Abstract
Background
Few studies have comprehensively examined, within the same population, the roles of both prenatal and postnatal factors in the association between family socioeconomic status (SES) and early infant neurodevelopment.
Methods
Four thousand seven hundred ninety-one singleton live births from November 1, 2017, to December 31, 2022, were included. Family SES was classified as low or high using latent class analysis of family income, parental education, and occupation. Prenatal maternal psychological status, pregnancy complications, and birth outcomes were classified as favorable or unfavorable. Health-related parenting practices was scored based on breastfeeding duration (≥ 6 months), sleep duration (12–16 h/day), outdoor activity (≥ 0.5 h/day), and screen time (none), with each factor scored as 0 or 1 (higher scores indicating healthier parenting practices). Infant neurodevelopment at 1 year was assessed using Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) screening test, across five domains: cognition, receptive communication, expressive communication, fine motor, and gross motor. Each domain was categorized as either competent or ‘emerging or at risk’.
Results
Low family SES was associated with a higher risk of being classified as ‘emerging or at risk’ in receptive communication after adjusting for covariates (risk ratio [RR], 1.42; 95% confidence interval [CI], 1.24-1.62; P<.001), with health-related parenting practices partially mediating this association (mediated proportion, 6.3% (0.9% to 14.1%)).In infants from high-SES families, 0-2 health-related parenting practices scores exerted a more pronounced effect, accounting for nearly one-third of cases (attributable risk percentage, 30.97%). Joint analysis showed infants from low-SES families with 0-2 health-related parenting practices scores had the highest risk compared with infants from high-SES families with 3-4 health-related parenting practices scores (RR, 1.73; 95%CI, 1.41-2.12; P<.001).
Conclusions
Health-related parenting practices partially mediated the relationship between family SES and early infant neurodevelopment, suggesting that while improving health-related parenting practices may reduce disparities, additional strategies are required to address the elevated risk of receptive communication delays among infants from low-SES families. Additionally, unhealthy parenting practices were associated with a higher risk of neurodevelopmental delay even in high-SES families, underscoring the universal importance of early parenting interventions.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12884-026-09360-2.
Keywords: Family SES, Health-related parenting practices, Bayley-III screening test, Receptive communication
Background
From conception through early childhood, the developing brain is highly susceptible to influence from the family and social environment [1]. Socioeconomic status (SES) encompasses more than a family’s economic resources; it also reflects social standing, prestige, and the broader environmental context. Differences in SES-including disparities in income, education, occupation and wealth- exert a substantial impact on children’s neurodevelopment [2, 3]. These SES-related inequalities constitute major barriers to equitable infant development.
Previous researches have demonstrated associations between SES and both structural and functional brain development in infants, using developmental assessments and neuroimaging [4–7]. For instance, lower SES has been linked to reduced gray matter volume in key brain regions, particularly in the frontal and occipital lobes [8]. Although SES is a multifaceted construct encompassing income, education, occupation, and overall wealth [9, 10], many studies assessing its relationship with neurodevelopment capture only limited aspects of this complexity, thereby providing an incomplete picture [8].
SES is recognized as a critical social determinant of child development [11]. Much of the prior literature has focused on SES-related influences during childhood and adolescence [12, 13], and has proposed explanatory frameworks such as the parental investment model and the parental stress model [11, 14]. However, evidence suggests that SES begins to affect neurodevelopment as early as infancy, underscoring the importance of timely early interventions [8]. Yet, few studies have comprehensively examined, within the same population, the roles of both prenatal and postnatal factors in the association between family SES and infant neurodevelopment within the first 12 months of life. Accumulating evidence has documented that multiple modifiable prenatal and postnatal factors may lie on the pathway between family SES and early infant neurodevelopment. Prenatal maternal mental health problems, such as anxiety and depression, are more prevalent among women with disadvantaged socioeconomic backgrounds [15, 16] and have been recognized as critical predictors of impaired fetal brain development and subsequent neurodevelopmental performance in offspring [17, 18]. Similarly, women from low-SES backgrounds have higher risks of gestational hypertension, preeclampsia, and gestational diabetes [19]. These complications have, in turn, been consistently linked to adverse neurodevelopmental outcomes in offspring [20]. Suboptimal birth outcomes (e.g., preterm birth, low birth weight) are strong predictors of later neurodevelopmental impairments [21, 22]. Evidence suggests that socioeconomic disparities contribute to variations in birth outcomes [23], which may also be shaped by maternal pregnancy complications [24] and prenatal psychological status [25]. Beyond pregnancy, resource constraints in low-SES families often extend into the postpartum period, manifesting as less supportive caregiving environments, suboptimal parenting practices, and lower-quality parent-child interactions. These conditions can affect infant cognition, language, and motor development, thereby disrupting early neurodevelopment [26, 27]. Moreover, most existing studies on SES-related developmental disparities have been conducted in Western populations. Therefore, we investigated these associations in a Chinese population to inform culturally appropriate public health interventions.
In the present study, these four factors—prenatal maternal psychological status, pregnancy complications, birth outcomes, and health-related parenting practices—were selected as candidate variables based on existing epidemiological and theoretical evidence, and the present study conducted a secondary analysis of data from the Jiangsu Birth Cohort (JBC) to comprehensively investigate the complex relationships between family SES, candidate variables, and infant neurodevelopmental outcomes.
By identifying mechanisms through which social disadvantage contributes to disparities in early neurodevelopment, this study provides empirical evidence for understanding how socioeconomic inequalities shape child health. The findings are expected to inform targeted interventions and policies aimed to mitigating neurodevelopmental risks among infants from low-SES families. Furthermore, by highlighting modifiable mediators in the SES-neurodevelopmental pathway, this study offers both scientific and practical value for promoting health equity and supporting sustainable social development.
Methods
Study design and population
This study was based on the JBC, a prospective longitudinal birth cohort in China that enrolled couples who conceived either spontaneously or through assisted reproductive technology(ART) [28]. Participants were recruited from three institutions: the Women’s Hospital of Nanjing Medical University, Nanjing Medical University Affiliated Suzhou Hospital, and Nanjing Medical University Affiliated Changzhou Maternal and Child Health Hospital.
To minimize the potential influence of multiple pregnancies on neurodevelopment, and given that neurodevelopmental delays are more common in multiple births than in singletons [29], only singleton live births were included. Accordingly, the present analysis involved 9,021 singleton families with infants born between November 1, 2017, and December 31, 2022.
Screening was performed according to the following criteria (eFigure1): (1) infants who did not undergo or complete the Bayley-III Screening Test (n = 2,367); (2) infants who completed the Bayley-III Screening Test exceeded 12.5 months of age (n = 1,816); And (3) families with missing socioeconomic information (n = 47). After exclusions, 4,791 families were included, of whom 2,857 conceived spontaneously, and 1,934 conceived through ART.
Assessment of family SES
In the JBC study, couples who conceived through ART were enrolled before treatment, whereas those with spontaneous conception were enrolled during early pregnancy (8–14 weeks of gestation). During enrollment, both parents completed standardized questionnaires that collected key family information. Family SES was then assessed using latent class analysis (LCA) based on five indicators: family income, paternal occupation, paternal education, maternal occupation, and maternal education. Each variable was categorized with consideration of the existing grouping in the JBC study, interpretability, and subgroup sample size. Annual family income was classified as ≤ 100,000 CNY, 100,000-200,000 CNY, 200,000-300,000 CNY, and > 300,000 CNY. Education level was categorized into three groups: high school or below, college or polytechnic, and postgraduate or higher. Occupation was classified according to the International Socio-Economic Index of Occupational Status (ISEI-08) [30] and adapted to China’s domestic context, resulting in two groups: high-prestige and low-prestige occupations. LCA, which identifies unobserved subgroups based on multiple categorical indicators [10, 31], was conducted using the ‘poLCA’ package in R software [32]. Two latent classes were identified, representing high and low family SES according to the item-response probabilities (eTable1). The two-class solution model demonstrated good classification accuracy, with mean posterior probabilities for each class ≥ 0.90. In addition, the proportion of participants in each latent class exceeded 20%, indicating adequate class size and statistical stability. Therefore, the two-class model was considered reliable and appropriate for subsequent analyses.
Assessment of prenatal maternal psychological status, pregnancy complications, offspring birth outcomes, health-related parenting practices and other covariates
In the JBC, detailed clinical information [28], including physical examination, clinical tests, and pregnancy complications (e.g., gestational diabetes mellitus and hypertensive disorder of pregnancy) was extracted from electronic medical records (EMRs). Maternal psychological status during pregnancy was assessed using the 20-item Centre for Epidemiologic Studies Depression Scale (CES-D), the 10-item Perceived Stress Scale (PSS-10), and the 20-item Self-rating Anxiety Scale (SAS). In the present study, internal consistency reliability of the three psychological health scales was evaluated using Cronbach’s α coefficient. In the first trimester, Cronbach’s α values were 0.90 for the CES-D, 0.78 for the PSS, and 0.78 for the SAS. In the second trimester, the corresponding α coefficients were 0.90, 0.79, and 0.74, respectively. In the third trimester, Cronbach’s α values were 0.91, 0.81, and 0.79. Overall, all scales demonstrated good internal consistency and reliability across the three trimesters. At delivery, information on pregnancy complications and birth outcomes (e.g., premature delivery and birthweight) was obtained from EMRs. During the postnatal period, mothers completed a structured interview that collected information on infants parenting practices (e.g., feeding and sleep patterns).
Based on these data, prenatal maternal psychological status, pregnancy complications, and offspring birth outcomes were defined as binary variables (favorable or unfavorable), while health-related parenting practices was treated as an ordinal variable. Specifically, maternal psychological status was assessed in the first (8–14 weeks), second (22–26 weeks), and third (30–34 weeks) trimester, including measures of anxiety, depression, and stress. Pregnancy complications were defined as the presence of gestational diabetes mellitus or hypertensive disorders of pregnancy. Offspring birth outcomes include preterm birth and birth weight. Each factor was classified as favorable only if all components were within normal or healthy ranges; otherwise, it was classified as unfavorable. Health-related parenting practices included breastfeeding duration, sleep duration, outdoor activity, and screen time. For each factor, a score of 1 was assigned for a healthy level and 0 for an unhealthy level. resulting in a total score ranging from 0 to 4, with higher scores indicating healthier health-related parenting practices. Detailed definitions are provided in Table 1.
Table 1.
Definitions of potential risk factors
| Variables | Description | Definition | |
|---|---|---|---|
| Prenatal maternal psychological status | Prenatal anxiety (First trimester/Second trimester/Third trimester) | Assessed by validated scales: threshold ≥ 50 (abnormal)、<50 (normal) |
Defined as a binary variable: 1: Normal status in all three trimesters of pregnancy 0: Abnormal status in at least one trimester of pregnancy |
| Prenatal depression (First trimester/Second trimester/Third trimester) | Assessed by validated scales: threshold ≥ 16 (abnormal)、<16 (normal) | ||
| Prenatal stress (First trimester/Second trimester/Third trimester) | Assessed by validated scales: threshold ≥ 15 (abnormal)、<15 (normal) | ||
| Pregnancy complications | Diabetes in Pregnancy (Gestational Diabetes Mellitus、Pregestational Diabetes Mellitus) | Extracted by EMRs: Gestational diabetes or pre-gestational diabetes (presence/absence) |
Defined as a binary variable: 1: Absence of complications throughout pregnancy 0: Presence of at least one complication during pregnancy |
| Hypertensive Disorders of Pregnancy (Gestational Hypertension、Preeclampsia、Chronic Hypertension) | Extracted by EMRs: Gestational hypertension, preeclampsia, chronic hypertension (presence/absence) | ||
| Offspring birth outcomes | Gestational age | Extracted by EMRs: Preterm (GA < 37 weeks)、term (GA ≥ 37 weeks) |
Defined as a binary variable 1: Term infant (≥ 37 weeks) with normal birth weight (2500–4000 g) 0: Infant either preterm (< 37 weeks) or with abnormal birth weight (< 2500 g or > 4000 g) |
| Birth Weight | Extracted by EMRs: Normal birth weight: 2500–4000 g (yes/no) | ||
| Health-related parenting practices | Duration of breastfeeding | Extracted by Offspring follow-up questionnaire: ≥6 months (met/not met) |
Defined as an ordinal variable 0–4: The health-related parenting practices score was the sum of the points and ranged between 0 and 4, with higher scores indicating healthier health-related parenting practices. Each healthy practice was assigned 1 score. |
| Duration of sleep | Extracted by Offspring follow-up questionnaire: 12–16 h/day (met/not met) | ||
| Duration of outdoor activity | Extracted by Offspring follow-up questionnaire: ≥0.5 h/day (met/not met) | ||
| Duration of screen time | Extracted by Offspring follow-up questionnaire: Zero screen time (met/not met) | ||
Prenatal maternal psychological status during each trimester was assessed using the 20-item Centre for Epidemiologic Studies Depression Scale (CES-D), the 10-item Perceived Stress Scale (PSS-10), and the 20-item Self-Rating Anxiety Scale (SAS). Pregnancy complications and offspring birth outcomes were based on diagnoses recorded in the electronic medical records (EMRs). Health-related parenting practices, including duration of breastfeeding, duration of sleep, duration of outdoor activity, and duration of screen time, was categorized according to WHO recommendations
Other covariates were also collected via standardized questionnaires administered in person or by telephone, as well as through perinatal data extracted from the EMRs at delivery. Gestational age was determined from the last menstrual period (LMP) for spontaneous pregnancies, and for ART pregnancies, by calculating the embryo transfer date (subtracting 17 days for fresh or frozen cleavage-stage embryos and 20 days for fresh or frozen blastocyst transfer).
Assessment of outcome
At 1 year of age, infants were invited back to the hospital for comprehensive physical and neurodevelopmental assessments conducted by trained pediatricians or occupational therapists in the presence of a primary caregiver. Neurodevelopment was evaluated using the Bayley-III Screening Test, a shortened version of the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) [33]. This tool assesses five neurodevelopment domains: cognition, receptive communication, expressive communication, fine motor skills, and gross motor skills. Each domain yields a score (eTable2), which classifies infants into three diagnostic categories based on age: competent (normal development), emerging (borderline development), or at-risk (developmental delay). Due to the small number of infants in the “at-risk” category, the “at-risk” and “emerging” groups were combined as ‘emerging or at risk’ for analysis [34].
To ensure standardization and reliability, a developmental neuropsychologist trained examiners in the administration and scoring of the Bayley-III Screening Test. With guardian consent, all examinations were recorded. The Bayley-III Screening Test has demonstrated good reliability [35], with coefficients ranging from 0.82 to 0.88 across the five domains, and moderate efficacy (41.8%–92.1%) in identifying infants requiring further developmental evaluation [36].
During the neurodevelopmental assessments, examiners were blinded to the family socioeconomic status.
Statistical analysis
The relative risks (RRs) and 95% confidence intervals (CIs) for associations between family SES and infant neurodevelopment at 1 year of age were estimated using a generalized linear model with a robust Poisson distribution [37]. Model 1 was unadjusted, and Model 2 adjusted for maternal age at delivery (years), parity, study center, mode of conception, mode of delivery, infant sex, and gestational age at delivery. To account for multiple comparisons across the five neurodevelopmental domains, the false discovery rate (FDR) was controlled using the Benjamini-Hochberg procedure. Subgroup analyses were conducted by infant sex (boy, girl), study center (Nanjing, Suzhou, Changzhou), mode of conception (ART, spontaneous), and mode of delivery (cesarean, vaginal).
Mediation analyses were conducted using a regression-based methodology for causal mediation analysis. Direct counterfactual imputation was applied to estimate effects, with standard errors calculated through bootstrapping. Log-linear regression was applied for outcome models. Mediators defined as binary variables were modeled with logistic regressions, while health-related parenting practices, treated as an ordinal variable, was modeled with multinomial regression.
Family SES may influence infant neurodevelopment through multiple pathways, including prenatal maternal psychological status, pregnancy complications, offspring birth outcomes, and health-related parenting practices. To examine these potential pathways, we additionally applied structural equation modeling (SEM) based on seven predefined pathways (eTable3). SEM offers methodological advantages by simultaneously testing multiple parallel mediating pathways, accounting for measurement error, and modeling correlations among mediators. This approach provides more robust estimates of the mechanisms linking family SES to infant neurodevelopment than separate regression models. Given the inclusion of binary and ordinal variables, estimation was performed with diagonally weighted least squares (DWLS), and standard errors were obtained via bootstrapping.
For subsequent exploratory analyses, the health-related parenting practices was dichotomized into ‘0–2 scores’ and ‘3–4 scores’ groups due to the extremely low proportion of participants scoring 0 (1,0.0%) and 1(128,2.7%), the need to simplify complex effect interpretation, and to reduce analytical complexity and multiple comparison issues.
Potential interactions were evaluated by including product terms between family SES and covariates. Multiplicative interactions were assessed using the relative ratios (RRs) and 95% CIs for the product terms, while additive interactions were assessed with the relative excess risk due to interaction (RERI) and 95% CIs.
Stratified analyses were further conducted by family SES latent class to investigate associations of covariates with outcomes in different socioeconomic subgroups. Attributable risk percent (AR%) was calculated to quantify the proportion of risk among the exposed population attributable to the exposure.
Joint associations were examined by classifying participants according to family SES and covariates, and estimating the prevalence of neurodevelopmental delay across groups, using infants with high family SES and favorable conditions as the reference group.
Several sensitivity analyses were conducted. First, because rural residents and preterm infants accounted for only small proportions of the sample (4.4% and 4.6%, respectively), limiting statistical stability and representativeness, and to avoid potential bias from missing data imputation, all analyses were repeated among urban participants, full-term infants, and the complete cases (excluding missing data). Second, a weighted health-related parenting practices score was constructed to mitigate potential bias from the simple additive scoring method, which assumes equal associations of all components with the outcome. To obtain a weighted health-related parenting practices score, a regression-based weighting method was applied. The standardized partial regression coefficients (β) from this model were used as weights. Each indicator was multiplied by its corresponding β coefficient, and the products were summed to generate the weighted score.
All analyses were performed using R version 4.3.2, with a two-sided significance level of 0.05.
Missing data
With the exception of variables related to prenatal maternal psychological status, which had missing rates ranging from 13.8% to 20.4%, the proportion of missing data for the remaining variables was generally low, with most below 5%. To address missing values in the dataset, multiple imputation using Random Forest [38, 39] was applied to variables with incomplete data, including covariates and potential mediators. This approach leverages the predictive capability of Random Forest to estimate and replace missing values, ensuring the completeness and reliability of the dataset for subsequent analyses.
Ethics statement
All methods were carried out in accordance with relevant guidelines and regulations under ethics approval and consent to participate.
Results
Baseline characteristics of the study population
Table 2 summarizes the baseline characteristics of the pregnant families and their infants included in this study. A total of 4,791 infants were analyzed, comprising 2,935 from high-SES families and 1,856 from low-SES families. Compared with women from high-SES families, women from low-SES families had higher proportions of pre-pregnancy overweight or obesity, rural residence, multiparity, and conception through ART, as well as a higher incidence of hypertensive disorders of pregnancy. In addition, regarding prenatal maternal psychological status, women from low-SES families were more likely to experience stress throughout pregnancy (first, second, and third trimesters) and had a significantly higher prevalence of anxiety during the third trimester. At delivery, offspring from high‑SES families have a significantly higher rate of vaginal delivery and a slightly longer mean gestational age. Postnatally, infants from high-SES families experienced more favorable health-related parenting practices, including longer breastfeeding duration and less screen exposure.
Table 2.
Demographics of the participants according to family SES
| Characteristics | Total | High Family SES | Low Family SES | P Value |
|---|---|---|---|---|
| Parental characteristics | 4791 | 2935 | 1856 | |
| Family income (CNY) | < 0.001 | |||
| ≤100,000 | 1035 (21.6) | 233 (7.9) | 802 (43.2) | |
| 100,000-200,000 | 2052 (42.8) | 1382 (47.1) | 670 (36.1) | |
| 200,000-300,000 | 826 (17.2) | 631 (21.5) | 195 (10.5) | |
| >300,000 | 878 (18.3) | 689 (23.5) | 189 (10.2) | |
| Maternal education level | < 0.001 | |||
| high school or below | 826 (17.2) | 2 (0.1) | 824 (44.4) | |
| college or polytechnic graduate | 3420 (71.4) | 2388 (81.4) | 1032 (55.6) | |
| postgraduate or higher | 545 (11.4) | 545 (18.6) | 0 (0.0) | |
| Paternal education level | < 0.001 | |||
| high school or below | 736 (15.4) | 3 (0.1) | 733 (39.5) | |
| college or polytechnic graduate | 3453 (72.1) | 2331 (79.4) | 1122 (60.5) | |
| postgraduate or higher | 602 (12.6) | 601 (20.5) | 1 (0.1) | |
| Maternal occupation | < 0.001 | |||
| Low-prestige occupations | 1984 (41.4) | 492 (16.8) | 1492 (80.4) | |
| High-prestige occupations | 2807 (58.6) | 2443 (83.2) | 364 (19.6) | |
| Paternal occupation | < 0.001 | |||
| Low-prestige occupations | 1824 (38.1) | 444 (15.1) | 1380 (74.4) | |
| High-prestige occupations | 2967 (61.9) | 2491 (84.9) | 476 (25.6) | |
| Residential area | < 0.001 | |||
| Urban | 4580 (95.6) | 2868 (97.7) | 1712 (92.2) | |
| Rural | 210 (4.4) | 66 (2.2) | 144 (7.8) | |
| Missing | 1 (0.0) | 1 (0.0) | 0 (0.0) | |
| Center | < 0.001 | |||
| NanJing | 1896 (39.6) | 1265 (43.1) | 631 (34.0) | |
| ChangZhou | 1070 (22.3) | 520 (17.7) | 550 (29.6) | |
| SuZhou | 1825 (38.1) | 1150 (39.2) | 675 (36.4) | |
| Age at delivery(y), mean(SD) | 30.63 (3.89) | 30.63 (3.58) | 30.53 (4.35) | 0.392 |
| Pre-pregnancy BMI, kg/m2(%) | < 0.001 | |||
| <18.5 | 511 (10.7) | 339 (11.6) | 172 (9.3) | |
| 18.5 ~ 23.9 | 3227 (67.4) | 2063 (70.3) | 1164 (62.7) | |
| 24 ~ 27.9 | 799 (16.7) | 420 (14.3) | 379 (20.4) | |
| ≥28 | 252 (5.3) | 112 (3.8) | 140 (7.5) | |
| Missing | 2 (0.0) | 1 (0.0) | 1 (0.1) | |
| Parity | < 0.001 | |||
| Nulliparous | 3581 (74.7) | 2287 (77.9) | 1294 (69.7) | |
| Multiparous | 979 (20.4) | 556 (18.9) | 423 (22.8) | |
| Missing | 231 (4.8) | 92 (3.1) | 139 (7.5) | |
| Smoking during pregnancy | 15 (0.3) | 6 (0.2) | 9 (0.5) | 0.153 |
| Missing | 1 (0.0) | 0 (0.0) | 1 (0.1) | |
| Drinking during pregnancy | 36 (0.8) | 17 (0.6) | 19 (1.0) | 0.118 |
| Missing | 2 (0.0) | 1 (0.0) | 1 (0.1) | |
| Mode of conception | < 0.001 | |||
| ART | 1934 (40.4) | 982 (33.5) | 952 (51.3) | |
| Spontaneous | 2857 (59.6) | 1953 (66.5) | 904 (48.7) | |
| Pregnancy complications | ||||
| Diabetes mellitus a | 1425 (29.7) | 853 (29.1) | 572 (30.8) | 0.160 |
| Missing | 15 (0.3) | 2 (0.1) | 13 (0.7) | |
| Hypertension b | 327 (6.8) | 154 (5.2) | 173 (9.3) | < 0.001 |
| Missing | 32 (0.7) | 11 (0.4) | 21 (1.1) | |
| First trimester: Prenatal anxiety | 360 (7.5) | 205 (7.0) | 155 (8.4) | 0.124 |
| Missing | 867 (18.1) | 547 (18.6) | 320 (17.2) | |
| First trimester: Prenatal depression | 1402 (29.3) | 882 (30.1) | 520 (28.0) | 0.053 |
| Missing | 867 (18.1) | 547 (18.6) | 320 (17.2) | |
| First trimester: Prenatal stress | 1509 (31.5) | 859 (29.3) | 650 (35.0) | < 0.001 |
| Missing | 868 (18.1) | 548 (18.7) | 320 (17.2) | |
| Second trimester: Prenatal anxiety | 211 (4.4) | 120 (4.1) | 91 (4.9) | 0.097 |
| Missing | 664 (13.9) | 356 (12.1) | 308 (16.6) | |
| Second trimester: Prenatal depression | 897 (18.7) | 556 (18.9) | 341 (18.4) | 0.735 |
| Missing | 663 (13.8) | 354 (12.1) | 309 (16.6) | |
| Second trimester: Prenatal stress | 1105 (23.1) | 606 (20.6) | 499 (26.9) | < 0.001 |
| Missing | 664 (13.9) | 355 (12.1) | 309 (16.6) | |
| Third trimester: Prenatal anxiety | 280 (5.8) | 154 (5.2) | 126 (6.8) | 0.009 |
| Missing | 977 (20.4) | 554 (18.9) | 423 (22.8) | |
| Third trimester: Prenatal depression | 878 (18.3) | 555 (18.9) | 323 (17.4) | 0.620 |
| Missing | 978 (20.4) | 554 (18.9) | 424 (22.8) | |
| Third trimester: Prenatal stress | 981 (20.5) | 533 (18.2) | 448 (24.1) | < 0.001 |
| Missing | 977 (20.4) | 553 (18.8) | 424 (22.8) | |
| Infant characteristics | 4791 | 2935 | 1856 | |
| Mode of delivery | < 0.001 | |||
| Vaginal delivery | 2446 (51.1) | 1584 (54.0) | 862 (46.4) | |
| Cesarean delivery | 2319 (48.4) | 1336 (45.5) | 983 (53.0) | |
| Missing | 26 (0.5) | 15 (0.5) | 11 (0.6) | |
| Sex | 0.659 | |||
| Boys | 2545 (53.1) | 1567 (53.4) | 978 (52.7) | |
| Girls | 2246 (46.9) | 1368 (46.6) | 878 (47.3) | |
| Gestational age(wk), mean(SD) | 38.91 (1.42) | 38.97 (1.37) | 38.82 (1.47) | < 0.001 |
| PTB | 0.265 | |||
| Preterm | 221 (4.6) | 127 (4.3) | 94 (5.1) | |
| Term | 4570 (95.4) | 2808 (95.7) | 1762 (94.9) | |
| Birthweight | 0.112 | |||
| 2500 ~ 4000 g | 4306 (89.9) | 2653 (90.4) | 1653 (89.1) | |
| <2500 g and > 4000 g | 456 (9.5) | 263 (9.0) | 193 (10.4) | |
| Missing | 29 (0.6) | 19 (0.6) | 10 (0.5) | |
| Duration of breastfeeding (≥ 6 months) | 3938 (82.2) | 2485 (84.7) | 1453 (78.3) | < 0.001 |
| Missing | 10 (0.2) | 6 (0.2) | 4 (0.2) | |
| Duration of sleep, 6 m (12–16 h/day) | 3896(81.3) | 2418 (82.4) | 1478 (79.6) | 0.050 |
| Missing | 323 (6.7) | 187 (6.4) | 136 (7.3) | |
| Duration of sleep, 12 m (12–16 h/day) | 3877 (80.9) | 2392 (81.5) | 1485 (80.0) | 0.113 |
| Missing | 79 (1.6) | 53 (1.8) | 26 (1.4) | |
| Duration of outdoor activity (≥ 0.5 h/day) | 4558 (95.1) | 2790 (95.1) | 1768 (95.3) | 0.753 |
| Missing | 82 (1.7) | 55 (1.9) | 27 (1.5) | |
| Duration of screen time (no screen time) | 2785 (58.1) | 1805 (61.5) | 980 (52.8) | < 0.001 |
| Missing | 82 (1.7) | 55 (1.9) | 27 (1.5) | |
| Baley-III Screening Test | 4791 | 2935 | 1856 | |
| Cognition: Classified as ‘emerging or at risk’ | 521 (10.9) | 312 (10.6) | 209 (11.3) | 0.525 |
| Receptive communication: Classified as ‘emerging or at risk’ | 768 (16.0) | 412 (14.0) | 356 (19.2) | < 0.001 |
| Expressive communication: Classified as ‘emerging or at risk’ | 238 (5.0) | 140 (4.8) | 98 (5.3) | 0.469 |
| Fine motor: Classified as ‘emerging or at risk’ | 140 (2.9) | 80 (2.7) | 60 (3.2) | 0.354 |
| Gross motor: Classified as ‘emerging or at risk’ | 451 (9.4) | 266 (9.1) | 185 (10.0) | 0.320 |
Data are presented as number (percentage) or mean (SD)
y year, SD standard deviation, CNY Chinese yuan, ART assisted reproduction technology
a Diabetes mellitus includes chronic and gestational diabetes mellitus
b Hypertension includes chronic and gestational hypertension and preeclampsia
Table 2 also presents the rate of being classified as ‘emerging or at risk’ across the five neurodevelopmental domains. The rate was significantly higher in low‑SES than in high‑SES families for receptive communication. Although the differences for the other four domains were not statistically significant, infants in low-SES families showed a consistently higher rate in each of these domains compared to infants in high‑SES families. Similarly, eTable4 shows that, compared with infants from low-SES families, those from high-SES families had significantly higher raw scores in receptive communication (P < .001, SMD = 0.152). eTable5 presents a comparison of baseline characteristics between included, excluded, and overall samples.
Associations between family SES and infant neurodevelopment
As shown in Table 3, in the crude model, infants from low-SES families had an increased risk of being classified as ‘emerging or at risk’ in receptive communication. After adjusting for maternal age at delivery (years), parity, study center, mode of conception, mode of delivery, infant sex and gestational age at delivery, the association remained significant (RR, 1.42; 95%CI, 1.24–1.62; P<.001; q<0.001).
Table 3.
Associations between family SES and the risk of infants being classified as ‘emerging or at risk’ in the five domains
| Variable | High Family SES N (%) |
Low Family SES N (%) |
Crude RR (95% CI) |
P Value | Adjusted RR (95% CI) a |
P Value | FDR b |
|---|---|---|---|---|---|---|---|
| Cognition domain | 312/2935 (10.6) | 209/1856 (11.3) | 1.06 (0.90,1.25) | 0.494 | 1.19 (1.01,1.41) | 0.038 | 0.096 |
| Receptive communication domain | 412/2935 (14.0) | 356/1856 (19.2) | 1.37 (1.20,1.56) | < 0.001 | 1.42 (1.24,1.62) | < 0.001 | < 0.001 |
| Expressive communication domain | 140/2935 (4.8) | 98/1856 (5.3) | 1.11 (0.86,1.42) | 0.429 | 1.17 (0.91,1.51) | 0.220 | 0.274 |
| Fine motor domain | 80/2935 (2.7) | 60/1856 (3.2) | 1.19 (0.85,1.65) | 0.310 | 1.30 (0.93,1.80) | 0.122 | 0.204 |
| Gross motor domain | 266/2935 (9.1) | 185/1856 (10.0) | 1.10 (0.92,1.31) | 0.296 | 1.01 (0.84,1.21) | 0.954 | 0.954 |
CI confidence interval, RR risk ratio
a RRs, 95%CI, and P values were from a Generalized linear model with robust Poisson distribution, adjusted for maternal age at delivery, parity, study center, mode of conception, mode of delivery, infant sex and gestational weeks
b FDR, p-values were adjusted for multiple comparisons using the Benjamini-Hochberg false discovery rate (FDR) correction with a threshold of 0.05
Subgroup analyses by infant sex, study center, mode of conception, and mode of delivery (eFigure2) showed no significant heterogeneity, supporting the robustness of the association between family SES and receptive communication at 1 year of age. Sensitivity analyses restricted to urban, full-term, and complete-case populations (eTable6) produced consistent results.
Mediation analysis of health-related parenting practices on associations of family SES with receptive communication delays
The adjusted direct and indirect effects of family SES on receptive communication, accounting for potential confounders, are presented in Fig. 1. Family SES exhibited a significant direct effect, while health-related parenting practices demonstrated a notable indirect effect, mediating 6.3% of the total association (Fig. 1A). Other proposed factors did not demonstrate significant indirect effects (eFigure3). This mediation effect of health-related parenting practices was consistent in all sensitivity analyses. A weighted health-related parenting practices score was further tested to address the limitation of the additive scoring method, yielding similar results (eTable7).
Fig. 1.
Mediation analyses. A Causal Mediation analysis. PNDE: the average effect of the exposure through all other mechanisms, excluding the selected mediator; TNIE: the average effect of the exposure through the mediator; PM: Proportion of effect mediated. Adjusted for maternal age at delivery, parity, study center and mode of conception. B Diagram of multiple mediation analysis using SEM. SEM = structural equation modeling; DWLS = diagonally weighted least squares; Adjusted for maternal age at delivery, parity, study center and mode of conception; Model was estimated with DWLS and theta parameterization. Bootstrap SEs (N=1,000 resamples) address non-normality in ordered categorical variables
As demonstrated in Fig. 1B, structural equation modeling (SEM) based on the prespecified pathways identified only two significant paths: (1) family SES → receptive communication, and (2) family SES → health-related parenting practices → receptive communication, consistent with the mediation analysis.
Interaction and joint analysis of health-related parenting practices and family SES with receptive communication delays
No significant multiplicative or additive interaction between family SES and health-related parenting practices on receptive communication delay was observed (eTable8).
Table 4 indicates that among high-SES families, infants with 0–2 health-related parenting practices scores had a significantly higher risk of being classified as ‘emerging or at risk’ in receptive communication (RR, 1.48; 95% CI, 1.21–1.82; P<.001), with an attributable risk percent (AR%) of 30.97%. This association was not significant in low-SES families.
Table 4.
Associations of health-related parenting practices with the risk of infants being classified as ‘emerging or at risk’ in the receptive communication domain by family SES
| Crude RR (95% CI) |
P Value | Adjusted RR (95% CI) a |
P Value | AR% b | |
|---|---|---|---|---|---|
| High family SES | |||||
| 3–4 scores | 1 (reference) | 1 (reference) | 30.97 (15.05,43.62) | ||
| 0–2 scores | 1.44 (1.18,1.78) | < 0.001 | 1.48 (1.21,1.82) | < 0.001 | |
| Low family SES | |||||
| 3–4 scores | 1 (reference) | 1 (reference) | 12.89 (-7.14,28.83) | ||
| 0–2 scores | 1.17 (0.95,1.44) | 0.138 | 1.18 (0.96,1.45) | 0.116 | |
CI confidence interval, RR risk ratio, AR% attributable risk percentage
a RRs, 95%CI, and P values were from a Generalized linear model with robust Poisson distribution, adjusted for maternal age at delivery, parity, study center, mode of conception, mode of delivery, infant sex and gestational weeks
b AR% was defined as the proportion of risk among the exposed population that can be attributed to the exposure
Figure 2 shows the joint association of family SES and health-related parenting practices with receptive communication outcomes. Compared with the reference group (high family SES, and 3–4 health-related parenting practices scores), infants with low family SES and 0–2 health-related parenting practices scores had the highest risk of being classified as ‘emerging or at risk’ (RR, 1.73; 95%CI, 1.41–2.12; P<.001). The combined exposure to low family SES and suboptimal health-related parenting practices conferred a higher risk than either factor alone.
Fig. 2.
Joint effects of family SES and health-related parenting practices in the receptive communication domain. CI, confidence interval; RR, risk ratio. RRs, 95%CI, and P values were from a Generalized linear model with robust Poisson distribution, adjusted for maternal age at delivery, parity, study center, mode of conception, mode of delivery, infant sex and gestational weeks
Discussion
The present study demonstrated that low family SES was associated with poorer early infant neurodevelopment, particularly in receptive communication, with health-related parenting practices serving as a mediating factor. In addition, in infants from high-SES families, 0–2 health-related parenting practices scores were significantly associated with increased likelihood of being classified as ‘emerging or at risk’ in receptive communication. Infants from low-SES families with 0–2 health-related parenting practices scores exhibited the highest risk of delayed receptive communication.
Receptive communication appeared to be particularly sensitive to socioeconomic disadvantage in the present study. This finding is consistent with previous evidence suggesting that language-related domains are among the earliest developmental functions affected by socioeconomic inequalities [40]. Compared with expressive language, infants in early life rely primarily on passive language input and comprehension. Consequently, receptive communication development may be more vulnerable to differences in the early home environment. In a cohort study [41], infants from low-SES families performed more poorly than those from higher-SES families on the PLS-5 Total Language, Auditory Comprehension, and Expressive Communication. Similarly, by school entry, children from disadvantaged families perform well below their middle class peers on standardized tests of language production and comprehension [42]. Another study using the Bayley-III screening test reported that a relatively large proportion of infants (41%) were classified in the “Emerging” category for receptive communication, whereas the corresponding proportion ranged from only 1% to 12% across other developmental domains [43]. These results seem broadly similar to what we observed. Neuroimaging evidence further supports our findings. Previous studies have reported that lower SES is associated with reduced gray matter volume in key infant brain regions-particularly the frontal, parietal, and occipital lobes [8, 44] which play critical roles in language processing [45].
Our study explored potential pathways linking family SES to infant neurodevelopment. Unlike prior research that primarily examined social and familial environmental factors during childhood or adolescence [46–49], our study focused on the prenatal period-from conception to early childhood-addressing both prenatal and postnatal influences. Importantly, health-related parenting practices emerged as a significant mediator between family SES and early neurodevelopment. Positive health-related parenting practices, including longer breastfeeding duration, consistent sleep routines, outdoor engagement, and limited screen exposure, have been associated with improved cognitive and language outcomes in offspring [50–52]. These practices may support neurodevelopment by promoting high-quality parent–infant interactions during early life, which are critical for healthy brain and behavioral development [53, 54]. Families with lower SES are highly likely to experience persistent financial stress and reduced access to resources, which may make it more difficult to establish and maintain consistent daily health routines. Such challenges may, in turn, contribute to disparities in infant neurodevelopment.
The relatively small mediated proportion suggests that health-related parenting practices represent only one of multiple pathways linking socioeconomic disadvantage to infant neurodevelopment. A review [55] has suggested that childhood SES influences neurodevelopment, particularly in brain systems underlying language and executive function. Evidence from both human and animal studies further indicates that prenatal factors, parent–child interactions, and cognitive stimulation within the home environment may contribute to the association between SES and neural development. Therefore, interventions aimed at fully reducing SES-related inequalities in early neurodevelopment should combine parenting practices with other approaches targeting additional pathways. However, mediation effects through maternal psychological status, pregnancy complications, and birth outcomes were not detected. One possible explanation is that these pathways may be incomplete, as epigenetic mechanisms [56] were not considered. Such factors may influence neurodevelopment indirectly through molecular modifications rather than direct effects. Furthermore, these indicators may lack representativeness, potentially contributing to the absence of significant mediating effects.
Finally, in infants from high-SES families, lower health-related parenting practices scores were associated with an increased likelihood of being classified as ‘emerging or at risk’ in receptive communication. In this subgroup, approximately 30.97% of developmental delay cases could be attributed to unhealthy parenting practices. This finding may reflect that within high-SES families-where material and environmental conditions are relatively uniform-variation in health-related parenting practices becomes a particularly salient determinant of early neurodevelopment.
Strengths and limitations
First, given the limited research incorporating comprehensive measures of SES, this study developed a composite SES index based on family income, paternal occupation, paternal education, maternal occupation, and maternal education. Second, this study comprehensively examined the roles of both prenatal and postnatal factors in the association between family SES and infant neurodevelopment using data from a well-designed large-scale birth cohort, and the presence of significant mediation effects was robustly supported through multiple analytical approaches. Third, a series of sensitivity analyses were conducted to demonstrate the robustness of the findings.
Despite these strengths, several limitations should be acknowledged. First, the health-related parenting practices score was constructed using a simple additive method. Although a sensitivity analysis employing a weighted health-related parenting practices score yielded consistent results, this approach may not fully capture the complex interactions among parenting components, and the derived weights are specific to this study. Second, this study did not further examine potential interactions or moderating effects among the four mediators. Future research could consider the complex relationships among these variables to better elucidate the underlying mechanisms. Third, as most participants were from urban areas, the low-SES group represented relatively disadvantaged families within this context rather than those experiencing extreme poverty. Future studies should investigate populations facing more severe socioeconomic deprivation. Finally, the possibility of residual confounding cannot be entirely excluded, and causal inference should be made with caution given the observational nature of the study.
Conclusions
Infants from low-SES families demonstrated a significantly higher incidence of receptive communication deficits, with health-related parenting practices acting as a mediating factor. Improving parenting practices may therefore help reduce socioeconomic disparities in early child neurodevelopment. Moreover, in high-SES families, lower health-related parenting practices scores were more strongly associated with neurodevelopmental delay in receptive communication, suggesting that even among advantaged families, parenting quality remains a critical determinant. Finally, infants with both low family SES and suboptimal health-related parenting practices exhibited the highest risk of receptive communication delay, underscoring the need for integrated interventions targeting both socioeconomic disadvantage and parenting behaviors.
Supplementary Information
The Additional file is an updated version of the supplementary material. Reason: thousand separators have been added to the numerical values in eFigure 1 to improve readability. All other content remains unchanged. Please replace the original attachment with this version.
Acknowledgements
Our profound gratitude is extended to the devoted women who participated in this study and the dedicated staff who contributed their expertise, thereby facilitating the successful execution of this research.
Abbreviations
- SES
Socioeconomic status
- JBC
Jiangsu birth cohort study
- LCA
Latent class analysis
- SEM
Structural equation modeling
- ART
Assisted reproductive technology
- CES-D
Center for epidemiologic studies depression scale
- PSS
Perceived stress scale
- SAS
Self-rating anxiety scale
- Bayley-III
Bayley Scales of Infant and Toddler Development, Version-III
- LMP
Last menstrual period time
- SMD
Standardized mean difference
- FDR
False discovery rate
- DWLS
Diagonally weighted least squares
- RERI
Relative excess risk due to interaction
- AR%
Attributable risk percentage
- BMI
Body mass index
Authors’ contributions
T.Jiang, Hu and Chen had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis; Concept and design: T.Jiang, Chen; Acquisition and interpretation of data: All authors; Drafting of the manuscript: Chen, T.Jiang; Critical revision of the manuscript for important intellectual content: All authors; Statistical analysis: Chen, Huang, He; Obtained funding: T.Jiang, Lin; Administrative, technical, or material support: Hu, Jin, Ma, Du, Lin, T.Jiang; Supervision: T.Jiang, Du, Lin.
Funding
This study was funded by the National Natural Science Foundation of China (grant number 82373581), General Project of Basic Science (Natural Science) Research in Higher Education Institutions in Jiangsu Province (grant number 23KJB330002).
Data availability
The datasets generated and/or analysed during the current study are not publicly available due to our containing information that could compromise the privacy of research participants but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
All methods were carried out in accordance with relevant guidelines and regulations under ethics approval and consent to participate. All procedures were approved by the institutional review board of Nanjing Medical University, China NJMUIRB (2017) 002. All participants gave their written informed consent at the time of recruitment. Additionally, prior to administering the Bayley-III Screening Test, we secured an additional informed consent specifically for this assessment from each child’s guardian.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Jie Chen, Wanting Huang and Jiayi He contributed equally to this work.
Contributor Information
Zhibin Hu, Email: zhibin_hu@njmu.edu.cn.
Tao Jiang, Email: tao.chiang0923@njmu.edu.cn.
References
- 1.Miguel PM, Pereira LO, Silveira PP, Meaney MJ. Early environmental influences on the development of children’s brain structure and function. Dev Med Child Neurol. 2019;61(10):1127–33. 10.1111/dmcn.14182 . PubMed PMID: 30740660. [DOI] [PubMed] [Google Scholar]
- 2.Tomasi D, Volkow ND. Associations of family income with cognition and brain structure in USA children: prevention implications. Mol Psychiatry. 2021;26(11):6619–29. 10.1038/s41380-021-01130-0 . PubMed PMID: 33990770; PubMed Central PMCID: PMC8590701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Shi R, Chang X, Banaschewski T, Barker GJ, Bokde ALW, Desrivières S, et al. Gene-environment interactions in the influence of maternal education on adolescent neurodevelopment using ABCD study. Sci Adv. 2024;10(46):eadp3751. 10.1126/sciadv.adp3751 . PubMed PMID: 39546599; PubMed Central PMCID: PMC11567010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lu YC, Kapse K, Andersen N, Quistorff J, Lopez C, Fry A, et al. Association Between Socioeconomic Status and In Utero Fetal Brain Development. JAMA Netw Open. 2021;4(3):e213526. 10.1001/jamanetworkopen.2021.3526 . PubMed PMID: 33779746; PubMed Central PMCID: PMC8008281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tella P, Piccolo Lda, Rangel R, Rohde ML, Polanczyk LA, Miguel GV. Socioeconomic diversities and infant development at 6 to 9 months in a poverty area of São Paulo, Brazil. Trends Psychiatry Psychother. 2018;40(3):232–40. 10.1590/2237-6089-2017-0008 . PubMed PMID: 30156646. [DOI] [PubMed] [Google Scholar]
- 6.Ronfani L, Vecchi Brumatti L, Mariuz M, Tognin V, Bin M, Ferluga V, et al. The Complex Interaction between Home Environment, Socioeconomic Status, Maternal IQ and Early Child Neurocognitive Development: A Multivariate Analysis of Data Collected in a Newborn Cohort Study. PLoS ONE. 2015;10(5):e0127052. 10.1371/journal.pone.0127052. PubMed PMID: 25996934; PubMed Central PMCID: PMC4440732. [DOI] [PMC free article] [PubMed]
- 7.Mckinnon K, Galdi P, Blesa-Cábez M, Sullivan G, Vaher K, Corrigan A, et al. Association of Preterm Birth and Socioeconomic Status With Neonatal Brain Structure. JAMA Netw Open. 2023;6(5):e2316067. 10.1001/jamanetworkopen.2023.16067 . PubMed PMID: 37256618; PubMed Central PMCID: PMC10233421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rakesh D, Whittle S, Sheridan MA, McLaughlin KA. Childhood socioeconomic status and the pace of structural neurodevelopment: accelerated, delayed, or simply different? Trends Cogn Sci. 2023;27(9):833–51. 10.1016/j.tics.2023.03.011 . PubMed PMID: 37179140; PubMed Central PMCID: PMC10524122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Quaglia A, Lillini R, Mamo C, Ivaldi E, Vercelli M. Socio-Economic Indicators, Health) Working Group. Socio-economic inequalities: a review of methodological issues and the relationships with cancer survival. Crit Rev Oncol Hematol. 2013;85(3):266–77. 10.1016/j.critrevonc.2012.08. .007 PubMed PMID: 22999326. [DOI] [PubMed] [Google Scholar]
- 10.Zhang YB, Chen C, Pan XF, Guo J, Li Y, Franco OH, et al. Associations of healthy lifestyle and socioeconomic status with mortality and incident cardiovascular disease: two prospective cohort studies. BMJ. 2021;373:n604. 10.1136/bmj.n604 . PubMed PMID: 33853828; PubMed Central PMCID: PMC8044922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bradley RH, Corwyn RF. Socioeconomic status and child development. Annu Rev Psychol. 2002;53:371–99. 10.1146/annurev.psych.53.100901. .135233 PubMed PMID: 11752490. [DOI] [PubMed] [Google Scholar]
- 12.Noble KG, Houston SM, Brito NH, Bartsch H, Kan E, Kuperman JM, et al. Family income, parental education and brain structure in children and adolescents. Nat Neurosci. 2015;18(5):773–8. 10.1038/nn.3983 . PubMed PMID: 25821911; PubMed Central PMCID: PMC4414816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Whittle S, Vijayakumar N, Simmons JG, Dennison M, Schwartz O, Pantelis C, et al. Role of Positive Parenting in the Association Between Neighborhood Social Disadvantage and Brain Development Across Adolescence. JAMA Psychiatry. 2017;74(8):824–32. 10.1001/jamapsychiatry.2017.1558 . PubMed PMID: 28636697; PubMed Central PMCID: PMC5710640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Conger RD, Wallace LE, Sun Y, Simons RL, McLoyd VC, Brody GH. Economic pressure in African American families: a replication and extension of the family stress model. Dev Psychol. 2002;38(2):179–93. PubMed PMID: 11881755. [PubMed] [Google Scholar]
- 15.Gilman SE, Hornig M, Ghassabian A, Hahn J, Cherkerzian S, Albert PS, et al. Socioeconomic disadvantage, gestational immune activity, and neurodevelopment in early childhood. Proc Natl Acad Sci. 2017;114(26):6728–33. 10.1073/pnas.1617698114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Verbeek T, Bockting CLH, Beijers C, Meijer JL, van Pampus MG, Burger H. Low socioeconomic status increases effects of negative life events on antenatal anxiety and depression. Women Birth J Aust Coll Midwives. 2019;32(1):e138–43. 10.1016/j.wombi.2018.05.005 . PubMed PMID: 29887508. [DOI] [PubMed] [Google Scholar]
- 17.Cragoe N, Sprowles J, Eick SM, Harvey L, Ramirez XR, Sugg GA, et al. Associations of prenatal maternal psychosocial stress and depression with neurodevelopmental outcomes in 7.5-month-old infants in the ECHO.CA.IL prospective birth cohorts. Neurotoxicol Teratol. 2025;108:107431. 10.1016/j.ntt.2025.107431 . PubMed PMID: 39800169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Baroutis D, Sotiropoulou IM, Mantzioros R, Theodora M, Daskalakis G, Antsaklis P. Prenatal maternal stress and long-term neurodevelopmental outcomes: a narrative review. J Perinat Med. 2025;53(9):1159–71. 10.1515/jpm-2025-0297 . PubMed PMID: 40966711. [DOI] [PubMed] [Google Scholar]
- 19.Dickens LT. Disparities in Diabetes in Pregnancy and the Role of Social Determinants of Health. Curr Diab Rep. 2025;25(1):33. 10.1007/s11892-025-01587-1 . PubMed PMID: 40366501; PubMed Central PMCID: PMC12078402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kong L, Chen X, Liang Y, Forsell Y, Gissler M, Lavebratt C. Association of Preeclampsia and Perinatal Complications With Offspring Neurodevelopmental and Psychiatric Disorders. JAMA Netw Open. 2022;5(1):e2145719. 10.1001/jamanetworkopen.2021.45719 . PubMed PMID: 35089349; PubMed Central PMCID: PMC8800079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Upadhyay RP, Taneja S, Strand TA, Sommerfelt H, Hysing M, Mazumder S, et al. Early child stimulation, linear growth and neurodevelopment in low birth weight infants. BMC Pediatr. 2022;22(1):586. 10.1186/s12887-022-03579-6 . PubMed PMID: 36209050; PubMed Central PMCID: PMC9547474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hee Chung E, Chou J, Brown KA. Neurodevelopmental outcomes of preterm infants: a recent literature review. Transl Pediatr. 2020;9(Suppl 1):S3–8. 10.21037/tp. .2019.09.10 PubMed PMID: 32206579; PubMed Central PMCID: PMC7082240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.McHale P, Maudsley G, Pennington A, Schlüter DK, Barr B, Paranjothy S, et al. Mediators of socioeconomic inequalities in preterm birth: a systematic review. BMC Public Health. 2022;22(1):1134. 10.1186/s12889-022-13438-9 . PubMed PMID: 35668387; PubMed Central PMCID: PMC9172189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ye W, Luo C, Huang J, Li C, Liu Z, Liu F. Gestational diabetes mellitus and adverse pregnancy outcomes: systematic review and meta-analysis. BMJ. 2022;377:e067946. 10.1136/bmj-2021-067946 . PubMed PMID: 35613728; PubMed Central PMCID: PMC9131781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Dadi AF, Miller ER, Bisetegn TA, Mwanri L. Global burden of antenatal depression and its association with adverse birth outcomes: an umbrella review. BMC Public Health. 2020;20(1):173. 10.1186/s12889-020-8293-9 . PubMed PMID: 32019560; PubMed Central PMCID: PMC7001252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ke K, Chi X, Lv H, Zhao J, Jiang Y, Jiang T, et al. Association of Breastfeeding and Neonatal Jaundice With Infant Neurodevelopment. Am J Prev Med. 2024;66(4):698–706. PubMed PMID: 38052381. [DOI] [PubMed] [Google Scholar]
- 27.Sugiyama M, Tsuchiya KJ, Okubo Y, Rahman MS, Uchiyama S, Harada T, et al. Outdoor Play as a Mitigating Factor in the Association Between Screen Time for Young Children and Neurodevelopmental Outcomes. JAMA Pediatr. 2023;177(3):303–10. 10.1001/jamapediatrics.2022.5356 . PubMed PMID: 36689245; PubMed Central PMCID: PMC9871942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Du J, Lin Y, Xia Y, Ma H, Jiang Y, Lu C, et al. Cohort Profile: The Jiangsu Birth Cohort. Int J Epidemiol. 2023;52(6):e354–63. 10.1093/ije/dyad139 . PubMed PMID: 37935835. [DOI] [PubMed] [Google Scholar]
- 29.Pharoah POD. Risk of cerebral palsy in multiple pregnancies. Clin Perinatol. 2006;33(2):301–13. 10.1016/j.clp.2006. 03.017 PubMed PMID: 16765726. [DOI] [PubMed] [Google Scholar]
- 30.Ganzeboom HBG. A new international socio-economic index [ISEI] of occupational status for the International Standard Classification of Occupation 2008 [ISCO-08] constructed with data from the ISSP 2002–2007; with an analysis of quality of occupational measurement in ISSP. 2010 Jan 1. Available from: https://www.academia.edu/78346946/A_NEW_INTERNATIONAL_SOCIO-ECONOMIC_INDEX_ISEI_OF_OCCUPATIONAL_STATUS_FOR_THE_INTERNATIONAL_STANDARD_CLASSIFICATION_OF_OCCUPATION_2008_ISCO_08_CONSTRUCTED_WITH_DATA_FROM_THE_ISSP_2002_2007_With_an_analysis_of_quality_of_occupational_measurement_in_ISSP. [Cited 29 Aug 2025].
- 31.Li R, Li R, Xie J, Chen J, Liu S, Pan A, et al. Associations of socioeconomic status and healthy lifestyle with incident early-onset and late-onset dementia: a prospective cohort study. Lancet Healthy Longev. 2023;4(12):e693–702. 10.1016/S2666-7568. (23)00211-8 PubMed PMID: 38042162. [DOI] [PubMed] [Google Scholar]
- 32.Linzer DA, Lewis JB. poLCA: An R Package for Polytomous Variable Latent Class Analysis. J Stat Softw. 2011;42:1–29. 10.18637/jss.v042.i10. [Google Scholar]
- 33.Albers CA, Grieve AJ. Test Review: Bayley, N. (2006). Bayley Scales of Infant and Toddler Development– Third Edition. J Psychoeduc Assess. 2007;25(2):180–4. 10.1177/0734282906297199. [Google Scholar]
- 34.Frezzato RC, Santos DCC, Goto MMF, Ouro MPC, do, Santos CTMD, Dutra V, et al. Fine motor skills and expressive language: a study with children with congenital hypotyreoidism. CoDAS. 2017;29(1):e20160064. 10.1590/2317-1782/20172016064 . PubMed PMID: 28300959. [DOI] [PubMed] [Google Scholar]
- 35.Qing Y, Li SZ, Liu LL, Lin SR, Hua J. Predictive validity of Bayley Scale of Infant Development-3rd Edition-Screening Test in early term children. Chin J Child Health Care. 2018;26(7):729–32. 10.11852/zgetbjzz2018-26-07-09. [Google Scholar]
- 36.Tao S, Du J, Chi X, Zhu Y, Wang X, Meng Q, et al. Associations between antenatal corticosteroid exposure and neurodevelopment in infants. Am J Obstet Gynecol. 2022;227(5):759. e1-759.e15. [DOI] [PubMed] [Google Scholar]
- 37.Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702–6. 10.1093/aje/kwh090 . PubMed PMID: 15033648. [DOI] [PubMed] [Google Scholar]
- 38.Memon SMZ, Wamala R, Kabano IH. A comparison of imputation methods for categorical data. Inf Med Unlocked. 2023;42:101382. 10.1016/j.imu.2023.101382. [Google Scholar]
- 39.Shah AD, Bartlett JW, Carpenter J, Nicholas O, Hemingway H. Comparison of random forest and parametric imputation models for imputing missing data using MICE: a CALIBER study. Am J Epidemiol. 2014;179(6):764–74. 10.1093/aje/kwt312 . PubMed PMID: 24589914; PubMed Central PMCID: PMC3939843. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Pace A, Luo R, Hirsh-Pasek K, Golinkoff RM. Identifying Pathways Between Socioeconomic Status and Language Development. Annu Rev Linguist. 2017;3(1):285–308. 10.1146/annurev-linguistics-011516-034226. [Google Scholar]
- 41.Betancourt LM, Brodsky NL, Hurt H. Socioeconomic (SES) differences in language are evident in female infants at 7months of age. Early Hum Dev. 2015;91(12):719–24. 08.002 PubMed PMID: 26371987. [DOI] [PubMed] [Google Scholar]
- 42.Clegg J, Ginsborg J. Language and Social Disadvantage: Theory into Practice. Wiley; 2006.
- 43.de Paiva GS, Lima ACVM, de Lima S, de Eickmann M. The effect of poverty on developmental screening scores among infants. Sao Paulo Med J Rev Paul Med. 2010;128(5):276–83. 10.1590/s1516-31802010000500007 . PubMed PMID: 21181068; PubMed Central PMCID: PMC10948051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Hanson JL, Hair N, Shen DG, Shi F, Gilmore JH, Wolfe BL, et al. Family poverty affects the rate of human infant brain growth. PLoS ONE. 2013;8(12):e80954. 10.1371/journal.pone.0080954 . PubMed PMID: 24349025; PubMed Central PMCID: PMC3859472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Bzdok D, Hartwigsen G, Reid A, Laird AR, Fox PT, Eickhoff SB. Left inferior parietal lobe engagement in social cognition and language. Neurosci Biobehav Rev. 2016;68:319–34. 10.1016/j.neubiorev.2016.02.024 . PubMed PMID: 27241201; PubMed Central PMCID: PMC5441272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Assari S, Race, Ethnicity FS. Status, and Children’s Hippocampus Volume. Res Health Sci. 2020;5(4):25–45. 10.22158/rhs.v5n4p. 25 PubMed PMID: 33103023. [DOI] [PubMed] [Google Scholar]
- 47.Rakesh D, Zalesky A, Whittle S. Assessment of Parent Income and Education, Neighborhood Disadvantage, and Child Brain Structure. JAMA Netw Open. 2022;5(8):e2226208. 10.1001/jamanetworkopen.2022.26208 . PubMed PMID: 35980639; PubMed Central PMCID: PMC9389347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Dufford AJ, Bianco H, Kim P. Socioeconomic disadvantage, brain morphometry, and attentional bias to threat in middle childhood. Cogn Affect Behav Neurosci. 2019;19(2):309–26. 10.3758/s13415-018-00670-3 . PubMed PMID: 30460484; PubMed Central PMCID: PMC6951020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Turesky TK, Shama T, Kakon SH, Haque R, Islam N, Someshwar A, et al. Brain morphometry and diminished physical growth in Bangladeshi children growing up in extreme poverty: A longitudinal study. Dev Cogn Neurosci. 2021;52:101029. 10.1016/j.dcn.2021.101029 . PubMed PMID: 34801857; PubMed Central PMCID: PMC8605388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Leventakou V, Roumeliotaki T, Koutra K, Vassilaki M, Mantzouranis E, Bitsios P, et al. Breastfeeding duration and cognitive, language and motor development at 18 months of age: Rhea mother-child cohort in Crete, Greece. J Epidemiol Community Health. 2015;69(3):232–9. 10.1136/jech-2013-202500 . PubMed PMID: 24336236. [DOI] [PubMed] [Google Scholar]
- 51.Dionne G, Touchette E, Forget-Dubois N, Petit D, Tremblay RE, Montplaisir JY, et al. Associations between sleep-wake consolidation and language development in early childhood: a longitudinal twin study. Sleep. 2011;34(8):987–95. 10.5665/SLEEP.1148 . PubMed PMID: 21804661; PubMed Central PMCID: PMC3138173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Scott S, Gray T, Charlton J, Millard S. The Impact of Time Spent in Natural Outdoor Spaces on Children’s Language, Communication and Social Skills: A Systematic Review Protocol. Int J Environ Res Public Health. 2022;19(19):12038. 10.3390/ijerph191912038 . PubMed PMID: 36231338; PubMed Central PMCID: PMC9566327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Pierce LJ, Reilly E, Nelson CA. Associations Between Maternal Stress, Early Language Behaviors, and Infant Electroencephalography During the First Year of Life. J Child Lang. 2021;48(4):737–64. doi:10.1017/S0305000920000501 PubMed PMID: 32900397. [DOI] [PubMed] [Google Scholar]
- 54.Festante F, Antonelli C, Chorna O, Corsi G, Guzzetta A. Parent-Infant Interaction during the First Year of Life in Infants at High Risk for Cerebral Palsy: A Systematic Review of the Literature. Neural Plast. 2019;2019:5759694. 10.1155/2019/5759694 . PubMed PMID: 31178902; PubMed Central PMCID: PMC6501141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Hackman DA, Farah MJ, Meaney MJ. Socioeconomic status and the brain: mechanistic insights from human and animal research. Nat Rev Neurosci. 2010;11(9):651–9. 10.1038/nrn2897 . PubMed PMID: 20725096; PubMed Central PMCID: PMC2950073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Vaiserman AM, Koliada AK. Early-life adversity and long-term neurobehavioral outcomes: epigenome as a bridge? Hum Genomics. 2017;11(1):34. 10.1186/s40246-017-. 0129-z PubMed PMID: 29246185; PubMed Central PMCID: PMC5732459. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
The Additional file is an updated version of the supplementary material. Reason: thousand separators have been added to the numerical values in eFigure 1 to improve readability. All other content remains unchanged. Please replace the original attachment with this version.
Data Availability Statement
The datasets generated and/or analysed during the current study are not publicly available due to our containing information that could compromise the privacy of research participants but are available from the corresponding author on reasonable request.