudinally from early pregnancy through 1 month postpartum, as well as during early childhood (2–3 years of age), a period chosen to minimize the influence of camouflaging behaviors in girls. In addition, mothers who were prescribed fluoxetine—an antidepressant reported to potentially influence autism risk2—were consistently excluded. This design enabled the construction of a model that more accurately evaluates maternal perinatal depression.
Evaluation measures included maternal mental health during early, mid, and late pregnancy and at 1 month postpartum, mother–infant bonding, and autistic-related traits in toddlers. Data from 23,218 mother–child pairs were analyzed. Maternal history of psychiatric medication use and family history of neurodevelopmental disorders were also considered.
The finding in the Cohort Study
First, we observed an overall pattern: higher levels of maternal psychological distress during pregnancy and more severe depressive symptoms after childbirth were associated with higher autistic-related trait scores in toddlers. A more unexpected finding emerged when the data were stratified by child sex. The associations between maternal perinatal depression and autistic-related traits were substantially stronger in girls than in boys. At first glance, this may seem counterintuitive, as boys showed higher autistic-related trait levels overall, consistent with the well-established male predominance in autism prevalence. However, these findings are not contradictory. While boys exhibit higher baseline levels of autistic-related traits at the population level, maternal perinatal depression appeared to increase risk more markedly in girls than in boys. We also identified clinically relevant correlates. Female infants born to mothers with maternal perinatal depression tended to have lower birth weight. In addition, among girls, autistic-related traits were more strongly associated with maternal distress during mid-gestation and with difficulties in mother–infant bonding during the postpartum period. Although none of these findings alone establishes a causal relationship, together they point to a consistent pattern that supports a biologically plausible pathway and motivates further investigation using experimental models.
Animal Model to Clarify Biological Processes in the Brain
To explore the biological mechanisms underlying our cohort findings, we established a mouse model designed to approximate maternal psychological distress and to test potential causal links between prenatal stress exposure, oxytocin signaling—the hormone central to mother–infant bonding—and offspring developmental outcomes. Using a chronic unpredictable mild stress paradigm during late gestation, stressed dams showed depression-like behaviors and impaired maternal care. Their offspring showed reduced birth weight, mirroring the observations in the human cohort.
Autistic-Like Behaviors in the Animal Model: Evaluation of Grooming
To quantify autistic-like grooming behaviors, we developed an automated behavioral analysis pipeline using DeepLabCut3. Mice were labeled at 21 anatomical points, from the nose to the tip of the tail, allowing us to capture fine-grained changes in subtle movement distance and speed. Grooming behaviors were classified into distinct components, including stroking, bilateral face washing, and body licking. Behavioral outcomes were strongly sex dependent. Female juvenile mice born to stressed dams showed increased repetitive grooming and impaired social novelty recognition, whereas male offspring exhibited a different behavioral profile.
Potential Roles of Oxytocin and Microglia
Based on our previous studies4, 5, stressed dams exhibited reduced microglial oxytocin levels, accompanied by decreased expression of the oxytocin receptor in the prefrontal cortex of female juvenile offspring. We also observed reduced Bdnf expression in females, suggesting disrupted interactions between social signaling pathways and synaptic plasticity. Together, these findings provide a plausible biological framework for understanding why maternal perinatal depression may differentially shift neurodevelopmental vulnerability in girls and boys.
What we hope this work changes
Clinically, our findings reinforce an already important message: screening for and supporting maternal mental health during pregnancy and the postpartum period matters. Such support should not only be framed in terms of “preventing autism,” but also as a means of reducing developmental vulnerability or shaping early developmental trajectories. Scientifically, we hope this work encourages more explicit and thoughtful consideration of sex differences in experimental and epidemiological studies of neurodevelopmental disorders. Sex is often treated merely as a covariate to be adjusted away; here, it emerged as a central biological signal that uncovered a distinct pattern of vulnerability and offered new insight into the pathogenesis of autism in females.
Research News (Tohoku University):
Maternal Perinatal Depression May Increase the Risk of Autistic-related Traits in Girls
https://www.tohoku.ac.jp/en/press/maternal_depression_may_increase_risk_of_autism_in_girls.html
References
- Hull L, Petrides KV, Allison C, Smith P, Baron-Cohen S, Lai MC et al. "Putting on My Best Normal": Social Camouflaging in Adults with Autism Spectrum Conditions. J Autism Dev Disord 2017; 47(8): 2519-2534.
- Bhat RS, Alonazi M, Al-Daihan S, El-Ansary A. Prenatal SSRI Exposure Increases the Risk of Autism in Rodents via Aggravated Oxidative Stress and Neurochemical Changes in the Brain. Metabolites 2023; 13(2).
- Mathis A, Mamidanna P, Cury KM, Abe T, Murthy VN, Mathis MW et al. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci 2018; 21(9): 1281-1289.
- Maejima Y, Yokota S, Ono T, Yu Z, Yamachi M, Hidema S et al. Identification of oxytocin expression in human and murine microglia. Prog Neuropsychopharmacol Biol Psychiatry 2022; 119: 110600.
- Sakai M, Yu Z, Hirayama R, Nakasato M, Kikuchi Y, Ono C et al. Deficient Autophagy in Microglia Aggravates Repeated Social Defeat Stress-Induced Social Avoidance. Neural Plast 2022; 2022: 7503553.