How growth without siblings affects adult brain and behavior

How growth without siblings affects adult brain and behavior | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Biological Sciences - Article How growth without siblings affects adult brain and behavior Chunshui Yu, Jie Tang, Jing Zhang, Wei Li, Meiyun Wang, Jingliang Cheng, and 40 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3707132/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 31 Mar, 2025 Read the published version in Nature Human Behaviour → Version 1 posted You are reading this latest preprint version Abstract With the worldwide increase of only-child families, it is crucial to understand the influence of growth without siblings (GWS) on human health, but there is limited knowledge regarding the effects of GWS on human brain. Besides, existing studies have reported paradoxical associations between GWS and mental health, most likely due to mismatched confounders and overlooked growth environments. Here, using 2,305 pairs of individuals with and without siblings well matched in confounders, we comprehensively examined the impact of GWS on changes in adult brain structure, function, and behaviors, as well as pathways from GWS, growth environments to brain and behaviors. Our findings uncovered novel associations, including GWS being linked to higher language-fiber integrity, lower motor-fiber integrity, larger regional cerebellum volume, and lower frontotemporal spontaneous brain activity. Contrary to stereotypical associations between GWS and problem behaviors, we found positive correlations of GWS with neurocognition and mental health. Despite direct effects, GWS affects most adult brain and behavioral outcomes through modifiable environments, such as socioeconomic status, maternal care, and family support, suggesting targets for interventions to enhance children’s healthy growth. Biological sciences/Neuroscience/Social neuroscience Scientific community and society/Social sciences/Psychology/Human behaviour Earth and environmental sciences/Environmental sciences/Environmental impact Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Main Text With the surge of only-child families in the world, it is critical to understand how growth without siblings (GWS) affects the development of brain, behavior, and mental health ( 1-4 ). However, existing studies examining the associations between GWS and these outcomes have yielded inconsistent and even contradictory results. For instance, some studies have reported that only children (OC) exhibit more problem behaviors (more pessimistic and risk-averse and less trustworthy, competitive, and conscientious) ( 5 ), higher prevalence of anxiety and depression ( 6 ), and worse academic outcomes ( 7 ); some indicate comparable problem behaviors ( 8-10 ), anxiety and depression ( 11 ), and academic outcomes ( 12 ); while others suggest better academic outcomes ( 13 , 14 ), less anxiety and depression ( 15 ), and more prosocial behaviors in OC individuals ( 16 , 17 ). The divergent findings may arise from many reasons, such as the mismatched confounders ( 18 , 19 ) between OC and non-OC groups. The negative outcomes observed in OC are often attributed to the lack of interactions with siblings, which have been considered a crucial social environment during children’s growth ( 5 ). Whereas the positive outcomes are associated with receiving undivided care and support from their families, as well as growing up in families with higher socioeconomic status (SES) and in more developed areas ( 9 , 13 , 20 ), indicating that both GWS itself and other growth environments may contribute to the variations in behavior and mental health between OC and non-OC individuals. The critical impacts of other growth environments are supported by the observations that GWS only affects academic achievement and mental health in urban-living children ( 13 , 21 ), and that childhood problem behaviors in OC tend to diminish or even reverse after adolescence ( 17 ). However, the specific mechanisms by which modifiable growth environments mediate or moderate the effects of GWS on behavior and mental health remain largely unknown. From the perspective of environmental neuroscience, environmental exposures, such as GWS and modifiable growth environments, can affect human behavior and mental health by altering brain structure and function. However, only two neuroimaging studies have tested the effects of GWS on brain structure. Compared with non-OC, OC show thinner thickness and larger surface area in regional cerebral cortices of children ( 22 ). In college students, OC were found to have larger gray matter volume (GMV) in supra-marginal gyrus (SMG) and smaller GMV in medial prefrontal cortex ( 23 ). Nonetheless, it is still unclear about the effects of GWS on other brain imaging properties, such as brain white matter integrity and spontaneous neuronal activity. Furthermore, while modifiable growth environments may mediate or moderate the effects of GWS on both imaging-derived phenotypes (IDPs) and behavioral phenotypes, and IDPs may in turn mediate or moderate the effects of GWS on behavioral phenotypes, there is a lack of comprehensive knowledge regarding the direct and indirect effects of GWS on adult brain and behavior. In this study, we employed propensity score matching (PSM) to create well-matched pairs of OC and non-OC participants from 6,894 healthy young Chinese Han adults recruited through the Chinese Imaging Genetics (CHIMGEN) study ( 24 ). Then, we investigated the intergroup differences between the matched OC and non-OC in terms of growth environmental exposures, GMV, white matter integrity, spontaneous neuronal activity, cognitive performance, and mental health. We defined the growth environments associated with GWS as the proximal environmental exposures (PEEs), and conducted causal mediation analysis (CMA) and structural equation model (SEM) to investigate the mediation and moderation of PEEs on the associations between GWS and brain/behavioral phenotypes, as well as to quantify the relative contribution of GWS and PEEs to these phenotypes. We anticipate that the results could inform preventive interventions aimed at rectifying adverse or enhancing beneficial GWS effects on brain, cognition, and mental health, given that many of the PEEs are modifiable. Results Participants All participants were recruited from 32 sites by the CHIMGEN study ( 24 ) with the inclusion and exclusion criteria listed in Supplementary Table 1 . After sample selection ( Supplementary Fig. 1 ), we initially included 6,894 participants (2,521 OC and 4,373 non-OC) in this study. The distribution of these participants across sites is provided in Supplementary Table 2 . To balance the overall covariate distributions between groups, PSM was used to generate the matched OC and non-OC groups in 17 confounding covariates, including age, sex, interaction between age and sex, education, body mass index (BMI), 10 genetic principal components (PCs), total intracranial volume (TIV), and frame-wise displacement (FD). Based on the propensity scores of these participants, 1:1 nearest neighbor matching created 2,305 matched pairs of OC and non-OC. The standardized mean difference (SMD) between OC and non-OC was reduced from 0.006–0.434 to 0.000-0.044 after matching ( Supplementary Table 3, Supplementary Fig. 2 ), indicating that the matched OC and non-OC groups are well balanced in confounding covariates. We applied the pair assignment in subsequent analyses, and if any individual within a pair had missing data, the entire pair would be dropped from the analysis. Brain imaging differences between OC and non-OC We used GMV to evaluate gray matter macrostructural feature, fractional anisotropy (FA) to assess brain white matter microstructural integrity, and regional homogeneity (ReHo) to assess brain spontaneous neuronal activity. We applied Combat harmonization ( 25 ) to adjust for the scanner effects for all neuroimaging measures, which could effectively reduce the bias resulting from acquiring neuroimaging data with different magnetic resonance imaging (MRI) scanners. The effectiveness of Combat harmonization was confirmed by the improved consistency of voxel-wise brain imaging measures across scanners ( Supplementary Fig. 3 ) in two volunteers whose neuroimaging data were acquired at 28 different scanners. Of the 2,305 matched pairs, 2,295, 2,278, and 2,260 qualified pairs were finally included in the GMV, FA, and ReHo analyses, respectively. In all neuroimaging analyses, we used a two-sample t -test to examine voxel-wise or phenotype-wise differences in neuroimaging measures between OC and non-OC. We applied a family-wise error (FWE) correction ( P c < 0.05) and additionally corrected for three neuroimaging measures (GMV, FA, and ReHo) and three data types (brain, behavior, and environmental exposures), resulting in an FWE-corrected P c < 0.05/3/3 = 5.55×10 − 3 . In the voxel-wise GMV analysis, compared with non-OC (n = 2,295), OC (n = 2,295) showed smaller GMV in the medial frontal and parietal cortex (MFPC), left inferior temporal gyrus (ITG_L), and right orbitofrontal cortex (OFC_R), and showed larger GMV in the bilateral cerebellar posterior lobes (CPL_L and CPL_R) and vermis 4–5 (CV4-5) (Fig. 1 a, Supplementary Fig. 4 , and Supplementary Table 4 ). For each participant, we calculated the FA values of 87 pre-defined white matter fiber tracts ( 26 ). Compared with non-OC (n = 2,278), OC (n = 2,278) showed lower FA in motor-related fiber tracts including the superior cerebellar peduncle (SCP), left corticobulbar tract (CBT_L), right dentato-rubro-thalamic tracts (DRTT_R), right corticospinal tracts (CST_R), and bilateral reticulospinal tracts (RST_L and RST_R) (Fig. 1 b and Supplementary Table 5 ). Conversely, OC showed higher FA in the right parietal aslant tract (PAT_R), which connects ITG_R and SMG_R, a subdivision of the arcuate fasciculus related to language ( 27 ). Considering the special function of the connection, we compared the functional connectivity difference between ITG_R and SMG_R, as defined by the automated anatomical labeling (AAL) atlas ( 26 ), and found stronger functional connectivity (Cohen’s d = 0.16, P = 1.83×10 − 7 , Supplementary Fig. 5 ) in OC (n = 2,260) compared to non-OC (n = 2,260). In the voxel-wise ReHo analysis, compared with non-OC (n = 2,260), OC (n = 2,260) showed lower ReHo in the bilateral ITG and the medial orbitofrontal cortex (MOFC) (Fig. 1 c, Supplementary Fig. 6 , and Supplementary Table 4 ). Behavioral differences between OC and non-OC A two-sample t -test was used to examine the differences between OC and non-OC in 34 behavioral phenotypes. We corrected for 34 phenotypes and three data types using a Bonferroni-corrected threshold of P < 0.05/34/3 = 4.90 × 10 − 4 . The behavioral differences between OC (n = 1,617-2,269) and non-OC (n = 1,617-2,269) are shown in Fig. 1 d and Supplementary Table 6 . Compared to non-OC, OC performed better in working memory (WM; Cohen’s d = 0.20, P = 1.16 × 10 − 10 ) assessed by the accuracy of the 3-back task, in immediate free recall (IFR; Cohen’s d = 0.14; P = 3.16 × 10 − 6 ) assessed by the total number of correct words of list A in five trials of the California verbal learning test II, in executive control (EC; Cohen’s d = 0.16; P = 1.32 × 10 − 7 ) assessed by the accuracy of no-go trails in the go/no-go task. In addition, OC showed more satisfaction with life (SWL; Cohen’s d = 0.13; P = 1.51 × 10 − 4 ), more self-image goals (IGS-SI; Cohen’s d = 0.13; P = 2.72 × 10 − 4 ) assessed by the interpersonal goals scale, higher perspective-taking (IRI-PT; Cohen’s d = 0.16; P = 9.50 × 10 − 6 ) and lower personal distress (IRI-PD; Cohen’s d = -0.13; P = 1.80 × 10 − 4 ) assessed by the interpersonal reactivity index scale, more openness (BFI-O; Cohen’s d = 0.32; P = 6.67 × 10 − 20 ) assessed by the big five inventory, more novelty-seeking (TPQ-NS; Cohen’s d = 0.15; P = 6.41 × 10 − 7 ) and less reward-dependence (TPQ-RD; Cohen’s d = -0.13; P = 7.09 × 10 − 6 ) assessed by the tridimensional personality questionnaire, as well as less cognitive impulsiveness (BIS-CI; Cohen’s d = -0.18; P = 6.84 × 10 − 7 ) assessed by the Barratt impulsiveness scale. PEE differences between OC and non-OC From the 32 early-life environmental exposures, we identified 23 PEEs (Fig. 2 a, Supplementary Table 7) by comparing the differences in growth environmental exposures ( P < 0.05/32/3 = 5.20 × 10 − 4 , Bonferroni correction for 32 exposures and three data types) between OC (n = 1,455-2,267) and non-OC (n = 1,455-2,267) using a two-sample t -test. Compared to non-OC, OC had parents with better education and occupation and families with higher income-to-need, larger house, more resources, less unemployment stress, financial difficulty and crisis; lived in cities with greater gross domestic product (GDP), more education expense and hospital beds; lived in safer community and more harmonious neighborhood; lived in environment with poorer indoor and outdoor air quality; received more maternal care, control, family and friend support; and experienced less physical neglect. PEE factors and their differences between OC and non-OC In the 2,902 participants with all 23 PEE assessments, the Kaiser-Meyer-Olkin (KMO) test and the Bartlett's chi-square test of sphericity were used to evaluate the suitability of the PEE correlation structure ( Supplementary Fig. 7 ) for conducting an exploratory factor analysis (EFA). The KMO = 0.83 and chi-square = 57,026 ( P < 0.001) indicate that the PEE correlation structure is suitable for EFA. We obtained an optimal five-factor EFA model with very simple structure (VSS) = 0.8 and minimum average partial (MAP) = 0.02. Based on the five-factor EFA model, we defined PEEs with factor loading > 0.45 as the main contributors for each factor. According to the included contributors (PEEs), we named the five PEE factors as family (FSES), adverse (ASES) and city-level (CSES) socioeconomic status, air pollution (AP), and support and care (SC), respectively. FSES included paternal and maternal occupation, paternal and maternal education, family resources, and income-to-need with factor loadings of 0.72, 0.74, 0.82, 0.80, 0.51, and 0.45; ASES included financial difficulty, financial crisis, unemployment stress, and home inadequacy with factor loadings of 0.77, 0.65, 0.66, and 0.54; CSES included education expense, GDP, and hospital bed with factor loadings of 0.94, 0.90, and 0.60; AP included PM 2.5 and NO 2 with factor loadings of 0.81 and 0.79; and SC included maternal care, family support, and friend support with factor loadings of 0.61, 0.81 and 0.58 (Fig. 2 b). To account for uncertainty in the within-sample prediction, a 10-fold cross-validation CFA was conducted to predict out-of-sample PEE factor scores. In each iteration, 90% participants were used to estimate the factor loadings of the CFA model, which were then applied to calculate the PEE factor scores of other 10% participants. The goodness-of-fit statistics indicated a perfect fit for the models established for each iteration and the full sample ( Supplementary Table 8 ). PEE factor scores estimated in the 10-fold cross-validation were highly correlated with those estimated in the full sample (all rho = 0.99, P < 0.001) ( Supplementary Fig. 8 ), indicating that the five-factor CFA model is stable, and we used the out-of-sample PEE factor scores in the subsequent analyses. We also compared PEE factor scores between OC (n = 1,451) and non-OC (n = 1,451), and found significant intergroup differences in all five factor scores ( P < 0.05/5/3 = 3.33 × 10 − 3 ) (Fig. 2 c ) . PEE factor scores mediate or moderate the effects of GWS on brain and behavior For the five GWS-related PEE factors, 16 IDPs, and 11 behavioral phenotypes, we used CMA to test the mediation and moderation of the five PEE factors on the associations of GWS with the 27 phenotypes (Fig. 3 a-b). Since the 16 IDPs were highly correlated ( Supplementary Fig. 9 ), we estimated the effective number of IDPs (n = 7.63) and set a Bonferroni corrected threshold of P < 0.05/5/ (11 + 7.63) = 5.36× 10 − 4 . The mediation and moderation effects of five PEE factors are presented in Fig. 3 c-d and the full statistics are provided in Supplementary Table 9 . Specifically, FSES mediated the positive associations of GWS with BFI-O ( β = 0.203), IFR ( β = 0.124), WM ( β = 0.088), SWL ( β = 0.104), TPQ-NS ( β = 0.112), IGS-SI ( β = 0.097), and GMV-CPL_L ( β = 0.079), and the negative associations of GWS with BIS-CI ( β = -0.089), GMV-MFPC ( β = -0.097), ReHo-ITG_L ( β = -0.136), ReHo-ITG_R ( β = -0.124), FA-CBT_L ( β = -0.103), FA-CST_R ( β = -0.085), and FA-RST_L ( β = -0.092). SC mediated the positive associations of GWS with SWL ( β = 0.099), IRI-PT ( β = 0.033), TPQ-RD ( β = 0.041), and IGS-SI ( β = 0.020), and the negative associations of GWS with BIS-CI ( β = -0.061) and IRI-PD ( β = -0.027). ASES mediated the positive associations of GWS with SWL ( β = 0.142), and the negative associations of GWS with BIS-CI ( β = -0.061), IRI-PD ( β = -0.088), and ReHo-ITG_L ( β = -0.059). CSES mediated the positive association between GWS and IGS-SI ( β = 0.025). AP mediated the negative associations of GWS with GMV-MFPC ( β = -0.018) (Fig. 3 c and Supplementary Table 9 ). All these mediation effects were significant at P < 5.36× 10 − 4 . However, we did not find any significant moderation ( P < 5.36× 10 − 4 ) of PEE factors on the associations between GWS and phenotypes (Fig. 3 d and Supplementary Table 9 ). IDPs mediate or moderate the effects of GWS on behavior We used CMA to test the mediation and moderation of 16 IDPs on the associations of GWS with 11 behavioral phenotypes ( P < 0.05/7.63/11 = 5.95 × 10 − 4 , Bonferroni corrected). The mediation and moderation effects of IDPs are presented in Fig. 3 e-f and the full statistics are provided in Supplementary Table 10 . We found that the positive association between GWS and IFR was mediated by GMV-MFPC ( β = 0.017), ReHo-ITG_R ( β = 0.016), and FA-SCP ( β = 0.008); the positive correlation between GWS and EC was mediated by GMV-CV4-5 ( β = 0.013); the positive correlation between GWS and IGS-SI was mediated by GMV-CV4-5 ( β = 0.014); the negative correlation between GWS and IRI-PD was mediated by GMV-CPL_R ( β = -0.011); and the positive correlation between GWS and TPQ-NS was mediated by GMV-ITG_L ( β = 0.012) and ReHo-ITG_R ( β = 0.019). Besides, we found GMV-ITG_L ( β = 0.014) and GMV-OFC_R ( β = 0.010) showed reverse mediation effects on the association between GWS and IRI-PD; and GMV-ITG_L showed reverse mediation effects on the association between GWS and BIS-CI ( β = 0.019) (Fig. 3 e and Supplementary Table 10 ). All these mediation effects were significant at P < 5.95 × 10 − 4 . However, we failed to find any significant moderation ( P < 5.95 × 10 − 4 ) of IDPs on the associations of GWS with behavioral phenotypes (Fig. 3 f and Supplementary Table 10 ). Pathways from GWS to brain and behavior CMA indicate that PEE factors mediate rather than moderate the associations of GWS with IDPs and behavioral phenotypes. However, CMA estimates the mediation without taking account of other predictors and cannot identify the complex pathways including two or more sequential mediators. Therefore, we performed structural equation model (SEM) to simultaneously model the complex relationships among the GWS, PEE factors, IDPs, and behavioral phenotypes in 2,618 participants without missing data in these variables. In the measurement model, we used the same structure as the five-factor-CFA model to construct five PEE latent variables (LVs) of FSES, ASES, CSES, SC, and AP. Based on the brain imaging measure (GMV, FA, or ReHo) and the direction of GWS effect (higher or lower in OC) of IDPs, we constructed four LVs: small GMV (SGMV) constructed by the three IDPs with smaller GMV in OC; large GMV (LGMV) by the three IDPs with larger GMV in OC; low FA (LFA) by the six IDPs with lower FA in OC; and low ReHo (LReHo) by the three IDPs with lower ReHo in OC. The right PAT with higher FA in OC was used as a single variable in the SEM analysis. We did not construct behavioral LVs owing to their weak correlations ( Supplementary Fig. 10 ). In the structural model, we defined GWS as the independent variable; FSES, ASES, CSES, SC, AP, SGMV, LGMV, LFA, FA-PAT_R, and LReHo as mediators; and IFR, EC, WM, TPQ-RD, TPQ-NS, BFI-O, BIS-CI, IGS-SI, IRI-PD, IRI-PT, and SWL as outcomes ( Supplementary Fig. 11 ). The constructed SEM showed an acceptable goodness-of-fit (CFI = 0.919, RMSEA = 0.042, and SRMR = 0.055). In SEM, we defined a two-sided P < 0.05 as the threshold for statistical significance. The model explained 0.8–13.3% variances of IDPs (highest for LReHo) and 2.0-21.1% variances of behavioral phenotypes (highest for SWL) (Fig. 4 a). The relative effects of GWS and PEEs on IDPs and behavioral phenotypes are summarized in Fig. 4 b, and all significant associations in SEM were shown in Fig. 4 c. For IDPs, FSES, AP, and SC showed comparable effects on SGMV ( β = -0.095, -0.094, and − 0.07, respectively); FSES, CSES, and GWS showed comparable effects on LGMV ( β = 0.102, 0.100, and 0.093, respectively); FSES was the strongest predictor of LFA ( β = -0.099) and LReHo ( β = -0.254); and GWS was the strongest predictor of FA-PAT_R ( β = 0.084) (Fig. 4 b-c). For the behavioral phenotypes, GWS was the strongest predictor of EC (β = 0.106); FSES was the strongest predictor of BFI-O ( β = 0.333), WM ( β = 0.153), and TPQ-NS ( β = 0.126); SC was the strongest predictor of SWL ( β = 0.442), BIS-CI ( β = -0.300), TPQ-RD ( β = 0.193), and IRI-PT ( β = 0.188); IFR was mainly affected by FSES ( β = 0.158), CSES ( β = 0.109), and AP ( β = -0.101); IGS-SI was mainly affected by FSES ( β = 0.125), and SC ( β = 0.119); and IRI-PD was mainly affected by ASES ( β = 0.142), and CSES ( β = 0.127) (Fig. 4 b-c). For the associations of IDP with behavioral phenotypes, SGMV was primarily associated with IFR ( β = -0.080) and BIS-CI ( β = -0.111); LGMV was mainly associated with IGS-SI ( β = 0.089); FA-PAT_R was mainly associated with EC ( β = 0.050); and LFA was associated with BIS-CI ( β = 0.046) (Fig. 4 c). The identified direct and indirect effects of GWS on IDPs and behavioral phenotypes are summarized in Fig. 4 d. GWS showed direct effects on FA-PAT_R ( β = 0.084), LGMV ( β = 0.093), LReHo ( β = -0.111), EC ( β = 0.106), and TPQ-RD ( β = -0.069). For the indirect effects (Fig. 4 d and Supplementary Table 11 ), GWS positively affected: (1) LGMV via the mediators of FSES ( β = 0.057) and CSES ( β = 0.022); (2) IFR via the mediators of FSES ( β = 0.089), CSES ( β = 0.024), and the sequential mediators of FSES-SGMV ( β = 0.004), SC-SGMV ( β = 0.0005) and AP-SGMV ( β = 0.001); (3) EC via the mediators of AP ( β = 0.007) and FA-PAT_R ( β = 0.004); (4) WM via the mediators of FSES ( β = 0.086) and CSES ( β = 0.019); (5) SWL via the mediators of ASES ( β = 0.049) and SC ( β = 0.040); (6) BFI-O via the mediators of FSES ( β = 0.188) and CSES ( β = 0.017); (7) IRI-PT via the mediators of FSES ( β = 0.057) and SC ( β = 0.017); (8) IGS-SI via the mediators of FSES ( β = 0.071), CSES ( β = 0.014), SC ( β = 0.011), LGMV ( β = 0.008), and CSES-LGMV ( β = 0.002); (9) TPQ-NS via the mediator of FSES ( β = 0.071) (Fig. 4 d and Supplementary Table 11 ). GWS negatively affected: (1) SGMV via the mediators of FSES ( β = -0.053), AP ( β = -0.011) and SC ( β = -0.007); (2) LReHo via the mediators of FSES ( β = -0.143) and CSES ( β = -0.036); (3) LFA via the mediator of FSES ( β = -0.056); (4) BIS-CI via the mediators of SC ( β = -0.027), AP ( β = -0.009), and FA-PAT_R ( β = -0.004); (5) TPQ-RD via the mediator of FSES ( β = -0.040); and (6) IRI-PD via the mediators of ASES ( β = -0.050) AP ( β = -0.010), and SC ( β = -0.007) (Fig. 4 d and Supplementary Table 11 ). Consistent pathways from GWS to brain and behavior CMA and SEM have identified several consistent mediation pathways from GWS to IDPs and behavioral phenotypes (Fig. 5 ). Specifically, GWS affected SGMV, LFA, IFR, WM, SWL, BIS-CI, IRI-PT, IRI-PD, IGS-SI, TPQ-NS, and BFI-O only through indirect pathways, but with different mediators. GWS affected FA-PAT_R only through direct pathway. GWS affected EC, TPQ-RD, LGMV, and LReHo through both direct and indirect pathways. Of the 16 brain and behavioral phenotypes, GWS affected 11 phenotypes via FSES, nine via SC, four via CSES, three via AP, and two via ASES. SGMV, LGMV, and FA-PAT_R were IDP mediators for the associations between GWS and behavioral phenotypes. In addition, CMA and SEM also revealed several indirect pathways whose effect directions were opposite to the total GWS effects, including the indirect effects of SC on TPQ-RD and TPQ-NS; AP on IFR; AP-SGMV, FSES-SGMV, and SC-SGMV on BIS-CI; and AP-SGMV and FSES-SGMV on IRI-PT. Discussion In up to 2,305 pairs of well-matched OC and non-OC, we investigated how GWS affects adult brain and behavior by testing their differences in terms of early-life growth environments, adult brain imaging and behavioral phenotypes, as well as the causal pathways from GWS to growth environments to brain and behavior. We identified novel associations of GWS with brain and behaviors, including white matter integrity, cerebellum volume, spontaneous neuronal activity, memory, executive control, and mental health. OC lived in better socioeconomic environment and received more maternal care, family and friend support than non-OC, in line with the resource dilution and attachment theory ( 9 , 28 ). We also found that most GWS effects on brain and behavior were mediated by modifiable growth environmental exposures, informing interventions to improve the healthy development of brain and behavior by increasing advantageous and reducing disadvantageous environmental exposures in early life. Consistent with an earlier study, OC showed smaller MFPC and ITG volumes ( 23 ), and the smaller MFPC volume was associated with better memory in the present study. The MFPC volume reaches its peak before puberty, followed by a nearly linear reduction ( 29 ). The volume reduction before adult is an integrated reflection of the synaptic pruning, axonal caliber alteration, and myelination ( 30 , 31 ) which is associated with higher neuronal efficiency ( 32 , 33 ). Thus, our results may be explained by accelerated normative growth of cerebral cortex ( 34-36 ) in OC individuals. As the GWS effect on MFPC volume was mainly mediated by family SES, family SES may be an intervention target for memory improvement. The larger cerebellar volume in OC was attributed to both direct effect and indirect effect mediated by family SES, in line with positive correlation between SES and cerebellar volume ( 36-38 ). We also extend previously reported association between family SES and FA (white matter integrity) ( 39 ) to the mediation of family SES in the association between GWS and lower FA in motor-related fibers. Furthermore, OC showed higher FA in the right PAT (a language-related tract) and higher functional connectivity between right ITG and SMG, supporting the superior language ability in OC ( 40 ). PAT integrity was only affected by the direct GWS effect, indicating that its integrity cannot be changed by improving these common growth environments. Although OC showed lower ReHo (spontaneous neuronal activity), we failed to find reliable association between ReHo and behavior, which cannot provide useful information for intervention. Contrary to the expectation of more behavioral problems in OC ( 5 ), we found that OC outperformed non-OC in various behavioral outcomes. In social-emotional skills, OC exhibited more prosocial (perspective-taking), and less distressful (personal distress), which were completely mediated by growth environments: family SES, support and care for perspective-taking, and adverse SES and support and care for personal distress. Similarly, OC showed higher levels of life satisfaction which was mediated by more support and care and less adverse SES. OC also showed more self-image goals in interpersonal relationship, which were mediated by family and city-level SES, support and care, and cerebellar volume, and the association between cerebellar volume and social cognition has been reported previously ( 36 , 41 ). These results suggest that the poorer social cognition and life satisfaction of non-OC could be enhanced by improving socioeconomic status and providing more care and support during early years of life. Regarding personality, OC showed more openness and novelty-seeking but less reward-dependence and cognitive impulsiveness. Both openness and novelty-seeking are associated with creativity, which is higher in OC college students ( 23 ). The GWS effects on these two personality traits were primarily mediated by family SES, suggesting that creativity may be facilitated by improving family socioeconomic circumstances during early development. Furthermore, the GWS effect on cognitive impulsiveness was completely mediated by support and care and family SES, in line with the findings that self-control could be enhanced by improving caregiving and family SES ( 42 , 43 ). The lower level of reward dependence observed in OC was mainly driven by the direct effect of GWS, but the effect was mitigated by support and care. Via this regulatory pathway, parents could help their children achieve the balance between dependency and independency by providing appropriate care and support while avoiding controlling behaviors. In terms of neurocognitive abilities, OC showed better performance in verbal learning and memory, working memory, and executive control than non-OC. However, GWS affects different neurocognitive domains through distinct pathways. For example, GWS positively affected verbal learning memory and working memory solely through indirect pathways with FSES, CSES, SC, and SGMV serving as mediators. On the other hand, the effect of GWS on executive control was primarily through the direct pathway, although indirect pathway also contributed to a smaller extent. These findings emphasize the importance of designing neurocognitive improvement strategies in a domain-specific manner, although a single mediator may involve in the regulation of multiple neurocognitive domains. Except for white matter integrity of the right PAT and executive control ability, the effects of GWS on brain and behavior were mediated by modifiable environmental exposures, which provides us opportunities to improve adult behavioral outcomes by modifying early-life environmental exposures according to the indirect pathways and their relative effects identified in this study. For instance, the improvement of family socioeconomic status may benefit for the development of brain and neurocognition, and the provision of more maternal care and support could benefit for life satisfaction and mental health These interventions should be implemented during the early years of life to help narrow the gaps between OC and non-OC individuals. Two limitations should be considered when one interprets our findings. First, all participants in this study were Chinese, which may limit the generalizability of the results to other populations. Second, as many environmental assessments relied on participant recall, there is a possibility of recall bias that may have affected our results. Methods Participants The CHIMGEN study recruited 7,306 healthy Chinese Han adult participants aged 18-30 years from 32 research centers in 22 provinces of China ( 24 ). After excluding those with unqualified data or missing covariate variables ( Supplementary Fig. 1 ), we initially included 6,894 participants, from which we generated 2,305 pairs of well-matched OC and non-OC participants. Written informed consent was obtained from each participant, and the CHIMGEN study was approved by the Medical Research Ethics Committees of Tianjin Medical University General Hospital (IRB2015-092-01) and other institutions. Neuroimaging data acquisition and preprocessing All the neuroimaging data were acquired by ten different types of 3T magnetic resonance imaging (MRI) scanners, and the scanning sequences and parameters for each type of MRI scanners are provided in Supplementary Table 12-14 . The structural MRI (sMRI) data were used to calculate gray matter volume (GMV) to assess brain macrostructural features; the diffusion tensor imaging (DTI) data were used to calculate fractional anisotropy (FA) to evaluate brain white matter integrity; and the resting-state functional MRI (rs-fMRI) data were used to calculate regional homogeneity (ReHo) and functional connectivity (FC), which were used to assess spontaneous neuronal activity within brain regions and temporal coherence of spontaneous neuronal activity between brain regions. sMRI data preprocessing . The sMRI data were preprocessed using computational anatomy toolbox (CAT12 v1364) (http://dbm.neuro.uni-jena.de/cat) implemented in statistical parametric mapping (SPM12; http://www.fil.ion.ucl.ac.uk/spm). After correcting for image inhomogeneity from B1-field bias, the structural images were segmented into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) using a model based on an adaptive Maximum A Posterior ( 44 ). To improve image normalization, the population-specific tissue probability templates for GM, WM, and CSF in Montreal Neurological Institute (MNI) space were derived from 6,000 CHIMGEN participants using the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) algorithm ( 45 ). The individual GM images were normalized to the GM template using a two-step DARTEL algorithm and resampled into a cubic voxel of 1.5-mm. Modulation was performed on the normalized GM images to preserve absolute GMV. The resulting voxel-wise GMV maps were smoothed with a Gaussian kernel of 8 × 8 × 8 mm 3 full width at half maximum (FWHM). The TIV of each participant was calculated as the sum of the volumes of GM, WM, and CSF. DTI data preprocessing. DTI data were preprocessed using FMRIB’s Software Library (FSL v5.0.10; www.fmrib.ox.ac.uk/fsl). Non-brain tissues were removed from b = 0 images using the brain extraction tool to generate a binary mask of brain tissue. The “EDDY_OPENMP” program was used to evaluate and repair image displacement, signal dropout caused by head motion, and image distortion caused by eddy current. A linear least square algorithm was applied to estimate diffusion tensor and to calculate FA value of each voxel using the DTIFIT program. A two-step procedure was conducted to estimate the normalization parameters between individual diffusion and MNI spaces. Specifically, individual b = 0 images were aligned to structural images using Boundary-Based Registration (BBR) ( 46 ). The obtained transformation matrix was concatenated with the DARTEL deformation field from individual to MNI space generated in the sMRI data preprocessing. The individual FA maps were normalized to the MNI space using the merged deformation field (BBR+DARTEL) and resampled into a cubic voxel of 2-mm. A mean FA image of all individuals was created and “thinned” to generate a mean white matter skeleton representing the centers of tracts common to all individuals. Each participant’s aligned FA images were projected onto the mean FA skeleton by filling the mean FA skeleton with FA values from the nearest tract center, achieving by searching perpendicular to the local skeleton structure for maximum value. We then calculated the mean FA values of 87 white matter tracts derived from 1,065 subjects of the human connectome project (HCP) ( 26 ). rs-fMRI data preprocessing. The rs-fMRI data were preprocessed using SPM12 and data processing assistant for resting-state fMRI analysis (DPARSFA v4.4) ( 47 ). As three different repetition times (TRs: 0.71, 0.8, and 2 s) were used to acquire rs-fMRI data and the functional images acquired at each TR was defined as a functional volume, the first several functional volumes (≈ 10 s: 15 volumes for TR = 0.71 s; 13 for TR = 0.8 s; and 5 for TR = 2 s) were discarded to allow signals to reach equilibrium. The remaining volumes were corrected for intra-volume temporal differences using sinc-interpolation and the inter-volume head motion was estimated and corrected by realigning each volume to the mean volume using rigid-body transformation. Based on the estimation of head motion, we excluded the participants whose fMRI data had the maximum displacement in any of the three orthogonal directions > 3 mm or a maximum rotation > 3.0 degree. In the remaining participants, we calculated FD to index volume-to-volume changes in head position using the Jenkinson method ( 48 ). We then conducted independent component analysis based automatic removal of motion artifacts (ICA-AROMA) ( 49 ), from which we identified and removed the independent components for motion artifacts. After removing non-brain tissues, the functional images were co-registered to the structural images using the BBR method. All co-registered functional volumes were normalized to MNI space using the deformation fields derived from the sMRI data preprocessing and were resampled to 3-mm isotropic voxels. After regressing out linear trend and signals from WM and CSF, a temporal band-pass filtering (0.01-0.08 Hz) was applied to reduce low-frequency drift and high-frequency noise. The unsmoothed fMRI data were used for computing ReHo. Here, ReHo of each voxel was defined as the Kendall’s coefficient concordance of the time series of this voxel with its nearest neighbors (26 voxels). The whole process was repeated for all GM voxels to generate the voxel-wise ReHo map of each participant. For standardization, ReHo of each voxel was subtracted from the mean and divided by the standard deviation (SD) of ReHo values of all GM voxels. Then, the ReHo maps were smoothed with a kernel of FWHM = 8 mm. For the special fiber tracts of interest, we also calculated the FCs between their connecting brain regions defined by the automated anatomical labeling (AAL) atlas ( 26 ). Here, FC was calculated based on the smoothed fMRI data (FWHM = 8 mm) and was defined as the Pearson correlation coefficient between the mean time-courses of two brain regions. To improve normality, the correlation coefficients were transformed to z value using Fisher r-to-z transformation. N euroimaging data q uality control (QC). A series of QC procedures were applied both before and after the acquisition of brain MRI data. For example, we optimized scanning parameters for each MRI scanner before acquisition. We identified and excluded MRI data with lesions, anatomical abnormalities, imaging artefacts, parameter inconsistency, and incomplete whole-brain coverage immediately after acquisition. We also checked the preprocessed MRI data to identify errors or imperfections that may have emerged during imaging data preprocessing steps. If they were identified in the preprocessed MRI data of a participant, we tried to identify reasons and re-run the pipeline after fixing the identified problems. If the reprocessed MR images were still problematic, the participant would be excluded from the analyses. The checked error and imperfection included incorrect non-brain tissue removal, bad tissue segmentation, imperfect spatial normalization, and intensity normalization error. Following the QC procedures, we finally included voxel-wise GMV maps from 6,852 participants, tract-based FA measurements from 6,844 participants, and voxel-wise ReHo maps and FCs between brain regions from 6,698 participants. Combat harmonization was applied to MRI metrics to adjust for the scanner effect ( 25 ). The effectiveness of harmonization was assessed by within-subject consistency of each imaging measure between scanners in two participants who have traveled to the research sites and were scanned by 28 MRI scanners used in the CHIMGEN study. Behavioral assessments In the present study, we included 34 behavioral variables that were used to measure neurocognition, social cognition, personality, and mental health. Verbal learning and memory . The second edition of the California verbal learning test (CVLT-II) was used to test verbal learning and memory ( 50 ). The test includes two lists of words: the list A consists of 16 words of four semantic categories (furniture, vegetable, vehicle, and animal) and the list B also includes 16 words of four categories (vegetable, animal, musical instrument, and house part name; the former two from the word list A). There were five trials for the word list A and one trial for the word list B and words were randomly presented, irrespective of semantic categories. In each trial, the experimenter read aloud the word list A at a rate of one word per second, and the participant was asked to recall as many words as possible. The experimenter recorded the number of correct words of list A in the five trials, which was used to assess immediate free recall (IFR). Immediately after the five trials, the participant was asked to learn and recall word list B once. Then, the participant was asked to engage in short delay free and cued recall of the word list A immediately, as well as long delay free and cued recall of the word list A after a 20-minute of the short delay recall. The participant was asked to recall in no order in free recall and to categorize the words into four categories in cued recall. The numbers of correct words in the four recall tasks were used to assess episodic memory, including the short delayed free (SDFR) and cued (SDCR) recall and the long delayed free (LDFR) and cued (LDCR) recall. Working memory . The letter n-back test (1-back and 3-back) was used to assess working memory ( 51 ). In the 1-back task, the participant was required to infer whether the current stimulus matched the previous one in the sequence of stimuli. In the 3-back task, the participant was required to infer whether the current stimulus matched the one three steps earlier. Both tasks included one block with 60 trials. Each stimulus was presented for 200ms, and the stimulation interval was 1800ms. Before the formal tasks, participants underwent a practice session of the 1-back task, and only those who achieved an accuracy rate higher than 75% were allowed to proceed to the formal task. E-Prime 2.0 software (Psychology Software Tools) was used to present stimuli and record participant responses. The accuracy for 3-back tasks were used to assess working memory (WM). Visuospatial memory. The Rey-Osterrieth complex figure test (ROCFT) was used to assess visuospatial capability ( 52 ). In ROCFT, the participant was asked to copy and recall a complex line graph at three time points. At the beginning, the participant was asked to copy the graph onto a blank sheet of paper. Immediately after copying, the participant was asked to redraw the graph from memory (immediate recall). Twenty-five minutes later, the participant was asked to reproduce the graph again (delayed recall). A 36-point scoring system was used to assess the accuracy of immediate (RO-IR) and delayed (RO-DR) recall tasks based on the scoring guideline for each element of the complex figure. To minimize inter-rater variability, all ROCFT assessments were conducted by a trained group of raters. Information processing speed . The symbol digit modalities test (SDMT) was used to assess information processing ability ( 53 ). The test involves pairing of single digits with abstract symbols. The keys consist of nine abstract symbols, each representing a specific digit. The participant was presented with these symbols and asked to write down the corresponding digits as fast as possible within 90 seconds. The information processing performance was assessed based on the number of correct responses within the given time frame. Executive control . The go/no-go test was used to evaluate sustained attention and behavior inhibition ( 54 ). During “go” trial (90% of 210 trials), the current letter was different from the previous letter and the participants were required to respond quickly by pressing the button. In contrast, during the “no-go” trial (10% trials), the current letter was the same as the previous letter and the participant were required not to press any button. The E-Prime 2.0 was used to present stimuli and collect responses. The mean accuracy of “no-go” trials was used to assess executive control (EC). Empathy. The interpersonal reactivity index (IRI) questionnaire was used to test dispositional empathy ( 55 ). The IRI includes four distinct conducts: perspective-taking (IRI-PT) measures the tendency to spontaneously adopt the psychological point of others; fantasy (IRI-F) tests the tendency to imaginatively transpose themselves into the feelings and actions of fictitious characters; empathic concern (IRI-EMC) reflects other-oriented feelings of sympathy and concern towards others who are experiencing unfortunate circumstances; and personal distress (IRI-PD) assesses self-oriented feelings of personal anxiety and unease in tense interpersonal settings. Interpersonal goals. The interpersonal goals scale (IGS) was used to test one’s interpersonal goals for their relationships ( 56 ). IGS assesses the compassionate goals (IGS-C) to support others and self-image goals (IGS-SI) to create and maintain desired self-images. Mental health . The second version of Beck depressive inventory (BDI) was used to assess the severity of depression ( 57 ). The state-trait anxiety inventory was used to measure state anxiety (SAI) and trait anxiety (TAI) ( 58 ). The positive and negative affect scale was used to measure positive affect (PNAS-PA) and negative affect (PNAS-NA) ( 59 ). The perceived stress scale was used to measure perceived stress (PSS) ( 60 ). The satisfaction with life scale was used to assess satisfaction with life (SWL) ( 61 ). Personality. The big-five inventory (BFI) was used to assess five personality traits, including the agreeableness (BFI-A), conscientiousness (BFI-C), extraversion (BFI-E), neuroticism (BFI-N), and openness (BFI-O) ( 62 ). The tridimensional personality questionnaire (TPQ) was used to assess three personality traits associated with distinct neurotransmitters: novelty-seeking (TPQ-NS) with dopamine; harm-avoidance (TPQ-HA) with serotonin; and reward dependence (TPQ-RD) with norepinephrine ( 63 ). The Barratt impulsiveness scale (BIS) was used to assess impulsiveness ( 64 ), including the motor impulsiveness (BIS-MI), cognitive impulsiveness (BIS-CI), and no-planning impulsiveness (BIS-NPI). B ehavioral data Q Cs . Each behavioral assessment underwent the following QC procedures: (a) we confirmed the consistency between input data and raw data to ensure that the input data were free of any input errors; (b) we excluded participants with accuracy < 75% on “1-back” or “go” trails, indicating the lack of compliance of the participants; (c) we further performed median-based outlier removal (discarding values greater than 3 times the median absolute deviation from the overall median) for each measure. Finally, a total of 34 behavioral variables from 4,828-6,768 participants were used in following analyses (Supplementary Fig. 1) . Growth environmental exposures The 32 growth environmental exposures potentially related to growth without siblings (GWS) were evaluated from the following questionnaires and approaches: The socioeconomic status (SES) questionnaire was used to assess family income, parental occupation and education, family resources, unemployment stress, financial difficulties and crisis, home inadequacy, community safety, and neighbor indifference. Each measure was assessed at three age windows of 0-10, 11-20, and > 21 years, and the mean value of age windows was used to represent the corresponding measure. Family income was represented by income to need, which was defined as the annual household income per capita. Parental education was divided into six grades: 1 = primary school or below, 2 = middle school, 3 = high or technical school, 4 = junior college, 5 = bachelor, 6 = master or doctor. Parental occupation included eight classes: 1 = day laborer or unemployed, 2 = manual worker or self-employed, 3 = operators of equipment, 4 = production personnel, 5 = business or service staff, 6 = civil servants or employees of companies, 7 = professional and technical personnel, 8 = manager of government, institution, or company. Family resources were defined as the sum of 14 binary items (0 = no and 1 = yes) including private room, car, computer, study desk, learning place, Internet, educational software, reference books, television, refrigerator, washing machine, masterpieces, collect books, and artistic works ( 65 ). A three-scoring system was applied to assess family unemployment stress (1 = never, 2 = occasional, 3 = often), financial difficulties (1 = never, 2 = occasional, 3 = often), financial crisis (1 = never, 2 = occasional, 3 = often), home inadequacy (1 = spacious, 2 = moderate, 3 = crowd), and community safety (1 = safe, 2 = moderate, 3 = violent) and neighbor indifference (1 = harmonious, 2 = moderate, 3 = indifferent). To get comparable assessments, we provided a standard set of instructions to the participants A five-scoring system was used to assess indoor air quality (1 = very good; 2 = good; 3 = moderate; 4 = poor; 5 = very poor) at six age windows: 0-5, 6-10, 11-15, 16-20, 21-25, and > 26 years. The mean value of age windows was used to represent the indoor air quality. Each participant was asked to recall residential coordinates of the years from birth to recruitment. If the participant moved in a year, the address where they lived for the longest duration was defined as the residential coordinate of that year. Using annual PM 2.5 and NO 2 data provided by two prior studies( 66 , 67 ), we extracted two measurements for each participant from birth year to the year of recruitment. The average values of the annual data for the available years were used to represent PM 2.5 and NO 2 exposures of the participant. Based on each residential address, we extracted city-level statistical data on GDP per capita, the number of hospital beds per 10,000 people, and education expense per 10,000 people from the annual report of the National Bureau of Statistics of China (http://www.stats.gov.cn/). The annual data of each city-level indicator were obtained for each participant from their birth year to the year of recruitment, and the average value of the lifetime data was used to represent the exposure. The maternal (PBIM) and paternal (PBIF) versions of 23-item parental bonding instrument (PBI) were used to evaluate parenting styles perceived by the participant before 16 years old ( 68 ). A four-scoring system (0 = strongly disagree, 1 = disagree, 2 = agree, and 3 = strongly agree) was applied to each item. Based on the combination of these items, we obtained the assessments of care, encouragement, and control from mother and father, respectively. The 12-item multidimensional scale of perceived social support (MSPSS) was used to assess the self-perceived social support of each participant ( 69 ). Each item has seven levels (1-7: very strongly disagree, strongly disagree, disagree, neutral, agree, strongly agree, very strongly agree), and higher score represents greater social support. Based on the combination of these items, we obtained the assessments of supports from family, friends, and others. The childhood trauma questionnaire (CTQ) was applied to assess childhood abuse experiences ( 70 ). CTQ consists of 28 items reflecting life experience before the age of 16 years and adopts a five-point scoring system (never, occasionally, sometimes, often, and always). The CTQ subscales include emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect. The scores of each subscale range from 5 to 25 points, with higher scores indicating more severe trauma. Confounding covariates We included the following 17 covariates in the PSM analysis: age, sex, the interaction of age and sex, education, BMI, genetic population stratification, TIV and mean FD. In this study, sex was confirmed using genotyping data. The methodology for calculating genetic population stratification for the CHIMGEN participants has been described in our previous study ( 71 ). We found that 10 PCs derived from principal component analysis of genomic data could account for the major variance of population stratification, which were used as the covariates to control for genetic population stratification in the current study. Statistical analysis Propensity score matching (PSM) From 2,521 OC and 4,373 non-OC participants, PSM was used to create well-matched OC and non-OC groups across 17 confounding covariates. The PSM analysis was performed using the “MatchIt” package from R 4.1.2 ( 72 ). Logistic regression was used to estimate the individual’s propensity score (PS), which represents the probability of being in the OC group given the covariates. Based on the propensity scores of each participant, we conducted one-to-one matching between OC and non-OC using the greedy nearest-neighbor method (without replacement and caliper = 0.1), resulting in 2,305 pairs of matched OC and non-OC participants. The standardized mean difference (SMD) was used to assess the matching quality, and an SMD < 0.10 indicated a balanced distribution between two groups ( 73 ). We applied the pair assignment in subsequent analyses, and if any individual in one pair had missing data, the entire pair would be dropped from the analysis. Comparisons of brain imaging phenotypes between OC and non-OC With SPM12, a two-sample t -test was performed to identify voxel-wise GMV differences between the well-matched OC and non-OC pairs. Multiple comparisons were corrected using a hybrid family-wise error (FWE) method, where we corrected for voxel-wise comparisons based on random field theory (RFT) ( 74 ), and further corrected for the three neuroimaging features (GMV, FA, and ReHo) and the three data types (environmental, neuroimaging and behavioral data). Thus, a hybrid FWE-corrected threshold was P < 2.17 × 10 -6 /3/3 = 2.41 × 10 -7 . We defined each cluster with significant GMV difference as an imaging-derived phenotype (IDP), and defined the mean GMV of the cluster in each participant as the IDP value of the participant. The same approach was used to identify voxel-wise differences in ReHo between the matched OC and non-OC pairs, and used a hybrid FWE-corrected threshold of P < 3.29 × 10 -6 /3/3 = 3.65 × 10 -7 . The difference in the voxel-wise thresholds between GMV and ReHo was caused by the voxel size difference in the two parameter maps. Similarly, we extracted IDP values for ReHo from each participant. A two-sample t-test was also used to test the FA differences between matched OC and non-OC pairs. In these analyses, we corrected for the 87 white matter fiber tracts, three neuroimaging features, and three data types, resulting in an FWE-corrected threshold of P < 0.05/87/3/3 = 6.38 × 10 -5 . We also defined the fiber tracts with significant FA difference as IDPs for the subsequent analyses. For the special fiber tracts of interest that exhibited FA differences between OC and non-OC groups, we further calculated the FC between the connected brain regions of each fiber tract. A two-sample t-test was used to compare the FC difference between OC and non-OC. Comparisons of behavioral measures between OC and non-OC A two-sample t-test was applied to test the differences in behavioral variables between the well-matched OC and non-OC groups. We corrected for the three data types and 34 behavioral measurements by setting a Bonferroni-corrected threshold of P < 0.05/34/3 = 4.90 × 10 -4 . We defined behavioral traits with significant intergroup differences as significant behavioral phenotypes. Comparison of growth environmental exposures between OC and non-OC A two-sample t-test was applied to test the differences in the 32 growth environmental exposures between OC and non-OC. We corrected for three data types and 32 exposures by setting a Bonferroni-corrected threshold of P < 0.05/32/3 = 5.20 × 10 -4 . The growth environmental exposures with significant differences between OC and non-OC were defined as proximal environmental exposures (PEEs). PEE factor construction and comparison Due to the high correlations between PEEs, we performed exploratory factor analysis (EFA) implemented in the “psych” 2.2.9 (https://cran.r-project.org/package=psych) from R 4.1.2 to categorize PEEs into factors. Spearman correlations were calculated between PEEs and the appropriateness for EFA was confirmed by the KMO test and Bartlett’s test of sphericity. The criteria of very simple structure (VSS) ( 75 ) and minimum average partial (MAP) ( 76 ) were used to select the optimal number of factors. The principal axis factoring was used to identify a latent factor structure of PEEs, and varimax rotation was then employed to clarify the relationship among latent variables. Based on the optimal EFA model, we defined PEE indicators with factor loading > 0.45 as the main contributors for each PEE factor. We named each PEE factor according to the main contributors (PEEs) of the factor. To account for uncertainty in within-sample prediction, the 10-fold cross-validation confirmatory factor analysis (CFA) was performed to obtain the out-of-sample PEE factor scores. In each iteration, 90% participants were used to estimate factor loadings of the CFA model, and the obtained factor loadings were used to predict PEE factor scores of other 10% participants. This process was iterated 10 times to obtain the out-of-sample PEE factor scores of all participants. The model fit was evaluated by a combination of root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMR). To test the stability of CFA models, we calculated Spearman rank correlation between PEE factor scores estimated in the 10-fold cross-validation and those estimated using the full sample. A two-sample t-test was employed to compare the out-of-sample PEE factor scores between OC and non-OC ( P c < 0.05, adjusting for the number of PEE factors and three data types). Causal mediation analysis (CMA) CMA was implemented using the R package “mediation” 4.5.0 (https://cran.r-project.org/web/packages/mediation/index.html) ( 77 ) to test (a) whether a PEE factor mediated or moderated an association of GWS (treatment, X) with IDP or behavioral phenotype (outcome, Y); (b) whether an IDP mediated or moderated the association between GWS (treatment, X) and behavioral phenotype (outcome, Y). The direct, mediation and moderation effects were estimated using a quasi-Bayesian Monte Carlo simulation with 10,000 iterations. As IDPs were highly correlated, we estimated the effective number of IDPs ( 78 ). In (a), we corrected for the number of PEE factors and the sum of the effective number of IDPs and the number of behavioral variables. In (b), we corrected for the effective number of IDPs and the number of behavioral variables. Structural equation model (SEM) SEM was implemented using the R package “lavaan” 0.6-15 (https://cran.r-project.org/web/packages/lavaan/index.html) ( 79 ) based on the presumed effects of GWS and PEE factors on IDPs and behavioral phenotypes ( Supplementary Fig. 11 ). SEM can estimate effect while taking other predictors into account and allow for simultaneously estimating the associations among multiple variables while reducing the measurement error by using latent variables ( 79 ). SEM consists of two sub-models: the measurement model is used to define latent variables, and the structure model is used to specify direct and indirect pathways. A two-sided P < 0.05 was considered statistically significant in SEM. The model fit was evaluated by a combination of CFI, RMSEA, and SRMR. From SEM, we obtained the portion of the variance of each phenotype explained by the model, the relative effect of PEEs, as well as the direct, indirect, and total effects of GWS on outcomes. We integrated the results from CMA and SEM to identify consistent pathways from GWS to IDPs and behavioral phenotypes. As CMA used individual IDP while SEM used the latent variable of IDPs, we considered a causal pathway consistent between CMA and SEM when the individual IDP in CMA was one of the IDPs of the corresponding latent variable in SEM, and the indirect effect was in the same direction. Declarations Acknowledgements: Funding: This research was partially supported by the National Natural Science Foundation of China (grant 82030053 and 81425013 to Chunshui Yu, and grant 82202093 to Jie Tang). Authors contributions: Conceptualization: C.Y, J.T, J.Z, W.L, and M.W Methodology: C.Y, J.T, J-H.G, Z.G and W.Q Investigation: J.T, J.G, B.Z, W.Z, S,Q, G.C, Y.Y, W.L, H.Z, B.G, X.X, Y.Y, T.H, Z.Y, Q.Z, F.L, M.L, S.W, Q.X, J.X, J.F, Y.J, N.L, P.Z, D.S, C.W, S.L, Z.Y, F.C, W.S, W.M, D.W, J.X, X.Z, K.X, X-N.Z, L.Z, and Z.Y Visualization: J.T and W.Q Funding acquisition: C.Y, J.T, and W.L Project administration: C.Y Supervision: C.Y, J-H.G and W.Q Writing – original draft: J.T, J-H.G, and J.Z Writing – review & editing: C.Y, J-H.G and J.T Competing interests: The authors have declared that no competing interests exist. Data and materials availability: We made use of publicly available software and tools. The publicly available tools used in our analyses are described in the Methods section. The voxel-wise neuroimaging statistical maps are available at https://figshare.com/articles/dataset/Voxelwise_difference/24716832. Individual-level data will be shared with the researchers after their proposal has been approved by the committee. We encourage researchers or parties interested in collaboration for non-commercial use to email to C.Y. ( [email protected] ). All data requests will be reviewed by the committee of CHIMGEN to verify whether the request is subject to any intellectual property or confidentiality obligations. The use of data must comply with the requirements of the Human Genetics Resources Administration of China and other country-specific or region-specific regulations. References T. Hesketh, X. Zhou, Y. Wang, The End of the One-Child Policy: Lasting Implications for China. JAMA 314 , 2619-2620 (2015). Y. Wang, V. L. Fong, Little Emperors and the 4:2:1 Generation: China's Singletons. Journal of the American Academy of Child & Adolescent Psychiatry 48 , 1137-1139 (2009). T. Hesketh, L. Lu, Z. W. Xing, The effect of China's one-child family policy after 25 years. N Engl J Med 353 , 1171-1176 (2005). M. Hvistendahl, China. Making a selfish generation by fiat. Science 339 , 131 (2013). L. Cameron, N. Erkal, L. Gangadharan, X. Meng, Little emperors: behavioral impacts of China's One-Child Policy. Science 339 , 953-957 (2013). S. Cheng, C. Jia, Y. Wang, Only Children Were Associated with Anxiety and Depressive Symptoms among College Students in China. Int J Environ Res Public Health 17 , 4035 (2020). D. L. Poston, A. T. Falbo, Academic performance and personality traits of Chinese children: "Onlies" versus others. American Journal of Sociology 96 , 433-451 (1990). A. J. Fuligni, W. Zhang, Attitudes toward family obligation among adolescents in contemporary urban and rural China. Child Dev 75 , 180-192 (2004). Z.-y. Chen, R. X. Liu, Comparing adolescent only children with those who have siblings on academic related outcomes and psychosocial adjustment. Child Development Research 2014 , 1-10 (2014). W. Wang, J. Zhang, D. A. Hennessy, W. Yin, Do only-children communicate better than non-only children? A study of medical students in China. J. Comp. Fam. Stud 51 , 84–109 (2020). Y. Guo, J. Zhang, N. Zhang, Character strengths and their influencing factors among nursing students in Changsha, China: the only-child versus non-only-child. Arch Psychiatr Nurs 29 , 365-371 (2015). R. X. Liu, W. Lin, Z.-y. Chen, School performance, peer association, psychological and behavioral adjustments: A comparison between Chinese adolescents with and without siblings. Journal of Adolescence 33 , 411-417 (2010). C. Jia, Z. Yang, T. Xin, Y. Li, Y. Wang, T. Yang, Differences in school performance between only children and non-only children: evidence from China. Front Psychol 12 , 608704 (2021). N. Liu, Y. Chen, X. Yang, Y. Hu, Do demographic characteristics make differences? Demographic characteristics as moderators in the associations between only child status and cognitive/non-cognitive outcomes in China. Front Psychol 8 , 423 (2017). Y. Cao, L. Huang, T. Si, N. Q. Wang, M. Qu, X. Y. Zhang, The role of only-child status in the psychological impact of COVID-19 on mental health of Chinese adolescents. J Affect Disord 282 , 316-321 (2021). T. Falbo, Only Children: An Updated Review. Journal of Individual Psychology 68 , 38-49 (2012). X. Zhao, X. Ma, Y. Yao, C. Wan, E. Ng, China's little emperors show signs of success. Science 339 , 905-906 (2013). Y. Chen, F. Dangardt, P. Friberg, Association between childhood BMI trajectories and cardiometabolic risk and mental health problems at the age of 13 years: the cohort STudy of Adolescence Resilience and Stress (STARS). Lancet Glob Health 11 Suppl 1 , S3 (2023). E. Dennis, P. Manza, N. D. Volkow, Socioeconomic status, BMI, and brain development in children. Transl Psychiatry 12 , 33 (2022). T. Falbo, S. Y. Hooper, China's only children and psychopathology: A quantitative synthesis. Am J Orthopsychiatry 85 , 259-274 (2015). C. Liu, T. Munakata, F. N. Onuoha, Mental health condition of the only-child: a study of urban and rural high school students in China. Adolescence 40 , 831-845 (2005). D. Tomasi, N. D. Volkow, Associations of family income with cognition and brain structure in USA children: prevention implications. Mol Psychiatry 26 , 6619-6629 (2021). J. Yang, X. Hou, D. Wei, K. Wang, Y. Li, J. Qiu, Only-child and non-only-child exhibit differences in creativity and agreeableness: evidence from behavioral and anatomical structural studies. Brain Imaging Behav 11 , 493-502 (2017). Q. Xu, L. Guo, J. Cheng, M. Wang, Z. Geng, W. Zhu, B. Zhang, W. Liao, S. Qiu, H. Zhang, X. Xu, Y. Yu, B. Gao, T. Han, Z. Yao, G. Cui, F. Liu, W. Qin, Q. Zhang, M. J. Li, M. Liang, F. Chen, J. Xian, J. Li, J. Zhang, X. N. Zuo, D. Wang, W. Shen, Y. Miao, F. Yuan, S. Lui, X. Zhang, K. Xu, L. J. Zhang, Z. Ye, C. Yu, C. Consortium, CHIMGEN: a Chinese imaging genetics cohort to enhance cross-ethnic and cross-geographic brain research. Mol Psychiatry 25 , 517-529 (2020). J. P. Fortin, D. Parker, B. Tunc, T. Watanabe, M. A. Elliott, K. Ruparel, D. R. Roalf, T. D. Satterthwaite, R. C. Gur, R. E. Gur, R. T. Schultz, R. Verma, R. T. Shinohara, Harmonization of multi-site diffusion tensor imaging data. Neuroimage 161 , 149-170 (2017). F. C. Yeh, Population-based tract-to-region connectome of the human brain and its hierarchical topology. Nat Commun 13 , 4933 (2022). S. S. Panesar, J. T. A. Belo, F. C. Yeh, J. C. Fernandez-Miranda, Structure, asymmetry, and connectivity of the human temporo-parietal aslant and vertical occipital fasciculi. Brain Struct Funct 224 , 907-923 (2019). D. B. Downey, Number of siblings and intellectual development: The resource dilution explanation. American Psychologist 56 , 497-504 (2001). R. A. I. Bethlehem, J. Seidlitz, S. R. White, J. W. Vogel, K. M. Anderson, C. Adamson, S. Adler, G. S. Alexopoulos, E. Anagnostou, A. Areces-Gonzalez, D. E. Astle, B. Auyeung, M. Ayub, J. Bae, G. Ball, S. Baron-Cohen, R. Beare, S. A. Bedford, V. Benegal, F. Beyer, J. Blangero, M. Blesa Cabez, J. P. Boardman, M. Borzage, J. F. Bosch-Bayard, N. Bourke, V. D. Calhoun, M. M. Chakravarty, C. Chen, C. Chertavian, G. Chetelat, Y. S. Chong, J. H. Cole, A. Corvin, M. Costantino, E. Courchesne, F. Crivello, V. L. Cropley, J. Crosbie, N. Crossley, M. Delarue, R. Delorme, S. Desrivieres, G. A. Devenyi, M. A. Di Biase, R. Dolan, K. A. Donald, G. Donohoe, K. Dunlop, A. D. Edwards, J. T. Elison, C. T. Ellis, J. A. Elman, L. Eyler, D. A. Fair, E. Feczko, P. C. Fletcher, P. Fonagy, C. E. Franz, L. Galan-Garcia, A. Gholipour, J. Giedd, J. H. Gilmore, D. C. Glahn, I. M. Goodyer, P. E. Grant, N. A. Groenewold, F. M. Gunning, R. E. Gur, R. C. Gur, C. F. Hammill, O. Hansson, T. Hedden, A. Heinz, R. N. Henson, K. Heuer, J. Hoare, B. Holla, A. J. Holmes, R. Holt, H. Huang, K. Im, J. Ipser, C. R. Jack, Jr., A. P. Jackowski, T. Jia, K. A. Johnson, P. B. Jones, D. T. Jones, R. S. Kahn, H. Karlsson, L. Karlsson, R. Kawashima, E. A. Kelley, S. Kern, K. W. Kim, M. G. Kitzbichler, W. S. Kremen, F. Lalonde, B. Landeau, S. Lee, J. Lerch, J. D. Lewis, J. Li, W. Liao, C. Liston, M. V. Lombardo, J. Lv, C. Lynch, T. T. Mallard, M. Marcelis, R. D. Markello, S. R. Mathias, B. Mazoyer, P. McGuire, M. J. Meaney, A. Mechelli, N. Medic, B. Misic, S. E. Morgan, D. Mothersill, J. Nigg, M. Q. W. Ong, C. Ortinau, R. Ossenkoppele, M. Ouyang, L. Palaniyappan, L. Paly, P. M. Pan, C. Pantelis, M. M. Park, T. Paus, Z. Pausova, D. Paz-Linares, A. Pichet Binette, K. Pierce, X. Qian, J. Qiu, A. Qiu, A. Raznahan, T. Rittman, A. Rodrigue, C. K. Rollins, R. Romero-Garcia, L. Ronan, M. D. Rosenberg, D. H. Rowitch, G. A. Salum, T. D. Satterthwaite, H. L. Schaare, R. J. Schachar, A. P. Schultz, G. Schumann, M. Scholl, D. Sharp, R. T. Shinohara, I. Skoog, C. D. Smyser, R. A. Sperling, D. J. Stein, A. Stolicyn, J. Suckling, G. Sullivan, Y. Taki, B. Thyreau, R. Toro, N. Traut, K. A. Tsvetanov, N. B. Turk-Browne, J. J. Tuulari, C. Tzourio, E. Vachon-Presseau, M. J. Valdes-Sosa, P. A. Valdes-Sosa, S. L. Valk, T. van Amelsvoort, S. N. Vandekar, L. Vasung, L. W. Victoria, S. Villeneuve, A. Villringer, P. E. Vertes, K. Wagstyl, Y. S. Wang, S. K. Warfield, V. Warrier, E. Westman, M. L. Westwater, H. C. Whalley, A. V. Witte, N. Yang, B. Yeo, H. Yun, A. Zalesky, H. J. Zar, A. Zettergren, J. H. Zhou, H. Ziauddeen, A. Zugman, X. N. Zuo, B. R, Aibl, I. Alzheimer's Disease Neuroimaging, I. Alzheimer's Disease Repository Without Borders, C. Team, C. A. N. Cam, Ccnp, Cobre, cVeda, E. D. B. A. W. Group, P. Developing Human Connectome, FinnBrain, S. Harvard Aging Brain, Imagen, Kne, A. Mayo Clinic Study of, Nspn, Pond, P.-A. R. Group, Vetsa, E. T. Bullmore, A. F. Alexander-Bloch, Brain charts for the human lifespan. Nature 604 , 525-533 (2022). T. Paus, M. Keshavan, J. N. Giedd, Why do many psychiatric disorders emerge during adolescence? Nat Rev Neurosci 9 , 947-957 (2008). E. R. Sowell, P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, A. W. Toga, Longitudinal mapping of cortical thickness and brain growth in normal children. J Neurosci 24 , 8223-8231 (2004). D. A. Lewis, Development of the prefrontal cortex during adolescence: insights into vulnerable neural circuits in schizophrenia. Neuropsychopharmacology 16 , 385-398 (1997). L. C. Rutherford, S. B. Nelson, G. G. Turrigiano, BDNF has opposite effects on the quantal amplitude of pyramidal neuron and interneuron excitatory synapses. Neuron 21 , 521-530 (1998). S. Whittle, N. Vijayakumar, J. G. Simmons, M. Dennison, O. Schwartz, C. Pantelis, L. Sheeber, M. L. Byrne, N. B. Allen, Role of Positive Parenting in the Association Between Neighborhood Social Disadvantage and Brain Development Across Adolescence. JAMA Psychiatry 74 , 824-832 (2017). S. Whittle, J. G. Simmons, M. Dennison, N. Vijayakumar, O. Schwartz, M. B. Yap, L. Sheeber, N. B. Allen, Positive parenting predicts the development of adolescent brain structure: a longitudinal study. Dev Cogn Neurosci 8 , 7-17 (2014). J. Xu, X. Liu, Q. Li, R. Goldblatt, W. Qin, F. Liu, C. Chu, Q. Luo, A. Ing, L. Guo, N. Liu, H. Liu, C. Huang, J. Cheng, M. Wang, Z. Geng, W. Zhu, B. Zhang, W. Liao, S. Qiu, H. Zhang, X. Xu, Y. Yu, B. Gao, T. Han, G. Cui, F. Chen, J. Xian, J. Li, J. Zhang, X. N. Zuo, D. Wang, W. Shen, Y. Miao, F. Yuan, S. Lui, X. Zhang, K. Xu, L. Zhang, Z. Ye, T. Banaschewski, G. J. Barker, A. L. W. Bokde, H. Flor, A. Grigis, H. Garavan, P. Gowland, A. Heinz, R. Brühl, J. L. Martinot, E. Artiges, F. Nees, D. P. Orfanos, H. Lemaitre, T. Paus, L. Poustka, L. Robinson, S. Hohmann, J. H. Fröhner, M. N. Smolka, H. Walter, R. Whelan, J. Winterer, K. Patrick, V. Calhoun, M. J. Li, M. Liang, P. Gong, E. D. Barker, N. Clinton, A. Marquand, L. Yu, C. Yu, G. Schumann, Global urbanicity is associated with brain and behaviour in young people. Nat Hum Behav 6 , 279-293 (2022). L. Machlin, K. A. McLaughlin, M. A. Sheridan, Brain structure mediates the association between socioeconomic status and attention-deficit/hyperactivity disorder. Dev Sci 23 , e12844 (2020). Y. C. Lu, K. Kapse, N. Andersen, J. Quistorff, C. Lopez, A. Fry, J. Cheng, N. Andescavage, Y. Wu, K. Espinosa, G. Vezina, A. du Plessis, C. Limperopoulos, Association between socioeconomic status and in utero fetal brain development. JAMA Netw Open 4 , e213526 (2021). A. J. Dufford, G. W. Evans, J. Dmitrieva, J. E. Swain, I. Liberzon, P. Kim, Prospective associations, longitudinal patterns of childhood socioeconomic status, and white matter organization in adulthood. Hum Brain Mapp 41 , 3580-3593 (2020). D. Zhao, Z. He, Y. Tian, H. Liu, Differences in cognitive and non-cognitive results between only-child and non-only-child children: analysis of popensity scores based on large-scale assessment. Children 9 , 807 (2022). F. Van Overwalle, K. Baetens, P. Marien, M. Vandekerckhove, Social cognition and the cerebellum: a meta-analysis of over 350 fMRI studies. Neuroimage 86 , 554-572 (2014). D. Reiss, J. M. Ganiban, L. D. Leve, J. M. Neiderhiser, D. S. Shaw, M. N. Natsuaki, Parenting in the Context of the Child: Genetic and Social Processes. Monogr Soc Res Child Dev 87 , 7-188 (2022). Z. H. He, M. D. Li, X. Y. Ma, C. J. Liu, Family socioeconomic status and social media addiction in female college students: the mediating role of impulsiveness and inhibitory control. J Genet Psychol 182 , 60-74 (2021). J. C. Rajapakse, J. N. Giedd, J. L. Rapoport, Statistical approach to segmentation of single-channel cerebral MR images. IEEE Trans Med Imaging 16 , 176-186 (1997). J. Ashburner, A fast diffeomorphic image registration algorithm. Neuroimage 38 , 95-113 (2007). S. M. Smith, M. Jenkinson, H. Johansen-Berg, D. Rueckert, T. E. Nichols, C. E. Mackay, K. E. Watkins, O. Ciccarelli, M. Z. Cader, P. M. Matthews, T. E. Behrens, Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31 , 1487-1505 (2006). Y. Chao-Gan, Z. Yu-Feng, DPARSF: A MATLAB Toolbox for "Pipeline" Data Analysis of Resting-State fMRI. Front Syst Neurosci 4 , 13 (2010). M. Jenkinson, P. Bannister, M. Brady, S. Smith, Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 17 , 825-841 (2002). R. H. R. Pruim, M. Mennes, D. van Rooij, A. Llera, J. K. Buitelaar, C. F. Beckmann, ICA-AROMA: A robust ICA-based strategy for removing motion artifacts from fMRI data. Neuroimage 112 , 267-277 (2015). D. C. Delis, J. Freeland, J. H. Kramer, E. Kaplan, Integrating clinical assessment with cognitive neuroscience: construct validation of the California Verbal Learning Test. J Consult Clin Psychol 56 , 123-130 (1988). A. M. Owen, K. M. McMillan, A. R. Laird, E. Bullmore, N-back working memory paradigm: a meta-analysis of normative functional neuroimaging studies. Hum Brain Mapp 25 , 46-59 (2005). M. S. Shin, S. Y. Park, S. R. Park, S. H. Seol, J. S. Kwon, Clinical and empirical applications of the Rey-Osterrieth Complex Figure Test. Nat Protoc 1 , 892-899 (2006). Y. XIN, MCCB Chinese norm Manual(Chinese Edition) . (Peking University Medical Press, 2014). J. N. Kaufman, T. J. Ross, E. A. Stein, H. Garavan, Cingulate hypoactivity in cocaine users during a GO-NOGO task as revealed by event-related functional magnetic resonance imaging. J Neurosci 23 , 7839-7843 (2003). S. Pulos, Elison, J., Lennon, R.J, The hierarchical structure of the Interpersonal Reactivity Index. Soc Behav Pers 32 , 16 (2004). A. Canevello, J. Crocker, Creating good relationships: responsiveness, relationship quality, and interpersonal goals. J Pers Soc Psychol 99 , 78-106 (2010). X. Wang, Y. Wang, T. Xin, The Psychometric Properties of the Chinese Version of the Beck Depression Inventory-II With Middle School Teachers. Front Psychol 11 , 548965 (2020). D. T. Shek, The Chinese version of the State-Trait Anxiety Inventory: its relationship to different measures of psychological well-being. J Clin Psychol 49 , 349-358 (1993). D. Watson, L. A. Clark, A. Tellegen, Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol 54 , 1063-1070 (1988). K. T. Cohen S, Mermelstein R, A Global Measure of Perceived Stress. J Health Soc Behav 24 , 12 (1983). W. Pavot, & Diener, E, The Satisfaction With Life Scale and the emerging construct of life satisfaction. Journal of Positive Psychology 3 , 16 (2008). P. T. Costa Jr, R. R. McCrae, in The SAGE handbook of personality theory and assessment, Vol 2: Personality measurement and testing. (Sage Publications, Inc, Thousand Oaks, CA, US, 2008), pp. 179-198. C. R. Cloninger, T. R. Przybeck, D. M. Svrakic, The Tridimensional Personality Questionnaire: U.S. normative data. Psychol Rep 69 , 1047-1057 (1991). J. H. Patton, M. S. Stanford, E. S. Barratt, Factor structure of the Barratt impulsiveness scale. J Clin Psychol 51 , 768-774 (1995). V. Rolfe, Tailoring a measurement model of socioeconomic status: Applying the alignment optimization method to 15 years of PISA. International Journal of Educational Research 106 , 101723 (2021). A. van Donkelaar, R. V. Martin, M. Brauer, N. C. Hsu, R. A. Kahn, R. C. Levy, A. Lyapustin, A. M. Sayer, D. M. Winker, Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors. Environmental Science & Technology 50 , 3762-3772 (2016). J. A. Geddes, R. V. Martin, B. L. Boys, A. van Donkelaar, Long-Term Trends Worldwide in Ambient NO2 Concentrations Inferred from Satellite Observations. Environ Health Perspect 124 , 281-289 (2016). G. Parker, H. Tupling, L. B. Brown, A parental bonding instrument. British Journal of Medical Psychology 52 , 1-10 (1979). J. Canty-Mitchell, G. D. Zimet, Psychometric properties of the Multidimensional Scale of Perceived Social Support in urban adolescents. Am J Community Psychol 28 , 391-400 (2000). X. F. Zhao, Y. L. Zhang, L. F. Li, Y. F. Zhou, H. Z. Li, S. C. Yang, Reliability and validity of the Chinese version of childhood trauma questionnaire. Chinese Journal of Clinical Rehabilitation 9 , 105-107 (2005). N. Liu, L. Zhang, T. Tian, J. Cheng, B. Zhang, S. Qiu, Z. Geng, G. Cui, Q. Zhang, W. Liao, Y. Yu, H. Zhang, B. Gao, X. Xu, T. Han, Z. Yao, W. Qin, F. Liu, M. Liang, Q. Xu, J. Fu, J. Xu, W. Zhu, P. Zhang, W. Li, D. Shi, C. Wang, S. Lui, Z. Yan, F. Chen, J. Li, J. Zhang, D. Wang, W. Shen, Y. Miao, J. Xian, J. H. Gao, X. Zhang, M. J. Li, K. Xu, X. N. Zuo, M. Wang, Z. Ye, C. Yu, C. Consortium, Cross-ancestry genome-wide association meta-analyses of hippocampal and subfield volumes. Nat Genet 55 , 1126-1137 (2023). E. A. Stuart, G. King, K. Imai, D. Ho, MatchIt: nonparametric preprocessing for parametric causal inference. Journal of statistical software 42 , 1-28 (2011). P. C. Austin, Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 28 , 3083-3107 (2009). K. J. Worsley, J. E. Taylor, F. Tomaiuolo, J. Lerch, Unified univariate and multivariate random field theory. Neuroimage 23 Suppl 1 , S189-195 (2004). W. F. Velicer, Determining the number of components from the matrix of partial correlations. Psychometrika 41 , 321-327 (1976). W. Revelle, T. Rocklin, Very simple structure: an alternative procedure for estimating the optimal number of interpretable factors. Multivariate Behav Res 14 , 403-414 (1979). D. Tingley, T. Yamamoto, K. Hirose, L. Keele, K. Imai, mediation: R Package for Causal Mediation Analysis. Journal of Statistical Software 59 , 1-38 (2014). C. S. Bretherton, M. Widmann, V. P. Dymnikov, J. M. Wallace, I. Bladé, The effective number of spatial degrees of freedom of a time-varying field. Journal of Climate 12 , 1990-2009 (1999). Y. Rosseel, lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software 48 , 1-36 (2012). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryTables.xlsx Supplementary Tables SupplementaryFigures.docx Supplementary Figures Cite Share Download PDF Status: Published Journal Publication published 31 Mar, 2025 Read the published version in Nature Human Behaviour → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3707132","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Biological Sciences - Article","associatedPublications":[],"authors":[{"id":257627510,"identity":"e1d5b62d-41a5-46d2-8018-1c4a51aa80fd","order_by":0,"name":"Chunshui Yu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA00lEQVRIiWNgGAWjYDACCSBmbGDgYWBgPgAROUC8FrbEBpK0AAGPIXFa+Gf3mD38ueOwjMGNnO+PbrYxyPHdSGD8XIDPkjtnzI15zxzmMbiRu7E5t43BWPJGArP0DDxaDCRyzKQZ2xBaEjfcSGBj5iGgRfInWEvOQ5CWeqK0SPBCtDCCtCQYENIicSOtTJq3LZ1H8swzw9k55yQMZ5552CyNTwv/jORtQIdZ2/MdT37wOafMRh7IOPgZnxY4UDgAsZUBFk2EgTyR6kbBKBgFo2AEAgCXqEy+lZ+xdAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-5648-5199","institution":"Tianjin Medical University General Hospital","correspondingAuthor":true,"prefix":"","firstName":"Chunshui","middleName":"","lastName":"Yu","suffix":""},{"id":257627511,"identity":"47255a80-1670-4b6c-8a76-e9c9c5b133a9","order_by":1,"name":"Jie Tang","email":"","orcid":"","institution":"Laboratory of Functional Imaging, Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Tang","suffix":""},{"id":257627512,"identity":"aa3ff321-7a41-43a4-a226-fe72aad87917","order_by":2,"name":"Jing Zhang","email":"","orcid":"","institution":"Lanzhou University Second Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jing","middleName":"","lastName":"Zhang","suffix":""},{"id":257627513,"identity":"0e6d3702-8ac8-47ba-818e-2680608b9a52","order_by":3,"name":"Wei Li","email":"","orcid":"","institution":"Tianjin Medical University Cancer Institute and Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Li","suffix":""},{"id":257627514,"identity":"99b16462-903e-4777-9590-8a4ce4f1e171","order_by":4,"name":"Meiyun Wang","email":"","orcid":"https://orcid.org/0000-0003-4454-5005","institution":"Henan Provincial People's Hospital \u0026 the People's Hospital of Zhengzhou University","correspondingAuthor":false,"prefix":"","firstName":"Meiyun","middleName":"","lastName":"Wang","suffix":""},{"id":257627515,"identity":"5bf3382c-6006-4175-938f-9c9b49ea543d","order_by":5,"name":"Jingliang Cheng","email":"","orcid":"","institution":"The First Affiliated Hospital of Zhengzhou University","correspondingAuthor":false,"prefix":"","firstName":"Jingliang","middleName":"","lastName":"Cheng","suffix":""},{"id":257627516,"identity":"f066b3c9-31fa-4715-93a1-887b1ae1e5f2","order_by":6,"name":"Bing Zhang","email":"","orcid":"","institution":"Drum Tower Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Bing","middleName":"","lastName":"Zhang","suffix":""},{"id":257627517,"identity":"b04c1d69-cb4f-47d0-99a2-7083bdc8c535","order_by":7,"name":"Wenzhen Zhu","email":"","orcid":"","institution":"Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Wenzhen","middleName":"","lastName":"Zhu","suffix":""},{"id":257627518,"identity":"c9ba9703-0daf-4880-9342-4ba6d1f7a908","order_by":8,"name":"Shijun Qiu","email":"","orcid":"","institution":"The First Affiliated Hospital of Guangzhou University of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shijun","middleName":"","lastName":"Qiu","suffix":""},{"id":257627519,"identity":"ae51f3a1-35e5-4fb8-a802-0b74772c4f42","order_by":9,"name":"Guangbin Cui","email":"","orcid":"","institution":"Tangdu Hospital, Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Guangbin","middleName":"","lastName":"Cui","suffix":""},{"id":257627520,"identity":"f113be55-c9a4-468e-82aa-1900320148b3","order_by":10,"name":"Yongqiang Yu","email":"","orcid":"","institution":"The First Affiliated Hospital of Anhui Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yongqiang","middleName":"","lastName":"Yu","suffix":""},{"id":257627521,"identity":"4a49870f-f98a-413b-95f0-5b0416f88fd1","order_by":11,"name":"Weihua Liao","email":"","orcid":"","institution":"Xiangya Hospital, Central South University","correspondingAuthor":false,"prefix":"","firstName":"Weihua","middleName":"","lastName":"Liao","suffix":""},{"id":257627522,"identity":"eaca703f-e983-4e59-b0ea-8857381baa8e","order_by":12,"name":"Hui Zhang","email":"","orcid":"","institution":"The First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Zhang","suffix":""},{"id":257627523,"identity":"179c34b1-63a1-4906-915d-545bdebeaccc","order_by":13,"name":"Bo Gao","email":"","orcid":"","institution":"The Affiliated Hospital of Guizhou Medical University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Gao","suffix":""},{"id":257627524,"identity":"c8f23b69-3e0e-488c-a796-567a5a65eb35","order_by":14,"name":"Xiaojun Xu","email":"","orcid":"","institution":"The Second Affiliated Hospital of Zhejiang University, School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Xiaojun","middleName":"","lastName":"Xu","suffix":""},{"id":257627525,"identity":"ee9c178d-8f25-4b3b-b16a-69218476b2fb","order_by":15,"name":"Yunjun Yang","email":"","orcid":"","institution":"Department of Radiology, The First Affiliated Hospital of Wenzhou Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yunjun","middleName":"","lastName":"Yang","suffix":""},{"id":257627526,"identity":"c0b0059d-84fa-4744-b8a8-ccc2b02e3d6f","order_by":16,"name":"Tong Han","email":"","orcid":"","institution":"Tianjin Huanhu Hospital","correspondingAuthor":false,"prefix":"","firstName":"Tong","middleName":"","lastName":"Han","suffix":""},{"id":257627527,"identity":"22b6e14f-703e-474f-bb71-eda0b524dcef","order_by":17,"name":"Zhenwei Yao","email":"","orcid":"","institution":"Huashan Hospital, Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Zhenwei","middleName":"","lastName":"Yao","suffix":""},{"id":257627528,"identity":"c885a686-5772-42ab-a98c-1afa95f3414e","order_by":18,"name":"Quan Zhang","email":"","orcid":"","institution":"Characteristic Medical Center of Chinese People's Armed Police Force","correspondingAuthor":false,"prefix":"","firstName":"Quan","middleName":"","lastName":"Zhang","suffix":""},{"id":257627529,"identity":"3f4587bc-7a93-4216-8dba-371c47b0b34d","order_by":19,"name":"Wen Qin","email":"","orcid":"https://orcid.org/0000-0002-9121-8296","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wen","middleName":"","lastName":"Qin","suffix":""},{"id":257627530,"identity":"c53a9559-548f-4415-a270-33e579806d06","order_by":20,"name":"Feng Liu","email":"","orcid":"https://orcid.org/0000-0002-3570-4222","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Liu","suffix":""},{"id":257627531,"identity":"6f4a0b59-4125-4e92-9dd8-fb337e270148","order_by":21,"name":"Meng Liang","email":"","orcid":"","institution":"Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Meng","middleName":"","lastName":"Liang","suffix":""},{"id":257627532,"identity":"5c6f9ef3-9a9c-4083-bf8e-f81c44f91ada","order_by":22,"name":"Sijia Wang","email":"","orcid":"https://orcid.org/0000-0001-8949-303X","institution":"School of Medical Imaging and Tianjin Key Laboratory of Functional Imaging, Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Sijia","middleName":"","lastName":"Wang","suffix":""},{"id":257627533,"identity":"c0429f49-c56e-45b6-a0bb-6951fd6c15f9","order_by":23,"name":"Qiang Xu","email":"","orcid":"","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Xu","suffix":""},{"id":257627534,"identity":"1b162994-2bff-43bd-9399-199ed12a5316","order_by":24,"name":"Jiayuan Xu","email":"","orcid":"","institution":"Department of Radiology and Tianjin Key Laboratory of Functional Imaging, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jiayuan","middleName":"","lastName":"Xu","suffix":""},{"id":257627535,"identity":"28042683-99e3-4667-8a0a-241e63eafe10","order_by":25,"name":"Jilian Fu","email":"","orcid":"","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jilian","middleName":"","lastName":"Fu","suffix":""},{"id":257627536,"identity":"07b72f91-aeaf-4463-bf9b-c72b0a3841eb","order_by":26,"name":"Yuan Ji","email":"","orcid":"https://orcid.org/0000-0003-2874-5904","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yuan","middleName":"","lastName":"Ji","suffix":""},{"id":257627537,"identity":"dbb3a188-8d12-4f32-a04e-5ad41c326d17","order_by":27,"name":"Nana Liu","email":"","orcid":"","institution":"Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Nana","middleName":"","lastName":"Liu","suffix":""},{"id":257627538,"identity":"08273038-5a74-4ed6-8aea-1ce2bf9e2723","order_by":28,"name":"Peng Zhang","email":"","orcid":"","institution":"Tianjin Medical University Cancer Institute and Hospital","correspondingAuthor":false,"prefix":"","firstName":"Peng","middleName":"","lastName":"Zhang","suffix":""},{"id":257627539,"identity":"a125c0d6-83ef-466c-bbaf-05fbe003b006","order_by":29,"name":"Dapeng Shi","email":"","orcid":"","institution":"Henan Provincial People's Hospital \u0026 Zhengzhou University People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Dapeng","middleName":"","lastName":"Shi","suffix":""},{"id":257627540,"identity":"26146dce-437e-453f-a7b6-5a2c75118753","order_by":30,"name":"Caihong Wang","email":"","orcid":"","institution":"The First Affiliated Hospital of Zhengzhou University","correspondingAuthor":false,"prefix":"","firstName":"Caihong","middleName":"","lastName":"Wang","suffix":""},{"id":257627541,"identity":"98b362ca-d8f4-4267-96ab-8e9461493ffe","order_by":31,"name":"Su Lui","email":"","orcid":"","institution":"Department of Radiology, the Center for Medical Imaging, West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Su","middleName":"","lastName":"Lui","suffix":""},{"id":257627542,"identity":"35d68f45-fd10-4d5e-b793-922eb7451481","order_by":32,"name":"Zhihan Yan","email":"","orcid":"","institution":"The Second Affiliated Hospital and Yuying Children's Hospital of Wenzhou Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhihan","middleName":"","lastName":"Yan","suffix":""},{"id":257627543,"identity":"773233c9-51da-477a-88a9-89de9ee8a154","order_by":33,"name":"Feng Chen","email":"","orcid":"https://orcid.org/0000-0002-9129-7895","institution":"Hainan General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Chen","suffix":""},{"id":257627544,"identity":"e43af2d9-8b0b-4ffe-bbeb-a02068d1210e","order_by":34,"name":"Wen Shen","email":"","orcid":"","institution":"Tianjin First Center Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wen","middleName":"","lastName":"Shen","suffix":""},{"id":257627545,"identity":"12ac9454-2a79-4d98-8443-beffe048872d","order_by":35,"name":"Yanwei Miao","email":"","orcid":"","institution":"The First Affiliated Hospital of Dalian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yanwei","middleName":"","lastName":"Miao","suffix":""},{"id":257627546,"identity":"dab849cb-ec00-4d76-b9a9-f35c98f28daa","order_by":36,"name":"Dawei Wang","email":"","orcid":"","institution":"Qilu Hospital of Shandong University","correspondingAuthor":false,"prefix":"","firstName":"Dawei","middleName":"","lastName":"Wang","suffix":""},{"id":257627547,"identity":"6d5cdc1f-d757-4a88-b970-c8900354e01a","order_by":37,"name":"Junfang Xian","email":"","orcid":"","institution":"Beijing Tongren Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junfang","middleName":"","lastName":"Xian","suffix":""},{"id":257627548,"identity":"ec768c0b-232f-4362-af1a-add729914acb","order_by":38,"name":"Xiaochu Zhang","email":"","orcid":"https://orcid.org/0000-0002-7541-0130","institution":"University of Science and Technology of China","correspondingAuthor":false,"prefix":"","firstName":"Xiaochu","middleName":"","lastName":"Zhang","suffix":""},{"id":257627549,"identity":"b8025ba6-1828-40a0-a01b-da5a4271d089","order_by":39,"name":"Kai Xu","email":"","orcid":"","institution":"The Affiliated Hospital of Xuzhou Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kai","middleName":"","lastName":"Xu","suffix":""},{"id":257627550,"identity":"9cb244e5-93a5-4629-a32e-fed763f5b5fc","order_by":40,"name":"Xi-Nian Zuo","email":"","orcid":"https://orcid.org/0000-0001-9110-585X","institution":"Beijing Normal University","correspondingAuthor":false,"prefix":"","firstName":"Xi-Nian","middleName":"","lastName":"Zuo","suffix":""},{"id":257627551,"identity":"923fea31-6f78-499d-8d7d-5c24ff4ea6d6","order_by":41,"name":"Long Jiang Zhang","email":"","orcid":"https://orcid.org/0000-0002-6664-7224","institution":"Jinling hospital, Affiliated Hospital of Medical School, Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Long","middleName":"Jiang","lastName":"Zhang","suffix":""},{"id":257627552,"identity":"7ba8adf4-90b2-4a8d-9e74-1390491227af","order_by":42,"name":"Zhaoxiang Ye","email":"","orcid":"https://orcid.org/0000-0003-3157-8393","institution":"Tianjin Medical University Cancer Institute and Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhaoxiang","middleName":"","lastName":"Ye","suffix":""},{"id":257627553,"identity":"1b62fe89-ea1a-4a93-8692-5fcd8496fac9","order_by":43,"name":"Zuojun Geng","email":"","orcid":"","institution":"Department of Medical Imaging, The Second Hospital of Hebei Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zuojun","middleName":"","lastName":"Geng","suffix":""},{"id":257627556,"identity":"71f4e55f-0aea-4688-b437-fb490aa07ada","order_by":44,"name":"Jia-Hong Gao","email":"","orcid":"https://orcid.org/0000-0002-9311-0297","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jia-Hong","middleName":"","lastName":"Gao","suffix":""},{"id":257627557,"identity":"5e98e728-e3cb-4423-b967-a5808f658496","order_by":45,"name":"The CHIMGEN Consortium","email":"","orcid":"","institution":"Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"The","middleName":"CHIMGEN","lastName":"Consortium","suffix":""}],"badges":[],"createdAt":"2023-12-05 01:21:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3707132/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3707132/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41562-025-02142-4","type":"published","date":"2025-03-31T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":49237998,"identity":"cc9a6534-55f8-4c55-8372-58efba391a38","added_by":"auto","created_at":"2024-01-05 18:10:10","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":542772,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eBrain imaging and behavioral differences between participants with and without siblings.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ea-c\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). Compared with non-OC, OC show smaller\u003c/em\u003e \u003cem\u003eGMV in the MFPC, ITG_L, and OFC_R, and larger GMV in the bilateral CPL and CV4-5 (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ea\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e), higher FA in language-related fibers (PAT_R) and lower FA in motor-related fibers (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e), and\u003c/em\u003e \u003cem\u003elower ReHo in the bilateral ITG and MOFC (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e) at an FWE-corrected threshold of P\u003c/em\u003e\u003csub\u003e\u003cem\u003ec\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e \u0026lt; 0.05/3/3. Color depicts Cohen’s d value of each IDP. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ed\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). Behavioral differences between OC and non-OC. The y-axis shows each behavioral phenotype, and the x-axis shows -log\u003c/em\u003e\u003csub\u003e\u003cem\u003e10\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e(P) values. The dashed line corresponds to a Bonferroni-corrected threshold of P = 4.90×10\u003c/em\u003e\u003csup\u003e\u003cem\u003e-4\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e. Dot sizes and color depict Cohen’s d. Abbreviations:\u003c/em\u003e \u003cem\u003eBDI, depression; BFI, big five inventory; BFI-A,\u003c/em\u003e a\u003cem\u003egreeableness; BFI-C, conscientiousness; BFI-E, extraversion;\u003c/em\u003e \u003cem\u003eBFI-N, neuroticism; BFI-O, openness; BIS, Barratt impulsiveness scale; BIS-CI, cognitive impulsiveness; BIS-MI, motor impulsiveness; BIS-NPI, no-planning impulsiveness; CBT, corticobulbar tract; CPL, cerebellar posterior lobe; CST, corticospinal tract; CV, cerebellar vermis; DRTT, dentato-rubro-thalamic tract; EC, executive control; FA, fractional anisotropy; GMV, gray matter volume; IFR, immediate free recall; IGS, interpersonal goals scale; IGS-C, compassionate; IGS-SI, self-image goals; IRI, interpersonal reactivity index; IRI-EMC, empathy concern; IRI-F, fantasy; IRI-PD, personal distress; IRI-PT, perspective-taking; ITG, inferior temporal gyrus; L, left; LDCR, number of correct words for long delayed cued recall; LDFR, number of correct words for long delayed free recall; MFPC, medial frontal and parietal cortices; MOFC, medial orbito-frontal cortex; OFC, orbitofrontal cortex; PAT, parietal aslant tract; PNAS, the positive and negative affect scale; PNAS-NA, negative affect; PNAS-PA, positive affect; PSS, perceived stress; R, right; ReHo, regional homogeneity; RST, reticulospinal tract; RO, Rey-Osterrieth complex figure test; RO-IR,\u003c/em\u003e \u003cem\u003eimmediate recall;\u003c/em\u003e \u003cem\u003eRO-DR, delay recall; SAI, state anxiety; SCP, superior cerebellar peduncle; SDCR, number of correct words for short delayed cued recall; SDFR, number of correct words for short delayed free recall; SDMT, correct number of digits; SWL, satisfaction with life;\u003c/em\u003e \u003cem\u003eTAI, trait anxiety; TPQ, tridimensional personality questionnaire; TPQ-HA, harm avoidance; TPQ-NS, novelty seeking; TPQ-RD, reward dependence; WM, working memory.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Binder11.png","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/0bbce73d591e314c228def02.png"},{"id":49237994,"identity":"899d1d9f-0d6f-43bd-b87e-b8d805ab1ae0","added_by":"auto","created_at":"2024-01-05 18:10:10","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":88316,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003ePEE and PEE factor score differences between participants with and without siblings.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ea\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). Intergroup differences in 32 early-life environmental exposures between OC and non-OC. The y-axis shows each environmental exposure, and the x-axis shows -log\u003c/em\u003e\u003csub\u003e\u003cem\u003e10\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e(P) values. The dashed line corresponds to the Bonferroni-corrected threshold of P = 5.20×10\u003c/em\u003e\u003csup\u003e\u003cem\u003e-4\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e. Dot sizes and color depict Cohen’s d. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). PEE factor loadings in the five-factor EFA model. For each factor, PEEs with factor loading \u0026gt; 0.45 are used to calculate the out-of-sample PEE factor scores for each participant by conducting 10-fold-cross validation CFA. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). PEE factor differences between OC and non-OC (all P \u0026lt; 0.05/5/3 = 3.33 × 10\u003c/em\u003e\u003csup\u003e\u003cem\u003e-3\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e, Bonferroni corrected); number above each line represents the Cohen’s d. Violin plot depicts density distribution of the data, and the black point and error bar represent for mean and standard deviation. OC show higher FSES, CSES, AP, SC scores and lower ASES scores than non-OC. Abbreviations: AP, air pollution; ASES, adverse socioeconomic status; CFA, confirmatory factor analysis; CS, community safety; CSES, city-level socioeconomic status; EA, emotion abuse; EE, education expense; EFA, exploratory factor analysis; EN, emotion neglect; FAS, family support; FIC, financial crisis; FID, financial difficulty; FR, family resources; FRS, friend support; FSES, family socioeconomic status; FUS, family unemployment stress; GDP, gross domestic product; HB, hospital beds; HI, home inadequacy; IAQ, indoor air quality; ITN, income to need; MCA, maternal care; MCO, maternal control; ME, maternal education; MEN, maternal encouragement; MO, maternal occupation; NI, neighbor indifference; NO\u003c/em\u003e\u003csub\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e, nitrogen dioxide; OC, only children; OS, other support; PA, physical abuse; PCA, paternal care; PCO, paternal control; PE, paternal education; PEE, proximal environmental exposure; PEN, paternal encouragement; PM\u003c/em\u003e\u003csub\u003e\u003cem\u003e2.5\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e, particulate matter 2.5; PN, physical neglect; PO, paternal occupation; SA, sexual abuse; SC, support and care.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Binder12.png","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/a4e9b73041cee565923d8758.png"},{"id":49237997,"identity":"0930ec2d-dee8-4b8b-8f5a-6601fc9a3894","added_by":"auto","created_at":"2024-01-05 18:10:10","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":112364,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMediation and moderation of PEE factors and IDPs on GWS effects.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e CMA is conducted to test the mediation and moderation of five PEE factors on the associations of GWS with IDPs and behavioral phenotypes and of IDPs on the associations of GWS with behavioral phenotypes. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ea\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). CMA mediation model. This model tests direct and indirect effects of GWS on phenotype with GWS as treatment (X), PEE factor or IDP as mediator (M), and IDP or behavioral phenotype as outcome (Y). (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). CMA moderation model. A moderation is considered significant if the effect of GWS on phenotype differs between the mediator levels. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec-d\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). Mediation (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e) and moderation (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ed\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e) effects of PEE factors on the associations of GWS with IDPs and behavioral phenotypes. Circos heatmaps show regression coefficient (β) for mediation and moderation effect. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ee-f\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). Mediation (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ee\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e) and moderation (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ef\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e) effects of IDPs on the associations of GWS with behavioral phenotypes.\u003c/em\u003e \u003cem\u003eThe heatmaps show β for mediation and moderation effect. ** P\u003c/em\u003e\u003csub\u003e\u003cem\u003ec\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e \u0026lt; 0.05, Bonferroni corrected; and * P \u0026lt; 0.05, uncorrected. Abbreviations: AP, air pollution; ASES, adverse socioeconomic status; BFI-O, openness of big five inventory; BIS-CI, cognitive impulsiveness of Barratt impulsiveness scale; CBT, corticobulbar tract; CPL, cerebellar posterior lobe; CSES, city-level socioeconomic status; CST, corticospinal tract; CV, cerebellar vermis; DRTT, dentato-rubro-thalamic tract; EC, executive control; FA, fractional anisotropy; FSES, family socioeconomic status; GMV, gray matter volume; IFR, immediate free recall; IGS-SI, self-image goals of interpersonal goals scale; IRI, interpersonal reactivity index; IRI-PD, personal distress; IRI-PT, perspective-taking; ITG, inferior temporal gyrus; L, left; MFPC, medial frontal and parietal cortices; MOFC, medial orbito-frontal cortex; OFC, orbito-frontal cortex; PAT, parietal aslant tract; R, right; ReHo, regional homogeneity; RST, reticulospinal tract; SCP, superior cerebellar peduncle; SWL, satisfaction with life; TPQ, tridimensional personality questionnaire; TPQ-NS, novelty seeking; TPQ-RD, reward-dependence; WM, working memory.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Binder13.png","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/510c253c19dee4c57ab2e720.png"},{"id":49239936,"identity":"37fe60eb-bc4b-4aa1-aa01-f10034319d8f","added_by":"auto","created_at":"2024-01-05 18:18:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":117544,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eStructural equation model shows the relationships of GWS, PEEs, IDPs and behavioral phenotypes.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ea\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). The variance of 16 brain and behavioral phenotypes explained by the SEM. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003eb\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). A summary of significant relative effects of GWS and five PEE latent variables on IDPs and behavioral phenotypes in SEM. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ec\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). The final SEM. Only significant associations are shown. Ellipses and rectangles represent latent variables and observed variables, respectively. Numbers adjacent to arrows are standardized path coefficients, analogous to relative regression weights. Solid and dashed arrows represent positive and negative relationships, respectively. Line thickness is proportional to the strength of the association. (\u003c/em\u003e\u003cem\u003e\u003cstrong\u003ed\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e). A summary of significant direct and indirect effects of GWS on IDPs and behavioral phenotypes. Direct and indirect effects of distinct mediators are shown in different colors, and the sequential mediators in a pathway are connected by “-”. Abbreviations: AP, the latent variable of air pollution; ASES, the\u003c/em\u003e\u003csub\u003e\u003cem\u003e \u003c/em\u003e\u003c/sub\u003e\u003cem\u003elatent variable of adverse SES; BFI-O, openness of big five inventory; BIS-CI, cognitive impulsiveness of Barratt impulsiveness scale; CSES, the latent variable of city-level SES; EC, executive control; FA, fractional anisotropy; GMV, gray matter volume;\u003c/em\u003e \u003cem\u003eGWS, growth without siblings; IFR, immediate free recall; IGS-SI, self-image goals of interpersonal goals scale; IRI, interpersonal reactivity index; IRI-PD, personal distress; IRI-PT, perspective-taking; ITG, inferior temporal gyrus; LFA, low FA latent variable; LGMV, large GMV latent variable; LReHo, low ReHo latent variable; PAT_R, right parietal aslant tract; ReHo, regional homogeneity; SC, the latent variable of support and care; SEM, structural equation model; SES, socioeconomic status; SGMV, small GMV latent variable; SWL, satisfaction with life; TPQ, tridimensional personality questionnaire; TPQ-NS, novelty seeking; TPQ-RD, reward dependence; WM, working memory.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Binder14.png","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/92b58551b9be209589d80287.png"},{"id":49239937,"identity":"59773937-710f-48fa-ae0e-47453c810947","added_by":"auto","created_at":"2024-01-05 18:18:10","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":94432,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eConsistent causal pathways identified by both CMA and SEM.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e The graphs show consistent causal pathways identified by both SEM and CMA. As CMA uses individual IDP while SEM uses the latent variable of IDPs, we consider the causal pathways consistent in both CMA and SEM when the individual IDP in CMA is one of the IDPs of the corresponding latent variable in SEM, and the indirect effect is in the same direction. The full statistics of indirect effects of CMA and SEM are provided in \u003c/em\u003e\u003cem\u003e\u003cstrong\u003eSupplementary Table 9-11\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e. Ellipses and rectangles represent latent and observed variables, respectively. Red, blue, and gray arrows represent positive, negative, and non-significant relationships, respectively. Curved arrows with dashed and dotted lines represent the total and direct effects of GWS, respectively. Line thickness is proportional to the strength of association in SEM. Abbreviations: AP, the latent variable of air pollution; ASES, the\u003c/em\u003e\u003csub\u003e\u003cem\u003e \u003c/em\u003e\u003c/sub\u003e\u003cem\u003elatent variable of adverse SES; BFI-O, openness of big five inventory; BIS-CI, cognitive impulsiveness of Barratt impulsiveness scale; CMA, causal mediation analysis; CSES, the latent variable of city-level SES; EC, executive control; FA, fractional anisotropy; GMV, gray matter volume; GWS, growth without siblings; IFR, immediate free recall; IGS-SI, self-image goals of interpersonal goals scale; IRI, interpersonal reactivity index; IRI-PD, personal distress; IRI-PT, perspective-taking; ITG, inferior temporal gyrus; LFA, low FA latent variable; LGMV, large GMV latent variable; LReHo, low ReHo latent variable; PAT, parietal aslant tract; ReHo, regional homogeneity; SC, the latent variable of support and care; SEM, structural equation model; SES, socioeconomic status; SGMV, small GMV latent variable; SWL, satisfaction with life; TPQ, tridimensional personality questionnaire; TPQ-NS, novelty seeking; TPQ-RD, reward dependence; WM, working memory.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Binder15.png","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/16647874bb763e6a19897feb.png"},{"id":79646830,"identity":"62e4b555-59f6-4ac6-80a5-46365bb15a64","added_by":"auto","created_at":"2025-04-01 07:13:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2992275,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/c6ca3a0c-ce62-4ca9-8e13-fbe07c956c42.pdf"},{"id":49237996,"identity":"e908b749-3e74-4152-bea4-ea014af988f0","added_by":"auto","created_at":"2024-01-05 18:10:10","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":189310,"visible":true,"origin":"","legend":"Supplementary Tables","description":"","filename":"SupplementaryTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/f87ae2308d518e64eb38f6a7.xlsx"},{"id":49238000,"identity":"5b3822c6-74dc-4377-a0c8-53225be5a583","added_by":"auto","created_at":"2024-01-05 18:10:10","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":4482462,"visible":true,"origin":"","legend":"Supplementary Figures","description":"","filename":"SupplementaryFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-3707132/v1/79ce3bb3abbd9c957390d607.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"How growth without siblings affects adult brain and behavior","fulltext":[{"header":"Main Text","content":"\u003cp\u003eWith the surge of only-child families in the world, it is critical to understand how growth without siblings (GWS) affects the development of brain, behavior, and mental health (\u003cem\u003e1-4\u003c/em\u003e). However, existing studies examining the associations between GWS and these outcomes have yielded inconsistent and even contradictory results. For instance, some studies have reported that only children (OC) exhibit more problem behaviors (more pessimistic and risk-averse and less trustworthy, competitive, and conscientious) (\u003cem\u003e5\u003c/em\u003e), higher prevalence of anxiety and depression (\u003cem\u003e6\u003c/em\u003e), and worse academic outcomes (\u003cem\u003e7\u003c/em\u003e); some indicate comparable problem behaviors (\u003cem\u003e8-10\u003c/em\u003e), anxiety and depression (\u003cem\u003e11\u003c/em\u003e), and academic outcomes (\u003cem\u003e12\u003c/em\u003e); while others suggest better academic outcomes (\u003cem\u003e13\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e14\u003c/em\u003e), less anxiety and depression (\u003cem\u003e15\u003c/em\u003e), and more prosocial behaviors in OC individuals (\u003cem\u003e16\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e17\u003c/em\u003e). The divergent findings may arise from many reasons, such as the mismatched confounders (\u003cem\u003e18\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e19\u003c/em\u003e) between OC and non-OC groups. The negative outcomes observed in OC are often attributed to the lack of interactions with siblings, which have been considered a crucial social environment during children\u0026rsquo;s growth (\u003cem\u003e5\u003c/em\u003e). Whereas the positive outcomes are associated with receiving undivided care and support from their families, as well as growing up in families with higher socioeconomic status (SES) and in more developed areas (\u003cem\u003e9\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e13\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e20\u003c/em\u003e), indicating that both GWS itself and other growth environments may contribute to the variations in behavior and mental health between OC and non-OC individuals. The critical impacts of other growth environments are supported by the observations that GWS only affects academic achievement and mental health in urban-living children (\u003cem\u003e13\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e21\u003c/em\u003e), and that childhood problem behaviors in OC tend to diminish or even reverse after adolescence (\u003cem\u003e17\u003c/em\u003e). However, the specific mechanisms by which modifiable growth environments mediate or moderate the effects of GWS on behavior and mental health remain largely unknown.\u003c/p\u003e\n\u003cp\u003eFrom the perspective of environmental neuroscience, environmental exposures, such as GWS and modifiable growth environments, can affect human behavior and mental health by altering brain structure and function. However, only two neuroimaging studies have tested the effects of GWS on brain structure. Compared with non-OC, OC show thinner thickness and larger surface area in regional cerebral cortices of children (\u003cem\u003e22\u003c/em\u003e). In college students, OC were found to have larger gray matter volume (GMV) in supra-marginal gyrus (SMG) and smaller GMV in medial prefrontal cortex (\u003cem\u003e23\u003c/em\u003e). Nonetheless, it is still unclear about the effects of GWS on other brain imaging properties, such as brain white matter integrity and spontaneous neuronal activity. Furthermore, while modifiable growth environments may mediate or moderate the effects of GWS on both imaging-derived phenotypes (IDPs) and behavioral phenotypes, and IDPs may in turn mediate or moderate the effects of GWS on behavioral phenotypes, there is a lack of comprehensive knowledge regarding the direct and indirect effects of GWS on adult brain and behavior.\u003c/p\u003e\n\u003cp\u003eIn this study, we employed propensity score matching (PSM) to create well-matched pairs of OC and non-OC participants from 6,894 healthy young Chinese Han adults recruited through the Chinese Imaging Genetics (CHIMGEN) study (\u003cem\u003e24\u003c/em\u003e). Then, we investigated the intergroup differences between the matched OC and non-OC in terms of growth environmental exposures, GMV, white matter integrity, spontaneous neuronal activity, cognitive performance, and mental health. We defined the growth environments associated with GWS as the proximal environmental exposures (PEEs), and conducted causal mediation analysis (CMA) and structural equation model (SEM) to investigate the mediation and moderation of PEEs on the associations between GWS and brain/behavioral phenotypes, as well as to quantify the relative contribution of GWS and PEEs to these phenotypes. We anticipate that the results could inform preventive interventions aimed at rectifying adverse or enhancing beneficial GWS effects on brain, cognition, and mental health, given that many of the PEEs are modifiable.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eAll participants were recruited from 32 sites by the CHIMGEN study (\u003cem\u003e24\u003c/em\u003e) with the inclusion and exclusion criteria listed in \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e. After sample selection (\u003cb\u003eSupplementary Fig.\u0026nbsp;1\u003c/b\u003e), we initially included 6,894 participants (2,521 OC and 4,373 non-OC) in this study. The distribution of these participants across sites is provided in \u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e. To balance the overall covariate distributions between groups, PSM was used to generate the matched OC and non-OC groups in 17 confounding covariates, including age, sex, interaction between age and sex, education, body mass index (BMI), 10 genetic principal components (PCs), total intracranial volume (TIV), and frame-wise displacement (FD). Based on the propensity scores of these participants, 1:1 nearest neighbor matching created 2,305 matched pairs of OC and non-OC. The standardized mean difference (SMD) between OC and non-OC was reduced from 0.006\u0026ndash;0.434 to 0.000-0.044 after matching (\u003cb\u003eSupplementary Table\u0026nbsp;3, Supplementary Fig.\u0026nbsp;2\u003c/b\u003e), indicating that the matched OC and non-OC groups are well balanced in confounding covariates. We applied the pair assignment in subsequent analyses, and if any individual within a pair had missing data, the entire pair would be dropped from the analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eBrain imaging differences between OC and non-OC\u003c/h2\u003e \u003cp\u003eWe used GMV to evaluate gray matter macrostructural feature, fractional anisotropy (FA) to assess brain white matter microstructural integrity, and regional homogeneity (ReHo) to assess brain spontaneous neuronal activity. We applied Combat harmonization (\u003cem\u003e25\u003c/em\u003e) to adjust for the scanner effects for all neuroimaging measures, which could effectively reduce the bias resulting from acquiring neuroimaging data with different magnetic resonance imaging (MRI) scanners. The effectiveness of Combat harmonization was confirmed by the improved consistency of voxel-wise brain imaging measures across scanners (\u003cb\u003eSupplementary Fig.\u0026nbsp;3\u003c/b\u003e) in two volunteers whose neuroimaging data were acquired at 28 different scanners. Of the 2,305 matched pairs, 2,295, 2,278, and 2,260 qualified pairs were finally included in the GMV, FA, and ReHo analyses, respectively. In all neuroimaging analyses, we used a two-sample \u003cem\u003et\u003c/em\u003e-test to examine voxel-wise or phenotype-wise differences in neuroimaging measures between OC and non-OC. We applied a family-wise error (FWE) correction (\u003cem\u003eP\u003c/em\u003e\u003csub\u003ec\u003c/sub\u003e \u0026lt; 0.05) and additionally corrected for three neuroimaging measures (GMV, FA, and ReHo) and three data types (brain, behavior, and environmental exposures), resulting in an FWE-corrected \u003cem\u003eP\u003c/em\u003e\u003csub\u003ec\u003c/sub\u003e \u0026lt; 0.05/3/3\u0026thinsp;=\u0026thinsp;5.55\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the voxel-wise GMV analysis, compared with non-OC (n\u0026thinsp;=\u0026thinsp;2,295), OC (n\u0026thinsp;=\u0026thinsp;2,295) showed smaller GMV in the medial frontal and parietal cortex (MFPC), left inferior temporal gyrus (ITG_L), and right orbitofrontal cortex (OFC_R), and showed larger GMV in the bilateral cerebellar posterior lobes (CPL_L and CPL_R) and vermis 4\u0026ndash;5 (CV4-5) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, \u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e, and \u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e). For each participant, we calculated the FA values of 87 pre-defined white matter fiber tracts (\u003cem\u003e26\u003c/em\u003e). Compared with non-OC (n\u0026thinsp;=\u0026thinsp;2,278), OC (n\u0026thinsp;=\u0026thinsp;2,278) showed lower FA in motor-related fiber tracts including the superior cerebellar peduncle (SCP), left corticobulbar tract (CBT_L), right dentato-rubro-thalamic tracts (DRTT_R), right corticospinal tracts (CST_R), and bilateral reticulospinal tracts (RST_L and RST_R) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb and \u003cb\u003eSupplementary Table\u0026nbsp;5\u003c/b\u003e). Conversely, OC showed higher FA in the right parietal aslant tract (PAT_R), which connects ITG_R and SMG_R, a subdivision of the arcuate fasciculus related to language (\u003cem\u003e27\u003c/em\u003e). Considering the special function of the connection, we compared the functional connectivity difference between ITG_R and SMG_R, as defined by the automated anatomical labeling (AAL) atlas (\u003cem\u003e26\u003c/em\u003e), and found stronger functional connectivity (Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.16, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.83\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e, \u003cb\u003eSupplementary Fig.\u0026nbsp;5\u003c/b\u003e) in OC (n\u0026thinsp;=\u0026thinsp;2,260) compared to non-OC (n\u0026thinsp;=\u0026thinsp;2,260). In the voxel-wise ReHo analysis, compared with non-OC (n\u0026thinsp;=\u0026thinsp;2,260), OC (n\u0026thinsp;=\u0026thinsp;2,260) showed lower ReHo in the bilateral ITG and the medial orbitofrontal cortex (MOFC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec, \u003cb\u003eSupplementary Fig.\u0026nbsp;6\u003c/b\u003e, and \u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eBehavioral differences between OC and non-OC\u003c/h2\u003e \u003cp\u003eA two-sample \u003cem\u003et\u003c/em\u003e-test was used to examine the differences between OC and non-OC in 34 behavioral phenotypes. We corrected for 34 phenotypes and three data types using a Bonferroni-corrected threshold of \u003cem\u003eP\u0026thinsp;\u0026lt;\u003c/em\u003e\u0026thinsp;0.05/34/3\u0026thinsp;=\u0026thinsp;4.90 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e. The behavioral differences between OC (n\u0026thinsp;=\u0026thinsp;1,617-2,269) and non-OC (n\u0026thinsp;=\u0026thinsp;1,617-2,269) are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed and \u003cb\u003eSupplementary Table\u0026nbsp;6\u003c/b\u003e. Compared to non-OC, OC performed better in working memory (WM; Cohen\u0026rsquo;s \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.20, \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;1.16 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;10\u003c/sup\u003e) assessed by the accuracy of the 3-back task, in immediate free recall (IFR; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.14; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;3.16 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) assessed by the total number of correct words of list A in five trials of the California verbal learning test II, in executive control (EC; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.16; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;1.32 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e) assessed by the accuracy of no-go trails in the go/no-go task. In addition, OC showed more satisfaction with life (SWL; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.13; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;1.51 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e), more self-image goals (IGS-SI; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.13; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;2.72 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) assessed by the interpersonal goals scale, higher perspective-taking (IRI-PT; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.16; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;9.50 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) and lower personal distress (IRI-PD; Cohen\u0026rsquo;s \u003cem\u003ed =\u003c/em\u003e -0.13; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;1.80 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) assessed by the interpersonal reactivity index scale, more openness (BFI-O; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.32; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;6.67 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;20\u003c/sup\u003e) assessed by the big five inventory, more novelty-seeking (TPQ-NS; Cohen\u0026rsquo;s \u003cem\u003ed\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.15; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;6.41 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e) and less reward-dependence (TPQ-RD; Cohen\u0026rsquo;s \u003cem\u003ed =\u003c/em\u003e -0.13; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;7.09 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) assessed by the tridimensional personality questionnaire, as well as less cognitive impulsiveness (BIS-CI; Cohen\u0026rsquo;s \u003cem\u003ed =\u003c/em\u003e -0.18; \u003cem\u003eP\u0026thinsp;=\u003c/em\u003e\u0026thinsp;6.84 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e) assessed by the Barratt impulsiveness scale.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003ePEE differences between OC and non-OC\u003c/h2\u003e \u003cp\u003eFrom the 32 early-life environmental exposures, we identified 23 PEEs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, \u003cb\u003eSupplementary Table\u0026nbsp;7)\u003c/b\u003e by comparing the differences in growth environmental exposures (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05/32/3\u0026thinsp;=\u0026thinsp;5.20 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e, Bonferroni correction for 32 exposures and three data types) between OC (n\u0026thinsp;=\u0026thinsp;1,455-2,267) and non-OC (n\u0026thinsp;=\u0026thinsp;1,455-2,267) using a two-sample \u003cem\u003et\u003c/em\u003e-test. Compared to non-OC, OC had parents with better education and occupation and families with higher income-to-need, larger house, more resources, less unemployment stress, financial difficulty and crisis; lived in cities with greater gross domestic product (GDP), more education expense and hospital beds; lived in safer community and more harmonious neighborhood; lived in environment with poorer indoor and outdoor air quality; received more maternal care, control, family and friend support; and experienced less physical neglect.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003ePEE factors and their differences between OC and non-OC\u003c/h2\u003e \u003cp\u003eIn the 2,902 participants with all 23 PEE assessments, the Kaiser-Meyer-Olkin (KMO) test and the Bartlett's chi-square test of sphericity were used to evaluate the suitability of the PEE correlation structure (\u003cb\u003eSupplementary Fig.\u0026nbsp;7\u003c/b\u003e) for conducting an exploratory factor analysis (EFA). The KMO\u0026thinsp;=\u0026thinsp;0.83 and chi-square\u0026thinsp;=\u0026thinsp;57,026 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) indicate that the PEE correlation structure is suitable for EFA. We obtained an optimal five-factor EFA model with very simple structure (VSS)\u0026thinsp;=\u0026thinsp;0.8 and minimum average partial (MAP)\u0026thinsp;=\u0026thinsp;0.02. Based on the five-factor EFA model, we defined PEEs with factor loading\u0026thinsp;\u0026gt;\u0026thinsp;0.45 as the main contributors for each factor. According to the included contributors (PEEs), we named the five PEE factors as family (FSES), adverse (ASES) and city-level (CSES) socioeconomic status, air pollution (AP), and support and care (SC), respectively. FSES included paternal and maternal occupation, paternal and maternal education, family resources, and income-to-need with factor loadings of 0.72, 0.74, 0.82, 0.80, 0.51, and 0.45; ASES included financial difficulty, financial crisis, unemployment stress, and home inadequacy with factor loadings of 0.77, 0.65, 0.66, and 0.54; CSES included education expense, GDP, and hospital bed with factor loadings of 0.94, 0.90, and 0.60; AP included PM\u003csub\u003e2.5\u003c/sub\u003e and NO\u003csub\u003e2\u003c/sub\u003e with factor loadings of 0.81 and 0.79; and SC included maternal care, family support, and friend support with factor loadings of 0.61, 0.81 and 0.58 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). To account for uncertainty in the within-sample prediction, a 10-fold cross-validation CFA was conducted to predict out-of-sample PEE factor scores. In each iteration, 90% participants were used to estimate the factor loadings of the CFA model, which were then applied to calculate the PEE factor scores of other 10% participants. The goodness-of-fit statistics indicated a perfect fit for the models established for each iteration and the full sample (\u003cb\u003eSupplementary Table\u0026nbsp;8\u003c/b\u003e). PEE factor scores estimated in the 10-fold cross-validation were highly correlated with those estimated in the full sample (all \u003cem\u003erho\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.99, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (\u003cb\u003eSupplementary Fig.\u0026nbsp;8\u003c/b\u003e), indicating that the five-factor CFA model is stable, and we used the out-of-sample PEE factor scores in the subsequent analyses. We also compared PEE factor scores between OC (n\u0026thinsp;=\u0026thinsp;1,451) and non-OC (n\u0026thinsp;=\u0026thinsp;1,451), and found significant intergroup differences in all five factor scores (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05/5/3\u0026thinsp;=\u0026thinsp;3.33 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePEE factor scores mediate or moderate the effects of GWS on brain and behavior\u003c/h2\u003e \u003cp\u003eFor the five GWS-related PEE factors, 16 IDPs, and 11 behavioral phenotypes, we used CMA to test the mediation and moderation of the five PEE factors on the associations of GWS with the 27 phenotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-b). Since the 16 IDPs were highly correlated (\u003cb\u003eSupplementary Fig.\u0026nbsp;9\u003c/b\u003e), we estimated the effective number of IDPs (n\u0026thinsp;=\u0026thinsp;7.63) and set a Bonferroni corrected threshold of \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05/5/ (11\u0026thinsp;+\u0026thinsp;7.63)\u0026thinsp;=\u0026thinsp;5.36\u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e. The mediation and moderation effects of five PEE factors are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec-d and the full statistics are provided in \u003cb\u003eSupplementary Table\u0026nbsp;9\u003c/b\u003e. Specifically, FSES mediated the positive associations of GWS with BFI-O (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.203), IFR (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.124), WM (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.088), SWL (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.104), TPQ-NS (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.112), IGS-SI (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.097), and GMV-CPL_L (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.079), and the negative associations of GWS with BIS-CI (\u003cem\u003eβ\u003c/em\u003e = -0.089), GMV-MFPC (\u003cem\u003eβ\u003c/em\u003e = -0.097), ReHo-ITG_L (\u003cem\u003eβ\u003c/em\u003e = -0.136), ReHo-ITG_R (\u003cem\u003eβ\u003c/em\u003e = -0.124), FA-CBT_L (\u003cem\u003eβ\u003c/em\u003e = -0.103), FA-CST_R (\u003cem\u003eβ\u003c/em\u003e = -0.085), and FA-RST_L (\u003cem\u003eβ\u003c/em\u003e = -0.092). SC mediated the positive associations of GWS with SWL (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.099), IRI-PT (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.033), TPQ-RD (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.041), and IGS-SI (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.020), and the negative associations of GWS with BIS-CI (\u003cem\u003eβ\u003c/em\u003e = -0.061) and IRI-PD (\u003cem\u003eβ\u003c/em\u003e = -0.027). ASES mediated the positive associations of GWS with SWL (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.142), and the negative associations of GWS with BIS-CI (\u003cem\u003eβ\u003c/em\u003e = -0.061), IRI-PD (\u003cem\u003eβ\u003c/em\u003e = -0.088), and ReHo-ITG_L (\u003cem\u003eβ\u003c/em\u003e = -0.059). CSES mediated the positive association between GWS and IGS-SI (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.025). AP mediated the negative associations of GWS with GMV-MFPC (\u003cem\u003eβ\u003c/em\u003e = -0.018) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec and \u003cb\u003eSupplementary Table\u0026nbsp;9\u003c/b\u003e). All these mediation effects were significant at \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.36\u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e. However, we did not find any significant moderation (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.36\u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) of PEE factors on the associations between GWS and phenotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed and \u003cb\u003eSupplementary Table\u0026nbsp;9\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eIDPs mediate or moderate the effects of GWS on behavior\u003c/h2\u003e \u003cp\u003eWe used CMA to test the mediation and moderation of 16 IDPs on the associations of GWS with 11 behavioral phenotypes (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05/7.63/11\u0026thinsp;=\u0026thinsp;5.95 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e, Bonferroni corrected). The mediation and moderation effects of IDPs are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee-f and the full statistics are provided in \u003cb\u003eSupplementary Table\u0026nbsp;10\u003c/b\u003e. We found that the positive association between GWS and IFR was mediated by GMV-MFPC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.017), ReHo-ITG_R (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.016), and FA-SCP (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008); the positive correlation between GWS and EC was mediated by GMV-CV4-5 (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.013); the positive correlation between GWS and IGS-SI was mediated by GMV-CV4-5 (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.014); the negative correlation between GWS and IRI-PD was mediated by GMV-CPL_R (\u003cem\u003eβ\u003c/em\u003e = -0.011); and the positive correlation between GWS and TPQ-NS was mediated by GMV-ITG_L (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.012) and ReHo-ITG_R (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.019). Besides, we found GMV-ITG_L (\u003cem\u003eβ\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.014) and GMV-OFC_R (\u003cem\u003eβ\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.010) showed reverse mediation effects on the association between GWS and IRI-PD; and GMV-ITG_L showed reverse mediation effects on the association between GWS and BIS-CI (\u003cem\u003eβ\u0026thinsp;=\u003c/em\u003e\u0026thinsp;0.019) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee and \u003cb\u003eSupplementary Table\u0026nbsp;10\u003c/b\u003e). All these mediation effects were significant at \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.95 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e. However, we failed to find any significant moderation (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.95 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) of IDPs on the associations of GWS with behavioral phenotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ef and \u003cb\u003eSupplementary Table\u0026nbsp;10\u003c/b\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003ePathways from GWS to brain and behavior\u003c/h2\u003e \u003cp\u003eCMA indicate that PEE factors mediate rather than moderate the associations of GWS with IDPs and behavioral phenotypes. However, CMA estimates the mediation without taking account of other predictors and cannot identify the complex pathways including two or more sequential mediators. Therefore, we performed structural equation model (SEM) to simultaneously model the complex relationships among the GWS, PEE factors, IDPs, and behavioral phenotypes in 2,618 participants without missing data in these variables. In the measurement model, we used the same structure as the five-factor-CFA model to construct five PEE latent variables (LVs) of FSES, ASES, CSES, SC, and AP. Based on the brain imaging measure (GMV, FA, or ReHo) and the direction of GWS effect (higher or lower in OC) of IDPs, we constructed four LVs: small GMV (SGMV) constructed by the three IDPs with smaller GMV in OC; large GMV (LGMV) by the three IDPs with larger GMV in OC; low FA (LFA) by the six IDPs with lower FA in OC; and low ReHo (LReHo) by the three IDPs with lower ReHo in OC. The right PAT with higher FA in OC was used as a single variable in the SEM analysis. We did not construct behavioral LVs owing to their weak correlations (\u003cb\u003eSupplementary Fig.\u0026nbsp;10\u003c/b\u003e). In the structural model, we defined GWS as the independent variable; FSES, ASES, CSES, SC, AP, SGMV, LGMV, LFA, FA-PAT_R, and LReHo as mediators; and IFR, EC, WM, TPQ-RD, TPQ-NS, BFI-O, BIS-CI, IGS-SI, IRI-PD, IRI-PT, and SWL as outcomes (\u003cb\u003eSupplementary Fig.\u0026nbsp;11\u003c/b\u003e). The constructed SEM showed an acceptable goodness-of-fit (CFI\u0026thinsp;=\u0026thinsp;0.919, RMSEA\u0026thinsp;=\u0026thinsp;0.042, and SRMR\u0026thinsp;=\u0026thinsp;0.055).\u003c/p\u003e \u003cp\u003eIn SEM, we defined a two-sided \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 as the threshold for statistical significance. The model explained 0.8\u0026ndash;13.3% variances of IDPs (highest for LReHo) and 2.0-21.1% variances of behavioral phenotypes (highest for SWL) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). The relative effects of GWS and PEEs on IDPs and behavioral phenotypes are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb, and all significant associations in SEM were shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec. For IDPs, FSES, AP, and SC showed comparable effects on SGMV (\u003cem\u003eβ\u003c/em\u003e = -0.095, -0.094, and \u0026minus;\u0026thinsp;0.07, respectively); FSES, CSES, and GWS showed comparable effects on LGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.102, 0.100, and 0.093, respectively); FSES was the strongest predictor of LFA (\u003cem\u003eβ\u003c/em\u003e = -0.099) and LReHo (\u003cem\u003eβ\u003c/em\u003e = -0.254); and GWS was the strongest predictor of FA-PAT_R (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.084) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb-c). For the behavioral phenotypes, GWS was the strongest predictor of EC (β\u0026thinsp;=\u0026thinsp;0.106); FSES was the strongest predictor of BFI-O (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.333), WM (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.153), and TPQ-NS (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.126); SC was the strongest predictor of SWL (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.442), BIS-CI (\u003cem\u003eβ\u003c/em\u003e = -0.300), TPQ-RD (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.193), and IRI-PT (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.188); IFR was mainly affected by FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.158), CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.109), and AP (\u003cem\u003eβ\u003c/em\u003e = -0.101); IGS-SI was mainly affected by FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.125), and SC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.119); and IRI-PD was mainly affected by ASES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.142), and CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.127) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb-c). For the associations of IDP with behavioral phenotypes, SGMV was primarily associated with IFR (\u003cem\u003eβ\u003c/em\u003e = -0.080) and BIS-CI (\u003cem\u003eβ\u003c/em\u003e = -0.111); LGMV was mainly associated with IGS-SI (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.089); FA-PAT_R was mainly associated with EC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.050); and LFA was associated with BIS-CI (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.046) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003eThe identified direct and indirect effects of GWS on IDPs and behavioral phenotypes are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed. GWS showed direct effects on FA-PAT_R (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.084), LGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.093), LReHo (\u003cem\u003eβ\u003c/em\u003e = -0.111), EC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.106), and TPQ-RD (\u003cem\u003eβ\u003c/em\u003e = -0.069). For the indirect effects (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed and \u003cb\u003eSupplementary Table\u0026nbsp;11\u003c/b\u003e), GWS positively affected: (1) LGMV via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.057) and CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.022); (2) IFR via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.089), CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.024), and the sequential mediators of FSES-SGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004), SC-SGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0005) and AP-SGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001); (3) EC via the mediators of AP (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.007) and FA-PAT_R (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004); (4) WM via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.086) and CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.019); (5) SWL via the mediators of ASES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.049) and SC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.040); (6) BFI-O via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.188) and CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.017); (7) IRI-PT via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.057) and SC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.017); (8) IGS-SI via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.071), CSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.014), SC (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.011), LGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008), and CSES-LGMV (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002); (9) TPQ-NS via the mediator of FSES (\u003cem\u003eβ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.071) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed and \u003cb\u003eSupplementary Table\u0026nbsp;11\u003c/b\u003e). GWS negatively affected: (1) SGMV via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e = -0.053), AP (\u003cem\u003eβ\u003c/em\u003e = -0.011) and SC (\u003cem\u003eβ\u003c/em\u003e = -0.007); (2) LReHo via the mediators of FSES (\u003cem\u003eβ\u003c/em\u003e = -0.143) and CSES (\u003cem\u003eβ\u003c/em\u003e = -0.036); (3) LFA via the mediator of FSES (\u003cem\u003eβ\u003c/em\u003e = -0.056); (4) BIS-CI via the mediators of SC (\u003cem\u003eβ\u003c/em\u003e = -0.027), AP (\u003cem\u003eβ\u003c/em\u003e = -0.009), and FA-PAT_R (\u003cem\u003eβ\u003c/em\u003e = -0.004); (5) TPQ-RD via the mediator of FSES (\u003cem\u003eβ\u003c/em\u003e = -0.040); and (6) IRI-PD via the mediators of ASES (\u003cem\u003eβ\u003c/em\u003e = -0.050) AP (\u003cem\u003eβ\u003c/em\u003e = -0.010), and SC (\u003cem\u003eβ\u003c/em\u003e = -0.007) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed and \u003cb\u003eSupplementary Table\u0026nbsp;11\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eConsistent pathways from GWS to brain and behavior\u003c/h2\u003e \u003cp\u003eCMA and SEM have identified several consistent mediation pathways from GWS to IDPs and behavioral phenotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Specifically, GWS affected SGMV, LFA, IFR, WM, SWL, BIS-CI, IRI-PT, IRI-PD, IGS-SI, TPQ-NS, and BFI-O only through indirect pathways, but with different mediators. GWS affected FA-PAT_R only through direct pathway. GWS affected EC, TPQ-RD, LGMV, and LReHo through both direct and indirect pathways. Of the 16 brain and behavioral phenotypes, GWS affected 11 phenotypes via FSES, nine via SC, four via CSES, three via AP, and two via ASES. SGMV, LGMV, and FA-PAT_R were IDP mediators for the associations between GWS and behavioral phenotypes. In addition, CMA and SEM also revealed several indirect pathways whose effect directions were opposite to the total GWS effects, including the indirect effects of SC on TPQ-RD and TPQ-NS; AP on IFR; AP-SGMV, FSES-SGMV, and SC-SGMV on BIS-CI; and AP-SGMV and FSES-SGMV on IRI-PT.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn up to 2,305 pairs of well-matched OC and non-OC, we investigated how GWS affects adult brain and behavior by testing their differences in terms of early-life growth environments, adult brain imaging and behavioral phenotypes, as well as the causal pathways from GWS to growth environments to brain and behavior. We identified novel associations of GWS with brain and behaviors, including white matter integrity, cerebellum volume, spontaneous neuronal activity, memory, executive control, and mental health. OC lived in better socioeconomic environment and received more maternal care, family and friend support than non-OC, in line with the resource dilution and attachment theory (\u003cem\u003e9\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e28\u003c/em\u003e). We also found that most GWS effects on brain and behavior were mediated by modifiable growth environmental exposures, informing interventions to improve the healthy development of brain and behavior by increasing advantageous and reducing disadvantageous environmental exposures in early life.\u003c/p\u003e\n\u003cp\u003eConsistent with an earlier study, OC showed smaller MFPC and ITG volumes (\u003cem\u003e23\u003c/em\u003e), and the smaller MFPC volume was associated with better memory in the present study. The MFPC volume reaches its peak before puberty, followed by a nearly linear reduction (\u003cem\u003e29\u003c/em\u003e). The volume reduction before adult is an integrated reflection of the synaptic pruning, axonal caliber alteration, and myelination (\u003cem\u003e30\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e31\u003c/em\u003e) which is associated with higher neuronal efficiency (\u003cem\u003e32\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e33\u003c/em\u003e). Thus, our results may be explained by accelerated normative growth of cerebral cortex (\u003cem\u003e34-36\u003c/em\u003e) in OC individuals. As the GWS effect on MFPC volume was mainly mediated by family SES, family SES may be an intervention target for memory improvement. The larger cerebellar volume in OC was attributed to both direct effect and indirect effect mediated by family SES, in line with positive correlation between SES and cerebellar volume (\u003cem\u003e36-38\u003c/em\u003e). We also extend previously reported association between family SES and FA (white matter integrity) (\u003cem\u003e39\u003c/em\u003e) to the mediation of family SES in the association between GWS and lower FA in motor-related fibers. Furthermore, OC showed higher FA in the right PAT (a language-related tract) and higher functional connectivity between right ITG and SMG, supporting the superior language ability in OC (\u003cem\u003e40\u003c/em\u003e). PAT integrity was only affected by the direct GWS effect, indicating that its integrity cannot be changed by improving these common growth environments. Although OC showed lower ReHo (spontaneous neuronal activity), we failed to find reliable association between ReHo and behavior, which cannot provide useful information for intervention.\u003c/p\u003e\n\u003cp\u003eContrary to the expectation of more behavioral problems in OC (\u003cem\u003e5\u003c/em\u003e), we found that OC outperformed non-OC in various behavioral outcomes. In social-emotional skills, OC exhibited more prosocial (perspective-taking), and less distressful (personal distress), which were completely mediated by growth environments: family SES, support and care for perspective-taking, and adverse SES and support and care for personal distress. Similarly, OC showed higher levels of life satisfaction which was mediated by more support and care and less adverse SES. OC also showed more self-image goals in interpersonal relationship, which were mediated by family and city-level SES, support and care, and cerebellar volume, and the association between cerebellar volume and social cognition has been reported previously (\u003cem\u003e36\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e41\u003c/em\u003e). These results suggest that the poorer social cognition and life satisfaction of non-OC could be enhanced by improving socioeconomic status and providing more care and support during early years of life. Regarding personality, OC showed more openness and novelty-seeking but less reward-dependence and cognitive impulsiveness. Both openness and novelty-seeking are associated with creativity, which is higher in OC college students (\u003cem\u003e23\u003c/em\u003e). The GWS effects on these two personality traits were primarily mediated by family SES, suggesting that creativity may be facilitated by improving family socioeconomic circumstances during early development. Furthermore, the GWS effect on cognitive impulsiveness was completely mediated by support and care and family SES, in line with the findings that self-control could be enhanced by improving caregiving and family SES (\u003cem\u003e42\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e43\u003c/em\u003e). The lower level of reward dependence observed in OC was mainly driven by the direct effect of GWS, but the effect was mitigated by support and care. Via this regulatory pathway, parents could help their children achieve the balance between dependency and independency by providing appropriate care and support while avoiding controlling behaviors. \u003c/p\u003e\n\u003cp\u003eIn terms of neurocognitive abilities, OC showed better performance in verbal learning and memory, working memory, and executive control than non-OC. However, GWS affects different neurocognitive domains through distinct pathways. For example, GWS positively affected verbal learning memory and working memory solely through indirect pathways with FSES, CSES, SC, and SGMV serving as mediators. On the other hand, the effect of GWS on executive control was primarily through the direct pathway, although indirect pathway also contributed to a smaller extent. These findings emphasize the importance of designing neurocognitive improvement strategies in a domain-specific manner, although a single mediator may involve in the regulation of multiple neurocognitive domains.\u003c/p\u003e\n\u003cp\u003eExcept for white matter integrity of the right PAT and executive control ability, the effects of GWS on brain and behavior were mediated by modifiable environmental exposures, which provides us opportunities to improve adult behavioral outcomes by modifying early-life environmental exposures according to the indirect pathways and their relative effects identified in this study. For instance, the improvement of family socioeconomic status may benefit for the development of brain and neurocognition, and the provision of more maternal care and support could benefit for life satisfaction and mental health These interventions should be implemented during the early years of life to help narrow the gaps between OC and non-OC individuals. Two limitations should be considered when one interprets our findings. First, all participants in this study were Chinese, which may limit the generalizability of the results to other populations. Second, as many environmental assessments relied on participant recall, there is a possibility of recall bias that may have affected our results.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch2\u003eParticipants\u003c/h2\u003e\n\u003cp\u003eThe CHIMGEN study recruited 7,306 healthy Chinese Han adult participants aged 18-30 years from 32 research centers in 22 provinces of China (\u003cem\u003e24\u003c/em\u003e). After excluding those with unqualified data or missing covariate variables (\u003cstrong\u003eSupplementary Fig. 1\u003c/strong\u003e), we initially included 6,894 participants, from which we generated 2,305 pairs of well-matched OC and non-OC participants. Written informed consent was obtained from each participant, and the CHIMGEN study was approved by the Medical Research Ethics Committees of Tianjin Medical University General Hospital (IRB2015-092-01) and other institutions.\u003c/p\u003e\n\u003ch2\u003eNeuroimaging data acquisition and preprocessing\u003c/h2\u003e\n\u003cp\u003eAll the neuroimaging data were acquired by ten different types of 3T magnetic resonance imaging (MRI) scanners, and the scanning sequences and parameters for each type of MRI scanners are provided in \u003cstrong\u003eSupplementary Table 12-14\u003c/strong\u003e. The structural MRI (sMRI) data were used to calculate gray matter volume (GMV) to assess brain macrostructural features; the diffusion tensor imaging (DTI) data were used to calculate fractional anisotropy (FA) to evaluate brain white matter integrity; and the resting-state functional MRI (rs-fMRI) data were used to calculate regional homogeneity (ReHo) and functional connectivity (FC), which were used to assess spontaneous neuronal activity within brain regions and temporal coherence of spontaneous neuronal activity between brain regions.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003esMRI data preprocessing\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe\u003cem\u003e \u003c/em\u003esMRI data were preprocessed using computational anatomy toolbox (CAT12 v1364) (http://dbm.neuro.uni-jena.de/cat) implemented in statistical parametric mapping (SPM12; http://www.fil.ion.ucl.ac.uk/spm). After correcting for image inhomogeneity from B1-field bias, the structural images were segmented into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) using a model based on an adaptive Maximum A Posterior (\u003cem\u003e44\u003c/em\u003e). To improve image normalization, the population-specific tissue probability templates for GM, WM, and CSF in Montreal Neurological Institute (MNI) space were derived from 6,000 CHIMGEN participants using the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) algorithm (\u003cem\u003e45\u003c/em\u003e). The individual GM images were normalized to the GM template using a two-step DARTEL algorithm and resampled into a cubic voxel of 1.5-mm. Modulation was performed on the normalized GM images to preserve absolute GMV. The resulting voxel-wise GMV maps were smoothed with a Gaussian kernel of 8 \u0026times; 8 \u0026times; 8 mm\u003csup\u003e3\u003c/sup\u003e full width at half maximum (FWHM). The TIV of each participant was calculated as the sum of the volumes of GM, WM, and CSF. \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDTI data preprocessing. \u003c/em\u003eDTI data were preprocessed using FMRIB\u0026rsquo;s Software Library (FSL v5.0.10; www.fmrib.ox.ac.uk/fsl). Non-brain tissues were removed from b = 0 images using the brain extraction tool to generate a binary mask of brain tissue. The \u0026ldquo;EDDY_OPENMP\u0026rdquo; program was used to evaluate and repair image displacement, signal dropout caused by head motion, and image distortion caused by eddy current. A linear least square algorithm was applied to estimate diffusion tensor and to calculate FA value of each voxel using the DTIFIT program. A two-step procedure was conducted to estimate the normalization parameters between individual diffusion and MNI spaces. Specifically, individual b = 0 images were aligned to structural images using Boundary-Based Registration (BBR) (\u003cem\u003e46\u003c/em\u003e). The obtained transformation matrix was concatenated with the DARTEL deformation field from individual to MNI space generated in the sMRI data preprocessing. The individual FA maps were normalized to the MNI space using the merged deformation field (BBR+DARTEL) and resampled into a cubic voxel of 2-mm. A mean FA image of all individuals was created and \u0026ldquo;thinned\u0026rdquo; to generate a mean white matter skeleton representing the centers of tracts common to all individuals. Each participant\u0026rsquo;s aligned FA images were projected onto the mean FA skeleton by filling the mean FA skeleton with FA values from the nearest tract center, achieving by searching perpendicular to the local skeleton structure for maximum value. We then calculated the mean FA values of 87 white matter tracts derived from 1,065 subjects of the human connectome project (HCP) (\u003cem\u003e26\u003c/em\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ers-fMRI data preprocessing. \u003c/em\u003eThe rs-fMRI data were preprocessed using SPM12 and data processing assistant for resting-state fMRI analysis (DPARSFA v4.4) (\u003cem\u003e47\u003c/em\u003e). As three different repetition times (TRs: 0.71, 0.8, and 2 s) were used to acquire rs-fMRI data and the functional images acquired at each TR was defined as a functional volume, the first several functional volumes (\u0026asymp; 10 s: 15 volumes for TR = 0.71 s; 13 for TR = 0.8 s; and 5 for TR = 2 s) were discarded to allow signals to reach equilibrium. The remaining volumes were corrected for intra-volume temporal differences using sinc-interpolation and the inter-volume head motion was estimated and corrected by realigning each volume to the mean volume using rigid-body transformation. Based on the estimation of head motion, we excluded the participants whose fMRI data had the maximum displacement in any of the three orthogonal directions \u0026gt; 3 mm or a maximum rotation \u0026gt; 3.0 degree. In the remaining participants, we calculated FD to index volume-to-volume changes in head position using the Jenkinson method (\u003cem\u003e48\u003c/em\u003e). We then conducted independent component analysis based automatic removal of motion artifacts (ICA-AROMA) (\u003cem\u003e49\u003c/em\u003e), from which we identified and removed the independent components for motion artifacts. After removing non-brain tissues, the functional images were co-registered to the structural images using the BBR method. All co-registered functional volumes were normalized to MNI space using the deformation fields derived from the sMRI data preprocessing and were resampled to 3-mm isotropic voxels. After regressing out linear trend and signals from WM and CSF, a temporal band-pass filtering (0.01-0.08 Hz) was applied to reduce low-frequency drift and high-frequency noise. The unsmoothed fMRI data were used for computing ReHo. Here, ReHo of each voxel was defined as the Kendall\u0026rsquo;s coefficient concordance of the time series of this voxel with its nearest neighbors (26 voxels). The whole process was repeated for all GM voxels to generate the voxel-wise ReHo map of each participant. For standardization, ReHo of each voxel was subtracted from the mean and divided by the standard deviation (SD) of ReHo values of all GM voxels. Then, the ReHo maps were smoothed with a kernel of FWHM = 8 mm. For the special fiber tracts of interest, we also calculated the FCs between their connecting brain regions defined by the automated anatomical labeling (AAL) atlas (\u003cem\u003e26\u003c/em\u003e). Here, FC was calculated based on the smoothed fMRI data (FWHM = 8 mm) and was defined as the Pearson correlation coefficient between the mean time-courses of two brain regions. To improve normality, the correlation coefficients were transformed to z value using Fisher r-to-z transformation.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eN\u003c/em\u003e\u003cem\u003eeuroimaging data \u003c/em\u003e\u003cem\u003eq\u003c/em\u003e\u003cem\u003euality control (QC). \u003c/em\u003eA series of QC procedures were applied both before and after the acquisition of brain MRI data. For example, we optimized scanning parameters for each MRI scanner before acquisition. We identified and excluded MRI data with lesions, anatomical abnormalities, imaging artefacts, parameter inconsistency, and incomplete whole-brain coverage immediately after acquisition. We also checked the preprocessed MRI data to identify errors or imperfections that may have emerged during imaging data preprocessing steps. If they were identified in the preprocessed MRI data of a participant, we tried to identify reasons and re-run the pipeline after fixing the identified problems. If the reprocessed MR images were still problematic, the participant would be excluded from the analyses. The checked error and imperfection included incorrect non-brain tissue removal, bad tissue segmentation, imperfect spatial normalization, and intensity normalization error. Following the QC procedures, we finally included voxel-wise GMV maps from 6,852 participants, tract-based FA measurements from 6,844 participants, and voxel-wise ReHo maps and FCs between brain regions from 6,698 participants. Combat harmonization was applied to MRI metrics to adjust for the scanner effect (\u003cem\u003e25\u003c/em\u003e). The effectiveness of harmonization was assessed by within-subject consistency of each imaging measure between scanners in two participants who have traveled to the research sites and were scanned by 28 MRI scanners used in the CHIMGEN study.\u003c/p\u003e\n\u003ch2\u003eBehavioral assessments\u003c/h2\u003e\n\u003cp\u003eIn the present study, we included 34 behavioral variables that were used to measure neurocognition, social cognition, personality, and mental health.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVerbal learning and memory\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe second edition of the California verbal learning test (CVLT-II) was used to test verbal learning and memory (\u003cem\u003e50\u003c/em\u003e). The test includes two lists of words: the list A consists of 16 words of four semantic categories (furniture, vegetable, vehicle, and animal) and the list B also includes 16 words of four categories (vegetable, animal, musical instrument, and house part name; the former two from the word list A). There were five trials for the word list A and one trial for the word list B and words were randomly presented, irrespective of semantic categories. In each trial, the experimenter read aloud the word list A at a rate of one word per second, and the participant was asked to recall as many words as possible. The experimenter recorded the number of correct words of list A in the five trials, which was used to assess immediate free recall (IFR). Immediately after the five trials, the participant was asked to learn and recall word list B once. Then, the participant was asked to engage in short delay free and cued recall of the word list A immediately, as well as long delay free and cued recall of the word list A after a 20-minute of the short delay recall. The participant was asked to recall in no order in free recall and to categorize the words into four categories in cued recall. The numbers of correct words in the four recall tasks were used to assess episodic memory, including the short delayed free (SDFR) and cued (SDCR) recall and the long delayed free (LDFR) and cued (LDCR) recall.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWorking memory\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe letter n-back test (1-back and 3-back) was used to assess working memory (\u003cem\u003e51\u003c/em\u003e). In the 1-back task, the participant was required to infer whether the current stimulus matched the previous one in the sequence of stimuli. In the 3-back task, the participant was required to infer whether the current stimulus matched the one three steps earlier. Both tasks included one block with 60 trials. Each stimulus was presented for 200ms, and the stimulation interval was 1800ms. Before the formal tasks, participants underwent a practice session of the 1-back task, and only those who achieved an accuracy rate higher than 75% were allowed to proceed to the formal task. E-Prime 2.0 software (Psychology Software Tools) was used to present stimuli and record participant responses. The accuracy for 3-back tasks were used to assess working memory (WM).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVisuospatial memory. \u003c/em\u003eThe Rey-Osterrieth complex figure test (ROCFT) was used to assess visuospatial capability (\u003cem\u003e52\u003c/em\u003e). In ROCFT, the participant was asked to copy and recall a complex line graph at three time points. At the beginning, the participant was asked to copy the graph onto a blank sheet of paper. Immediately after copying, the participant was asked to redraw the graph from memory (immediate recall). Twenty-five minutes later, the participant was asked to reproduce the graph again (delayed recall). A 36-point scoring system was used to assess the accuracy of immediate (RO-IR) and delayed (RO-DR) recall tasks based on the scoring guideline for each element of the complex figure. To minimize inter-rater variability, all ROCFT assessments were conducted by a trained group of raters.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInformation processing speed\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe symbol digit modalities test (SDMT) was used to assess information processing ability (\u003cem\u003e53\u003c/em\u003e). The test involves pairing of single digits with abstract symbols. The keys consist of nine abstract symbols, each representing a specific digit. The participant was presented with these symbols and asked to write down the corresponding digits as fast as possible within 90 seconds. The information processing performance was assessed based on the number of correct responses within the given time frame.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExecutive control\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe go/no-go test was used to evaluate sustained attention and behavior inhibition (\u003cem\u003e54\u003c/em\u003e). During \u0026ldquo;go\u0026rdquo; trial (90% of 210 trials), the current letter was different from the previous letter and the participants were required to respond quickly by pressing the button. In contrast, during the \u0026ldquo;no-go\u0026rdquo; trial (10% trials), the current letter was the same as the previous letter and the participant were required not to press any button. The E-Prime 2.0 was used to present stimuli and collect responses. The mean accuracy of \u0026ldquo;no-go\u0026rdquo; trials was used to assess executive control (EC).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eEmpathy. \u003c/em\u003eThe interpersonal reactivity index (IRI) questionnaire was used to test dispositional empathy (\u003cem\u003e55\u003c/em\u003e). The IRI includes four distinct conducts: perspective-taking (IRI-PT) measures the tendency to spontaneously adopt the psychological point of others; fantasy (IRI-F) tests the tendency to imaginatively transpose themselves into the feelings and actions of fictitious characters; empathic concern (IRI-EMC) reflects other-oriented feelings of sympathy and concern towards others who are experiencing unfortunate circumstances; and personal distress (IRI-PD) assesses self-oriented feelings of personal anxiety and unease in tense interpersonal settings.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInterpersonal goals. \u003c/em\u003eThe interpersonal goals scale (IGS) was used to test one\u0026rsquo;s interpersonal goals for their relationships (\u003cem\u003e56\u003c/em\u003e). IGS assesses the compassionate goals (IGS-C) to support others and self-image goals (IGS-SI) to create and maintain desired self-images.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMental health\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eThe second version of Beck depressive inventory (BDI) was used to assess the severity of depression (\u003cem\u003e57\u003c/em\u003e). The state-trait anxiety inventory was used to measure state anxiety (SAI) and trait anxiety (TAI) (\u003cem\u003e58\u003c/em\u003e). The positive and negative affect scale was used to measure positive affect (PNAS-PA) and negative affect (PNAS-NA) (\u003cem\u003e59\u003c/em\u003e). The perceived stress scale was used to measure perceived stress (PSS) (\u003cem\u003e60\u003c/em\u003e). The satisfaction with life scale was used to assess satisfaction with life (SWL) (\u003cem\u003e61\u003c/em\u003e). \u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePersonality.\u003c/em\u003e The big-five inventory (BFI) was used to assess five personality traits, including the agreeableness (BFI-A), conscientiousness (BFI-C), extraversion (BFI-E), neuroticism (BFI-N), and openness (BFI-O) (\u003cem\u003e62\u003c/em\u003e). The tridimensional personality questionnaire (TPQ) was used to assess three personality traits associated with distinct neurotransmitters: novelty-seeking (TPQ-NS) with dopamine; harm-avoidance (TPQ-HA) with serotonin; and reward dependence (TPQ-RD) with norepinephrine (\u003cem\u003e63\u003c/em\u003e). The Barratt impulsiveness scale (BIS) was used to assess impulsiveness (\u003cem\u003e64\u003c/em\u003e), including the motor impulsiveness (BIS-MI), cognitive impulsiveness (BIS-CI), and no-planning impulsiveness (BIS-NPI). \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eB\u003c/em\u003e\u003cem\u003eehavioral \u003c/em\u003e\u003cem\u003edata\u003c/em\u003e\u003cem\u003e Q\u003c/em\u003e\u003cem\u003eCs\u003c/em\u003e\u003cem\u003e. \u003c/em\u003eEach behavioral assessment underwent the following QC procedures: (a) we confirmed the consistency between input data and raw data to ensure that the input data were free of any input errors; (b) we excluded participants with accuracy \u0026lt; 75% on \u0026ldquo;1-back\u0026rdquo; or \u0026ldquo;go\u0026rdquo; trails, indicating the lack of compliance of the participants; (c) we further performed median-based outlier removal (discarding values greater than 3 times the median absolute deviation from the overall median) for each measure. Finally, a total of 34 behavioral variables from 4,828-6,768 participants were used in following analyses \u003cstrong\u003e(Supplementary Fig. 1)\u003c/strong\u003e.\u003c/p\u003e\n\u003ch2\u003eGrowth environmental exposures\u003c/h2\u003e\n\u003cp\u003eThe 32 growth environmental exposures potentially related to growth without siblings (GWS) were evaluated from the following questionnaires and approaches: \u003c/p\u003e\n\u003cp\u003eThe socioeconomic status (SES) questionnaire was used to assess family income, parental occupation and education, family resources, unemployment stress, financial difficulties and crisis, home inadequacy, community safety, and neighbor indifference. Each measure was assessed at three age windows of 0-10, 11-20, and \u0026gt; 21 years, and the mean value of age windows was used to represent the corresponding measure. Family income was represented by income to need, which was defined as the annual household income per capita. Parental education was divided into six grades: 1 = primary school or below, 2 = middle school, 3 = high or technical school, 4 = junior college, 5 = bachelor, 6 = master or doctor. Parental occupation included eight classes: 1 = day laborer or unemployed, 2 = manual worker or self-employed, 3 = operators of equipment, 4 = production personnel, 5 = business or service staff, 6 = civil servants or employees of companies, 7 = professional and technical personnel, 8 = manager of government, institution, or company. Family resources were defined as the sum of 14 binary items (0 = no and 1 = yes) including private room, car, computer, study desk, learning place, Internet, educational software, reference books, television, refrigerator, washing machine, masterpieces, collect books, and artistic works (\u003cem\u003e65\u003c/em\u003e). A three-scoring system was applied to assess family unemployment stress (1 = never, 2 = occasional, 3 = often), financial difficulties (1 = never, 2 = occasional, 3 = often), financial crisis (1 = never, 2 = occasional, 3 = often), home inadequacy (1 = spacious, 2 = moderate, 3 = crowd), and community safety (1 = safe, 2 = moderate, 3 = violent) and neighbor indifference (1 = harmonious, 2 = moderate, 3 = indifferent). To get comparable assessments, we provided a standard set of instructions to the participants\u003c/p\u003e\n\u003cp\u003eA five-scoring system was used to assess indoor air quality (1 = very good; 2 = good; 3 = moderate; 4 = poor; 5 = very poor) at six age windows: 0-5, 6-10, 11-15, 16-20, 21-25, and \u0026gt; 26 years. The mean value of age windows was used to represent the indoor air quality.\u003c/p\u003e\n\u003cp\u003eEach participant was asked to recall residential coordinates of the years from birth to recruitment. If the participant moved in a year, the address where they lived for the longest duration was defined as the residential coordinate of that year. Using annual PM\u003csub\u003e2.5 \u003c/sub\u003eand NO\u003csub\u003e2\u003c/sub\u003e data provided by two prior studies(\u003cem\u003e66\u003c/em\u003e\u003cem\u003e, \u003c/em\u003e\u003cem\u003e67\u003c/em\u003e), we extracted two measurements for each participant from birth year to the year of recruitment. The average values of the\u003csub\u003e \u003c/sub\u003eannual\u003csub\u003e \u003c/sub\u003edata for the available years were used to represent PM\u003csub\u003e2.5\u003c/sub\u003e and NO\u003csub\u003e2\u003c/sub\u003e exposures of the participant. Based on each residential address, we extracted city-level statistical data on GDP per capita, the number of hospital beds per 10,000 people, and education expense per 10,000 people from the annual report of the National Bureau of Statistics of China (http://www.stats.gov.cn/). The annual data of each city-level indicator were obtained for each participant from their birth year to the year of recruitment, and the average value of the lifetime data was used to represent the exposure.\u003c/p\u003e\n\u003cp\u003eThe maternal (PBIM) and paternal (PBIF) versions of 23-item parental bonding instrument (PBI) were used to evaluate parenting styles perceived by the participant before 16 years old (\u003cem\u003e68\u003c/em\u003e). A four-scoring system (0 = strongly disagree, 1 = disagree, 2 = agree, and 3 = strongly agree) was applied to each item. Based on the combination of these items, we obtained the assessments of care, encouragement, and control from mother and father, respectively.\u003c/p\u003e\n\u003cp\u003eThe 12-item multidimensional scale of perceived social support (MSPSS) was used to assess the self-perceived social support of each participant (\u003cem\u003e69\u003c/em\u003e). Each item has seven levels (1-7: very strongly disagree, strongly disagree, disagree, neutral, agree, strongly agree, very strongly agree), and higher score represents greater social support. Based on the combination of these items, we obtained the assessments of supports from family, friends, and others.\u003c/p\u003e\n\u003cp\u003eThe childhood trauma questionnaire (CTQ) was applied to assess childhood abuse experiences (\u003cem\u003e70\u003c/em\u003e). CTQ consists of 28 items reflecting life experience before the age of 16 years and adopts a five-point scoring system (never, occasionally, sometimes, often, and always). The CTQ subscales include emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect. The scores of each subscale range from 5 to 25 points, with higher scores indicating more severe trauma.\u003c/p\u003e\n\u003ch2\u003eConfounding covariates\u003c/h2\u003e\n\u003cp\u003eWe included the following 17 covariates in the PSM analysis: age, sex, the interaction of age and sex, education, BMI, genetic population stratification, TIV and mean FD. In this study, sex was confirmed using genotyping data. The methodology for calculating genetic population stratification for the CHIMGEN participants has been described in our previous study (\u003cem\u003e71\u003c/em\u003e). We found that 10 PCs derived from principal component analysis of genomic data could account for the major variance of population stratification, which were used as the covariates to control for genetic population stratification in the current study.\u003c/p\u003e\n\u003ch2\u003eStatistical analysis\u003c/h2\u003e\n\u003ch3\u003ePropensity score matching (PSM)\u003c/h3\u003e\n\u003cp\u003eFrom 2,521 OC and 4,373 non-OC participants, PSM was used to create well-matched OC and non-OC groups across 17 confounding covariates. The PSM analysis was performed using the \u0026ldquo;MatchIt\u0026rdquo; package from R 4.1.2 (\u003cem\u003e72\u003c/em\u003e). Logistic regression was used to estimate the individual\u0026rsquo;s propensity score (PS), which represents the probability of being in the OC group given the covariates. Based on the propensity scores of each participant, we conducted one-to-one matching between OC and non-OC using the greedy nearest-neighbor method (without replacement and caliper = 0.1), resulting in 2,305 pairs of matched OC and non-OC participants. The standardized mean difference (SMD) was used to assess the matching quality, and an SMD \u0026lt; 0.10 indicated a balanced distribution between two groups (\u003cem\u003e73\u003c/em\u003e). We applied the pair assignment in subsequent analyses, and if any individual in one pair had missing data, the entire pair would be dropped from the analysis.\u003c/p\u003e\n\u003ch3\u003eComparisons of brain imaging phenotypes between OC and non-OC\u003c/h3\u003e\n\u003cp\u003eWith SPM12, a two-sample \u003cem\u003et\u003c/em\u003e-test was performed to identify voxel-wise GMV differences between the well-matched OC and non-OC pairs. Multiple comparisons were corrected using a hybrid family-wise error (FWE) method, where we corrected for voxel-wise comparisons based on random field theory (RFT) (\u003cem\u003e74\u003c/em\u003e), and further corrected for the three neuroimaging features (GMV, FA, and ReHo) and the three data types (environmental, neuroimaging and behavioral data). Thus, a hybrid FWE-corrected threshold\u003cem\u003e \u003c/em\u003ewas\u003cem\u003e P \u0026lt;\u003c/em\u003e 2.17 \u0026times; 10\u003csup\u003e-6\u003c/sup\u003e/3/3 = 2.41 \u0026times; 10\u003csup\u003e-7\u003c/sup\u003e. We defined each cluster with significant GMV difference as an imaging-derived phenotype (IDP), and defined the mean GMV of the cluster in each participant as the IDP value of the participant. The same approach was used to identify voxel-wise differences in ReHo between the matched OC and non-OC pairs, and used a hybrid FWE-corrected threshold of \u003cem\u003eP \u0026lt; \u003c/em\u003e3.29 \u0026times; 10\u003csup\u003e-6\u003c/sup\u003e/3/3 = 3.65 \u0026times; 10\u003csup\u003e-7\u003c/sup\u003e. The difference in the voxel-wise thresholds between GMV and ReHo was caused by the voxel size difference in the two parameter maps. Similarly, we extracted IDP values for ReHo from each participant. A two-sample t-test was also used to test the FA differences between matched OC and non-OC pairs. In these analyses, we corrected for the 87 white matter fiber tracts, three neuroimaging features, and three data types, resulting in an FWE-corrected threshold of \u003cem\u003eP \u0026lt;\u003c/em\u003e 0.05/87/3/3 = 6.38 \u0026times; 10\u003csup\u003e-5\u003c/sup\u003e. We also defined the fiber tracts with significant FA difference as IDPs for the subsequent analyses. For the special fiber tracts of interest that exhibited FA differences between OC and non-OC groups, we further calculated the FC between the connected brain regions of each fiber tract. A two-sample t-test was used to compare the FC difference between OC and non-OC.\u003c/p\u003e\n\u003ch3\u003eComparisons of behavioral measures between OC and non-OC\u003c/h3\u003e\n\u003cp\u003eA two-sample t-test was applied to test the differences in behavioral variables between the well-matched OC and non-OC groups. We corrected for the three data types and 34 behavioral measurements by setting a Bonferroni-corrected threshold of \u003cem\u003eP \u0026lt;\u003c/em\u003e 0.05/34/3 = 4.90 \u0026times; 10\u003csup\u003e-4\u003c/sup\u003e. We defined behavioral traits with significant intergroup differences as significant behavioral phenotypes.\u003c/p\u003e\n\u003ch3\u003eComparison of growth environmental exposures between OC and non-OC\u003c/h3\u003e\n\u003cp\u003eA two-sample t-test was applied to test the differences in the 32 growth environmental exposures between OC and non-OC. We corrected for three data types and 32 exposures by setting a Bonferroni-corrected threshold of \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05/32/3 = 5.20 \u0026times; 10\u003csup\u003e-4\u003c/sup\u003e. The growth environmental exposures with significant differences between OC and non-OC were defined as proximal environmental exposures (PEEs).\u003c/p\u003e\n\u003ch3\u003ePEE factor construction and comparison\u003c/h3\u003e\n\u003cp\u003eDue to the high correlations between PEEs, we performed exploratory factor analysis (EFA) implemented in the \u0026ldquo;psych\u0026rdquo; 2.2.9 (https://cran.r-project.org/package=psych) from R 4.1.2 to categorize PEEs into factors. Spearman correlations were calculated between PEEs and the appropriateness for EFA was confirmed by the KMO test and Bartlett\u0026rsquo;s test of sphericity. The criteria of very simple structure (VSS) (\u003cem\u003e75\u003c/em\u003e) and minimum average partial (MAP) (\u003cem\u003e76\u003c/em\u003e) were used to select the optimal number of factors. The principal axis factoring was used to identify a latent factor structure of PEEs, and varimax rotation was then employed to clarify the relationship among latent variables. Based on the optimal EFA model, we defined PEE indicators with factor loading \u0026gt; 0.45 as the main contributors for each PEE factor. We named each PEE factor according to the main contributors (PEEs) of the factor. To account for uncertainty in within-sample prediction, the 10-fold cross-validation confirmatory factor analysis (CFA) was performed to obtain the out-of-sample PEE factor scores. In each iteration, 90% participants were used to estimate factor loadings of the CFA model, and the obtained factor loadings were used to predict PEE factor scores of other 10% participants. This process was iterated 10 times to obtain the out-of-sample PEE factor scores of all participants. The model fit was evaluated by a combination of root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMR). To test the stability of CFA models, we calculated Spearman rank correlation between PEE factor scores estimated in the 10-fold cross-validation and those estimated using the full sample. A two-sample t-test was employed to compare the out-of-sample PEE factor scores between OC and non-OC (\u003cem\u003eP\u003csub\u003ec\u003c/sub\u003e\u003c/em\u003e \u0026lt; 0.05, adjusting for the number of PEE factors and three data types).\u003c/p\u003e\n\u003ch3\u003eCausal mediation analysis (CMA)\u003c/h3\u003e\n\u003cp\u003eCMA was implemented using the R package \u0026ldquo;mediation\u0026rdquo; 4.5.0 (https://cran.r-project.org/web/packages/mediation/index.html) (\u003cem\u003e77\u003c/em\u003e) to test (a) whether a PEE factor mediated or moderated an association of GWS (treatment, X) with IDP or behavioral phenotype (outcome, Y); (b) whether an IDP mediated or moderated the association between GWS (treatment, X) and behavioral phenotype (outcome, Y). The direct, mediation and moderation effects were estimated using a quasi-Bayesian Monte Carlo simulation with 10,000 iterations. As IDPs were highly correlated, we estimated the effective number of IDPs (\u003cem\u003e78\u003c/em\u003e). In (a), we corrected for the number of PEE factors and the sum of the effective number of IDPs and the number of behavioral variables. In (b), we corrected for the effective number of IDPs and the number of behavioral variables.\u003c/p\u003e\n\u003ch3\u003eStructural equation model (SEM)\u003c/h3\u003e\n\u003cp\u003eSEM was implemented using the R package \u0026ldquo;lavaan\u0026rdquo; 0.6-15 (https://cran.r-project.org/web/packages/lavaan/index.html) (\u003cem\u003e79\u003c/em\u003e) based on the presumed effects of GWS and PEE factors on IDPs and behavioral phenotypes (\u003cstrong\u003eSupplementary Fig. 11\u003c/strong\u003e). SEM can estimate effect while taking other predictors into account and allow for simultaneously estimating the associations among multiple variables while reducing the measurement error by using latent variables (\u003cem\u003e79\u003c/em\u003e). SEM consists of two sub-models: the measurement model is used to define latent variables, and the structure model is used to specify direct and indirect pathways. A two-sided \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 was considered statistically significant in SEM. The model fit was evaluated by a combination of CFI, RMSEA, and SRMR. From SEM, we obtained the portion of the variance of each phenotype explained by the model, the relative effect of PEEs, as well as the direct, indirect, and total effects of GWS on outcomes. We integrated the results from CMA and SEM to identify consistent pathways from GWS to IDPs and behavioral phenotypes. As CMA used individual IDP while SEM used the latent variable of IDPs, we considered a causal pathway consistent between CMA and SEM when the individual IDP in CMA was one of the IDPs of the corresponding latent variable in SEM, and the indirect effect was in the same direction.\u003c/p\u003e\n"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was partially supported by the National Natural Science Foundation of China (grant 82030053 and 81425013 to Chunshui Yu, and grant 82202093 to Jie Tang).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: C.Y, J.T, J.Z, W.L, and M.W\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMethodology: C.Y, J.T, J-H.G, Z.G and W.Q\u003c/p\u003e\n\u003cp\u003eInvestigation: J.T, J.G, B.Z, W.Z, S,Q, G.C, Y.Y, W.L, H.Z, B.G, X.X, Y.Y, T.H, Z.Y, Q.Z, F.L, M.L, S.W, Q.X, J.X, J.F, Y.J, N.L, P.Z, D.S, C.W, S.L, Z.Y, F.C, W.S, W.M, D.W, J.X, X.Z, K.X, X-N.Z, L.Z, and Z.Y\u003c/p\u003e\n\u003cp\u003eVisualization: J.T and W.Q\u003c/p\u003e\n\u003cp\u003eFunding acquisition: C.Y, J.T, and W.L\u003c/p\u003e\n\u003cp\u003eProject administration: C.Y\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSupervision: C.Y, J-H.G and W.Q\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; original draft: J.T,\u0026nbsp;J-H.G, and J.Z\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; review \u0026amp; editing: C.Y,\u0026nbsp;J-H.G\u0026nbsp;and J.T\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have declared that no competing interests exist.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and materials availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe made use of publicly available software and tools. The publicly available tools used in our analyses are described in the Methods section. The voxel-wise neuroimaging statistical maps are available at https://figshare.com/articles/dataset/Voxelwise_difference/24716832. Individual-level data will be shared with the researchers after their proposal has been approved by the committee. We encourage researchers or parties interested in collaboration for non-commercial use to email to C.Y. ([email protected]). All data requests will be reviewed by the committee of CHIMGEN to verify whether the request is subject to any intellectual property or confidentiality obligations. The use of data must comply with the requirements of the Human Genetics Resources Administration of China and other country-specific or region-specific regulations.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eT. Hesketh, X. Zhou, Y. Wang, The End of the One-Child Policy: Lasting Implications for China. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e314\u003c/strong\u003e, 2619-2620 (2015).\u003c/li\u003e\n\u003cli\u003eY. Wang, V. L. Fong, Little Emperors and the 4:2:1 Generation: China\u0026apos;s Singletons. \u003cem\u003eJournal of the American Academy of Child \u0026amp; Adolescent Psychiatry\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 1137-1139 (2009).\u003c/li\u003e\n\u003cli\u003eT. Hesketh, L. Lu, Z. W. Xing, The effect of China\u0026apos;s one-child family policy after 25 years. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cstrong\u003e353\u003c/strong\u003e, 1171-1176 (2005).\u003c/li\u003e\n\u003cli\u003eM. Hvistendahl, China. Making a selfish generation by fiat. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e339\u003c/strong\u003e, 131 (2013).\u003c/li\u003e\n\u003cli\u003eL. Cameron, N. Erkal, L. Gangadharan, X. Meng, Little emperors: behavioral impacts of China\u0026apos;s One-Child Policy. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e339\u003c/strong\u003e, 953-957 (2013).\u003c/li\u003e\n\u003cli\u003eS. Cheng, C. Jia, Y. Wang, Only Children Were Associated with Anxiety and Depressive Symptoms among College Students in China. \u003cem\u003eInt J Environ Res Public Health\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 4035 (2020).\u003c/li\u003e\n\u003cli\u003eD. L. Poston, A. T. Falbo, Academic performance and personality traits of Chinese children: \u0026quot;Onlies\u0026quot; versus others. \u003cem\u003eAmerican Journal of Sociology\u003c/em\u003e \u003cstrong\u003e96\u003c/strong\u003e, 433-451 (1990).\u003c/li\u003e\n\u003cli\u003eA. J. Fuligni, W. Zhang, Attitudes toward family obligation among adolescents in contemporary urban and rural China. \u003cem\u003eChild Dev\u003c/em\u003e \u003cstrong\u003e75\u003c/strong\u003e, 180-192 (2004).\u003c/li\u003e\n\u003cli\u003eZ.-y. Chen, R. X. Liu, Comparing adolescent only children with those who have siblings on academic related outcomes and psychosocial adjustment. \u003cem\u003eChild Development Research\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e, 1-10 (2014).\u003c/li\u003e\n\u003cli\u003eW. Wang, J. Zhang, D. A. Hennessy, W. Yin, Do only-children communicate better than non-only children? A study of medical students in China. \u003cem\u003eJ. Comp. Fam. Stud\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 84\u0026ndash;109 (2020).\u003c/li\u003e\n\u003cli\u003eY. Guo, J. Zhang, N. Zhang, Character strengths and their influencing factors among nursing students in Changsha, China: the only-child versus non-only-child. \u003cem\u003eArch Psychiatr Nurs\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 365-371 (2015).\u003c/li\u003e\n\u003cli\u003eR. X. Liu, W. Lin, Z.-y. Chen, School performance, peer association, psychological and behavioral adjustments: A comparison between Chinese adolescents with and without siblings. \u003cem\u003eJournal of Adolescence\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 411-417 (2010).\u003c/li\u003e\n\u003cli\u003eC. Jia, Z. Yang, T. Xin, Y. Li, Y. Wang, T. Yang, Differences in school performance between only children and non-only children: evidence from China. \u003cem\u003eFront Psychol\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 608704 (2021).\u003c/li\u003e\n\u003cli\u003eN. Liu, Y. Chen, X. Yang, Y. Hu, Do demographic characteristics make differences? Demographic characteristics as moderators in the associations between only child status and cognitive/non-cognitive outcomes in China. \u003cem\u003eFront Psychol\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 423 (2017).\u003c/li\u003e\n\u003cli\u003eY. Cao, L. Huang, T. Si, N. Q. Wang, M. Qu, X. Y. Zhang, The role of only-child status in the psychological impact of COVID-19 on mental health of Chinese adolescents. \u003cem\u003eJ Affect Disord\u003c/em\u003e \u003cstrong\u003e282\u003c/strong\u003e, 316-321 (2021).\u003c/li\u003e\n\u003cli\u003eT. Falbo, Only Children: An Updated Review. \u003cem\u003eJournal of Individual Psychology\u003c/em\u003e \u003cstrong\u003e68\u003c/strong\u003e, 38-49 (2012).\u003c/li\u003e\n\u003cli\u003eX. Zhao, X. Ma, Y. Yao, C. Wan, E. Ng, China\u0026apos;s little emperors show signs of success. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e339\u003c/strong\u003e, 905-906 (2013).\u003c/li\u003e\n\u003cli\u003eY. Chen, F. Dangardt, P. Friberg, Association between childhood BMI trajectories and cardiometabolic risk and mental health problems at the age of 13 years: the cohort STudy of Adolescence Resilience and Stress (STARS). \u003cem\u003eLancet Glob Health\u003c/em\u003e \u003cstrong\u003e11 Suppl 1\u003c/strong\u003e, S3 (2023).\u003c/li\u003e\n\u003cli\u003eE. Dennis, P. Manza, N. D. Volkow, Socioeconomic status, BMI, and brain development in children. \u003cem\u003eTransl Psychiatry\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 33 (2022).\u003c/li\u003e\n\u003cli\u003eT. Falbo, S. Y. Hooper, China\u0026apos;s only children and psychopathology: A quantitative synthesis. \u003cem\u003eAm J Orthopsychiatry\u003c/em\u003e \u003cstrong\u003e85\u003c/strong\u003e, 259-274 (2015).\u003c/li\u003e\n\u003cli\u003eC. Liu, T. Munakata, F. N. Onuoha, Mental health condition of the only-child: a study of urban and rural high school students in China. \u003cem\u003eAdolescence\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, 831-845 (2005).\u003c/li\u003e\n\u003cli\u003eD. Tomasi, N. D. Volkow, Associations of family income with cognition and brain structure in USA children: prevention implications. \u003cem\u003eMol Psychiatry\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 6619-6629 (2021).\u003c/li\u003e\n\u003cli\u003eJ. Yang, X. Hou, D. Wei, K. Wang, Y. Li, J. Qiu, Only-child and non-only-child exhibit differences in creativity and agreeableness: evidence from behavioral and anatomical structural studies. \u003cem\u003eBrain Imaging Behav\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 493-502 (2017).\u003c/li\u003e\n\u003cli\u003eQ. Xu, L. Guo, J. Cheng, M. Wang, Z. Geng, W. Zhu, B. Zhang, W. Liao, S. Qiu, H. Zhang, X. Xu, Y. Yu, B. Gao, T. Han, Z. Yao, G. Cui, F. Liu, W. Qin, Q. Zhang, M. J. Li, M. Liang, F. Chen, J. Xian, J. Li, J. Zhang, X. N. Zuo, D. Wang, W. Shen, Y. Miao, F. Yuan, S. Lui, X. Zhang, K. Xu, L. J. Zhang, Z. Ye, C. Yu, C. Consortium, CHIMGEN: a Chinese imaging genetics cohort to enhance cross-ethnic and cross-geographic brain research. \u003cem\u003eMol Psychiatry\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 517-529 (2020).\u003c/li\u003e\n\u003cli\u003eJ. P. Fortin, D. Parker, B. Tunc, T. Watanabe, M. A. Elliott, K. Ruparel, D. R. Roalf, T. D. Satterthwaite, R. C. Gur, R. E. Gur, R. T. Schultz, R. Verma, R. T. Shinohara, Harmonization of multi-site diffusion tensor imaging data. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e161\u003c/strong\u003e, 149-170 (2017).\u003c/li\u003e\n\u003cli\u003eF. C. Yeh, Population-based tract-to-region connectome of the human brain and its hierarchical topology. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 4933 (2022).\u003c/li\u003e\n\u003cli\u003eS. S. Panesar, J. T. A. Belo, F. C. Yeh, J. C. Fernandez-Miranda, Structure, asymmetry, and connectivity of the human temporo-parietal aslant and vertical occipital fasciculi. \u003cem\u003eBrain Struct Funct\u003c/em\u003e \u003cstrong\u003e224\u003c/strong\u003e, 907-923 (2019).\u003c/li\u003e\n\u003cli\u003eD. B. Downey, Number of siblings and intellectual development: The resource dilution explanation. \u003cem\u003eAmerican Psychologist\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 497-504 (2001).\u003c/li\u003e\n\u003cli\u003eR. A. I. Bethlehem, J. Seidlitz, S. R. White, J. W. Vogel, K. M. Anderson, C. Adamson, S. Adler, G. S. Alexopoulos, E. Anagnostou, A. Areces-Gonzalez, D. E. Astle, B. Auyeung, M. Ayub, J. Bae, G. Ball, S. Baron-Cohen, R. Beare, S. A. Bedford, V. Benegal, F. Beyer, J. Blangero, M. Blesa Cabez, J. P. Boardman, M. Borzage, J. F. Bosch-Bayard, N. Bourke, V. D. Calhoun, M. M. Chakravarty, C. Chen, C. Chertavian, G. Chetelat, Y. S. Chong, J. H. Cole, A. Corvin, M. Costantino, E. Courchesne, F. Crivello, V. L. Cropley, J. Crosbie, N. Crossley, M. Delarue, R. Delorme, S. Desrivieres, G. A. Devenyi, M. A. Di Biase, R. Dolan, K. A. Donald, G. Donohoe, K. Dunlop, A. D. Edwards, J. T. Elison, C. T. Ellis, J. A. Elman, L. Eyler, D. A. Fair, E. Feczko, P. C. Fletcher, P. Fonagy, C. E. Franz, L. Galan-Garcia, A. Gholipour, J. Giedd, J. H. Gilmore, D. C. Glahn, I. M. Goodyer, P. E. Grant, N. A. Groenewold, F. M. Gunning, R. E. Gur, R. C. Gur, C. F. Hammill, O. Hansson, T. Hedden, A. Heinz, R. N. Henson, K. Heuer, J. Hoare, B. Holla, A. J. Holmes, R. Holt, H. Huang, K. Im, J. Ipser, C. R. Jack, Jr., A. P. Jackowski, T. Jia, K. A. Johnson, P. B. Jones, D. T. Jones, R. S. Kahn, H. Karlsson, L. Karlsson, R. Kawashima, E. A. Kelley, S. Kern, K. W. Kim, M. G. Kitzbichler, W. S. Kremen, F. Lalonde, B. Landeau, S. Lee, J. Lerch, J. D. Lewis, J. Li, W. Liao, C. Liston, M. V. Lombardo, J. Lv, C. Lynch, T. T. Mallard, M. Marcelis, R. D. Markello, S. R. Mathias, B. Mazoyer, P. McGuire, M. J. Meaney, A. Mechelli, N. Medic, B. Misic, S. E. Morgan, D. Mothersill, J. Nigg, M. Q. W. Ong, C. Ortinau, R. Ossenkoppele, M. Ouyang, L. Palaniyappan, L. Paly, P. M. Pan, C. Pantelis, M. M. Park, T. Paus, Z. Pausova, D. Paz-Linares, A. Pichet Binette, K. Pierce, X. Qian, J. Qiu, A. Qiu, A. Raznahan, T. Rittman, A. Rodrigue, C. K. Rollins, R. Romero-Garcia, L. Ronan, M. D. Rosenberg, D. H. Rowitch, G. A. Salum, T. D. Satterthwaite, H. L. Schaare, R. J. Schachar, A. P. Schultz, G. Schumann, M. Scholl, D. Sharp, R. T. Shinohara, I. Skoog, C. D. Smyser, R. A. Sperling, D. J. Stein, A. Stolicyn, J. Suckling, G. Sullivan, Y. Taki, B. Thyreau, R. Toro, N. Traut, K. A. Tsvetanov, N. B. Turk-Browne, J. J. Tuulari, C. Tzourio, E. Vachon-Presseau, M. J. Valdes-Sosa, P. A. Valdes-Sosa, S. L. Valk, T. van Amelsvoort, S. N. Vandekar, L. Vasung, L. W. Victoria, S. Villeneuve, A. Villringer, P. E. Vertes, K. Wagstyl, Y. S. Wang, S. K. Warfield, V. Warrier, E. Westman, M. L. Westwater, H. C. Whalley, A. V. Witte, N. Yang, B. Yeo, H. Yun, A. Zalesky, H. J. Zar, A. Zettergren, J. H. Zhou, H. Ziauddeen, A. Zugman, X. N. Zuo, B. R, Aibl, I. Alzheimer\u0026apos;s Disease Neuroimaging, I. Alzheimer\u0026apos;s Disease Repository Without Borders, C. Team, C. A. N. Cam, Ccnp, Cobre, cVeda, E. D. B. A. W. Group, P. Developing Human Connectome, FinnBrain, S. Harvard Aging Brain, Imagen, Kne, A. Mayo Clinic Study of, Nspn, Pond, P.-A. R. Group, Vetsa, E. T. Bullmore, A. F. Alexander-Bloch, Brain charts for the human lifespan. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e604\u003c/strong\u003e, 525-533 (2022).\u003c/li\u003e\n\u003cli\u003eT. Paus, M. Keshavan, J. N. Giedd, Why do many psychiatric disorders emerge during adolescence? \u003cem\u003eNat Rev Neurosci\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 947-957 (2008).\u003c/li\u003e\n\u003cli\u003eE. R. Sowell, P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, A. W. Toga, Longitudinal mapping of cortical thickness and brain growth in normal children. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 8223-8231 (2004).\u003c/li\u003e\n\u003cli\u003eD. A. Lewis, Development of the prefrontal cortex during adolescence: insights into vulnerable neural circuits in schizophrenia. \u003cem\u003eNeuropsychopharmacology\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 385-398 (1997).\u003c/li\u003e\n\u003cli\u003eL. C. Rutherford, S. B. Nelson, G. G. Turrigiano, BDNF has opposite effects on the quantal amplitude of pyramidal neuron and interneuron excitatory synapses. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 521-530 (1998).\u003c/li\u003e\n\u003cli\u003eS. Whittle, N. Vijayakumar, J. G. Simmons, M. Dennison, O. Schwartz, C. Pantelis, L. Sheeber, M. L. Byrne, N. B. Allen, Role of Positive Parenting in the Association Between Neighborhood Social Disadvantage and Brain Development Across Adolescence. \u003cem\u003eJAMA Psychiatry\u003c/em\u003e \u003cstrong\u003e74\u003c/strong\u003e, 824-832 (2017).\u003c/li\u003e\n\u003cli\u003eS. Whittle, J. G. Simmons, M. Dennison, N. Vijayakumar, O. Schwartz, M. B. Yap, L. Sheeber, N. B. Allen, Positive parenting predicts the development of adolescent brain structure: a longitudinal study. \u003cem\u003eDev Cogn Neurosci\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 7-17 (2014).\u003c/li\u003e\n\u003cli\u003eJ. Xu, X. Liu, Q. Li, R. Goldblatt, W. Qin, F. Liu, C. Chu, Q. Luo, A. Ing, L. Guo, N. Liu, H. Liu, C. Huang, J. Cheng, M. Wang, Z. Geng, W. Zhu, B. Zhang, W. Liao, S. Qiu, H. Zhang, X. Xu, Y. Yu, B. Gao, T. Han, G. Cui, F. Chen, J. Xian, J. Li, J. Zhang, X. N. Zuo, D. Wang, W. Shen, Y. Miao, F. Yuan, S. Lui, X. Zhang, K. Xu, L. Zhang, Z. Ye, T. Banaschewski, G. J. Barker, A. L. W. Bokde, H. Flor, A. Grigis, H. Garavan, P. Gowland, A. Heinz, R. Br\u0026uuml;hl, J. L. Martinot, E. Artiges, F. Nees, D. P. Orfanos, H. Lemaitre, T. Paus, L. Poustka, L. Robinson, S. Hohmann, J. H. Fr\u0026ouml;hner, M. N. Smolka, H. Walter, R. Whelan, J. Winterer, K. Patrick, V. Calhoun, M. J. Li, M. Liang, P. Gong, E. D. Barker, N. Clinton, A. Marquand, L. Yu, C. Yu, G. Schumann, Global urbanicity is associated with brain and behaviour in young people. \u003cem\u003eNat Hum Behav\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 279-293 (2022).\u003c/li\u003e\n\u003cli\u003eL. Machlin, K. A. McLaughlin, M. A. Sheridan, Brain structure mediates the association between socioeconomic status and attention-deficit/hyperactivity disorder. \u003cem\u003eDev Sci\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, e12844 (2020).\u003c/li\u003e\n\u003cli\u003eY. C. Lu, K. Kapse, N. Andersen, J. Quistorff, C. Lopez, A. Fry, J. Cheng, N. Andescavage, Y. Wu, K. Espinosa, G. Vezina, A. du Plessis, C. Limperopoulos, Association between socioeconomic status and in utero fetal brain development. \u003cem\u003eJAMA Netw Open\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, e213526 (2021).\u003c/li\u003e\n\u003cli\u003eA. J. Dufford, G. W. Evans, J. Dmitrieva, J. E. Swain, I. Liberzon, P. Kim, Prospective associations, longitudinal patterns of childhood socioeconomic status, and white matter organization in adulthood. \u003cem\u003eHum Brain Mapp\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 3580-3593 (2020).\u003c/li\u003e\n\u003cli\u003eD. Zhao, Z. He, Y. Tian, H. Liu, Differences in cognitive and non-cognitive results between only-child and non-only-child children: analysis of popensity scores based on large-scale assessment. \u003cem\u003eChildren\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 807 (2022).\u003c/li\u003e\n\u003cli\u003eF. Van Overwalle, K. Baetens, P. Marien, M. Vandekerckhove, Social cognition and the cerebellum: a meta-analysis of over 350 fMRI studies. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e86\u003c/strong\u003e, 554-572 (2014).\u003c/li\u003e\n\u003cli\u003eD. Reiss, J. M. Ganiban, L. D. Leve, J. M. Neiderhiser, D. S. Shaw, M. N. Natsuaki, Parenting in the Context of the Child: Genetic and Social Processes. \u003cem\u003eMonogr Soc Res Child Dev\u003c/em\u003e \u003cstrong\u003e87\u003c/strong\u003e, 7-188 (2022).\u003c/li\u003e\n\u003cli\u003eZ. H. He, M. D. Li, X. Y. Ma, C. J. Liu, Family socioeconomic status and social media addiction in female college students: the mediating role of impulsiveness and inhibitory control. \u003cem\u003eJ Genet Psychol\u003c/em\u003e \u003cstrong\u003e182\u003c/strong\u003e, 60-74 (2021).\u003c/li\u003e\n\u003cli\u003eJ. C. Rajapakse, J. N. Giedd, J. L. Rapoport, Statistical approach to segmentation of single-channel cerebral MR images. \u003cem\u003eIEEE Trans Med Imaging\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 176-186 (1997).\u003c/li\u003e\n\u003cli\u003eJ. Ashburner, A fast diffeomorphic image registration algorithm. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 95-113 (2007).\u003c/li\u003e\n\u003cli\u003eS. M. Smith, M. Jenkinson, H. Johansen-Berg, D. Rueckert, T. E. Nichols, C. E. Mackay, K. E. Watkins, O. Ciccarelli, M. Z. Cader, P. M. Matthews, T. E. Behrens, Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 1487-1505 (2006).\u003c/li\u003e\n\u003cli\u003eY. Chao-Gan, Z. Yu-Feng, DPARSF: A MATLAB Toolbox for \u0026quot;Pipeline\u0026quot; Data Analysis of Resting-State fMRI. \u003cem\u003eFront Syst Neurosci\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 13 (2010).\u003c/li\u003e\n\u003cli\u003eM. Jenkinson, P. Bannister, M. Brady, S. Smith, Improved optimization for the robust and accurate linear registration and motion correction of brain images. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 825-841 (2002).\u003c/li\u003e\n\u003cli\u003eR. H. R. Pruim, M. Mennes, D. van Rooij, A. Llera, J. K. Buitelaar, C. F. Beckmann, ICA-AROMA: A robust ICA-based strategy for removing motion artifacts from fMRI data. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e112\u003c/strong\u003e, 267-277 (2015).\u003c/li\u003e\n\u003cli\u003eD. C. Delis, J. Freeland, J. H. Kramer, E. Kaplan, Integrating clinical assessment with cognitive neuroscience: construct validation of the California Verbal Learning Test. \u003cem\u003eJ Consult Clin Psychol\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 123-130 (1988).\u003c/li\u003e\n\u003cli\u003eA. M. Owen, K. M. McMillan, A. R. Laird, E. Bullmore, N-back working memory paradigm: a meta-analysis of normative functional neuroimaging studies. \u003cem\u003eHum Brain Mapp\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 46-59 (2005).\u003c/li\u003e\n\u003cli\u003eM. S. Shin, S. Y. Park, S. R. Park, S. H. Seol, J. S. Kwon, Clinical and empirical applications of the Rey-Osterrieth Complex Figure Test. \u003cem\u003eNat Protoc\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 892-899 (2006).\u003c/li\u003e\n\u003cli\u003eY. XIN, \u003cem\u003eMCCB Chinese norm Manual(Chinese Edition)\u003c/em\u003e. (Peking University Medical Press, 2014).\u003c/li\u003e\n\u003cli\u003eJ. N. Kaufman, T. J. Ross, E. A. Stein, H. Garavan, Cingulate hypoactivity in cocaine users during a GO-NOGO task as revealed by event-related functional magnetic resonance imaging. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 7839-7843 (2003).\u003c/li\u003e\n\u003cli\u003eS. Pulos, Elison, J., Lennon, R.J, The hierarchical structure of the Interpersonal Reactivity Index. \u003cem\u003eSoc Behav Pers\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 16 (2004).\u003c/li\u003e\n\u003cli\u003eA. Canevello, J. Crocker, Creating good relationships: responsiveness, relationship quality, and interpersonal goals. \u003cem\u003eJ Pers Soc Psychol\u003c/em\u003e \u003cstrong\u003e99\u003c/strong\u003e, 78-106 (2010).\u003c/li\u003e\n\u003cli\u003eX. Wang, Y. Wang, T. Xin, The Psychometric Properties of the Chinese Version of the Beck Depression Inventory-II With Middle School Teachers. \u003cem\u003eFront Psychol\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 548965 (2020).\u003c/li\u003e\n\u003cli\u003eD. T. Shek, The Chinese version of the State-Trait Anxiety Inventory: its relationship to different measures of psychological well-being. \u003cem\u003eJ Clin Psychol\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 349-358 (1993).\u003c/li\u003e\n\u003cli\u003eD. Watson, L. A. Clark, A. Tellegen, Development and validation of brief measures of positive and negative affect: the PANAS scales. \u003cem\u003eJ Pers Soc Psychol\u003c/em\u003e \u003cstrong\u003e54\u003c/strong\u003e, 1063-1070 (1988).\u003c/li\u003e\n\u003cli\u003eK. T. Cohen S, Mermelstein R, A Global Measure of Perceived Stress. \u003cem\u003eJ Health Soc Behav\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 12 (1983).\u003c/li\u003e\n\u003cli\u003eW. Pavot, \u0026amp; Diener, E, The Satisfaction With Life Scale and the emerging construct of life satisfaction. \u003cem\u003eJournal of Positive Psychology\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 16 (2008).\u003c/li\u003e\n\u003cli\u003eP. T. Costa Jr, R. R. McCrae, in \u003cem\u003eThe SAGE handbook of personality theory and assessment, Vol 2: Personality measurement and testing.\u003c/em\u003e (Sage Publications, Inc, Thousand Oaks, CA, US, 2008), pp. 179-198.\u003c/li\u003e\n\u003cli\u003eC. R. Cloninger, T. R. Przybeck, D. M. Svrakic, The Tridimensional Personality Questionnaire: U.S. normative data. \u003cem\u003ePsychol Rep\u003c/em\u003e \u003cstrong\u003e69\u003c/strong\u003e, 1047-1057 (1991).\u003c/li\u003e\n\u003cli\u003eJ. H. Patton, M. S. Stanford, E. S. Barratt, Factor structure of the Barratt impulsiveness scale. \u003cem\u003eJ Clin Psychol\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 768-774 (1995).\u003c/li\u003e\n\u003cli\u003eV. Rolfe, Tailoring a measurement model of socioeconomic status: Applying the alignment optimization method to 15 years of PISA. \u003cem\u003eInternational Journal of Educational Research\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 101723 (2021).\u003c/li\u003e\n\u003cli\u003eA. van Donkelaar, R. V. Martin, M. Brauer, N. C. Hsu, R. A. Kahn, R. C. Levy, A. Lyapustin, A. M. Sayer, D. M. Winker, Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors. \u003cem\u003eEnvironmental Science \u0026amp; Technology\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 3762-3772 (2016).\u003c/li\u003e\n\u003cli\u003eJ. A. Geddes, R. V. Martin, B. L. Boys, A. van Donkelaar, Long-Term Trends Worldwide in Ambient NO2 Concentrations Inferred from Satellite Observations. \u003cem\u003eEnviron Health Perspect\u003c/em\u003e \u003cstrong\u003e124\u003c/strong\u003e, 281-289 (2016).\u003c/li\u003e\n\u003cli\u003eG. Parker, H. Tupling, L. B. Brown, A parental bonding instrument. \u003cem\u003eBritish Journal of Medical Psychology\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, 1-10 (1979).\u003c/li\u003e\n\u003cli\u003eJ. Canty-Mitchell, G. D. Zimet, Psychometric properties of the Multidimensional Scale of Perceived Social Support in urban adolescents. \u003cem\u003eAm J Community Psychol\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 391-400 (2000).\u003c/li\u003e\n\u003cli\u003eX. F. Zhao, Y. L. Zhang, L. F. Li, Y. F. Zhou, H. Z. Li, S. C. Yang, Reliability and validity of the Chinese version of childhood trauma questionnaire. \u003cem\u003eChinese Journal of Clinical Rehabilitation\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 105-107 (2005).\u003c/li\u003e\n\u003cli\u003eN. Liu, L. Zhang, T. Tian, J. Cheng, B. Zhang, S. Qiu, Z. Geng, G. Cui, Q. Zhang, W. Liao, Y. Yu, H. Zhang, B. Gao, X. Xu, T. Han, Z. Yao, W. Qin, F. Liu, M. Liang, Q. Xu, J. Fu, J. Xu, W. Zhu, P. Zhang, W. Li, D. Shi, C. Wang, S. Lui, Z. Yan, F. Chen, J. Li, J. Zhang, D. Wang, W. Shen, Y. Miao, J. Xian, J. H. Gao, X. Zhang, M. J. Li, K. Xu, X. N. Zuo, M. Wang, Z. Ye, C. Yu, C. Consortium, Cross-ancestry genome-wide association meta-analyses of hippocampal and subfield volumes. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e55\u003c/strong\u003e, 1126-1137 (2023).\u003c/li\u003e\n\u003cli\u003eE. A. Stuart, G. King, K. Imai, D. Ho, MatchIt: nonparametric preprocessing for parametric causal inference. \u003cem\u003eJournal of statistical software\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 1-28 (2011).\u003c/li\u003e\n\u003cli\u003eP. C. Austin, Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. \u003cem\u003eStat Med\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 3083-3107 (2009).\u003c/li\u003e\n\u003cli\u003eK. J. Worsley, J. E. Taylor, F. Tomaiuolo, J. Lerch, Unified univariate and multivariate random field theory. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e23 Suppl 1\u003c/strong\u003e, S189-195 (2004).\u003c/li\u003e\n\u003cli\u003eW. F. Velicer, Determining the number of components from the matrix of partial correlations. \u003cem\u003ePsychometrika\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, 321-327 (1976).\u003c/li\u003e\n\u003cli\u003eW. Revelle, T. Rocklin, Very simple structure: an alternative procedure for estimating the optimal number of interpretable factors. \u003cem\u003eMultivariate Behav Res\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 403-414 (1979).\u003c/li\u003e\n\u003cli\u003eD. Tingley, T. Yamamoto, K. Hirose, L. Keele, K. Imai, mediation: R Package for Causal Mediation Analysis. \u003cem\u003eJournal of Statistical Software\u003c/em\u003e \u003cstrong\u003e59\u003c/strong\u003e, 1-38 (2014).\u003c/li\u003e\n\u003cli\u003eC. S. Bretherton, M. Widmann, V. P. Dymnikov, J. M. Wallace, I. Blad\u0026eacute;, The effective number of spatial degrees of freedom of a time-varying field. \u003cem\u003eJournal of Climate\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 1990-2009 (1999).\u003c/li\u003e\n\u003cli\u003eY. Rosseel, lavaan: An R Package for Structural Equation Modeling. \u003cem\u003eJournal of Statistical Software\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 1-36 (2012).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3707132/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3707132/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWith the worldwide increase of only-child families, it is crucial to understand the influence of growth without siblings (GWS) on human health, but there is limited knowledge regarding the effects of GWS on human brain. Besides, existing studies have reported paradoxical associations between GWS and mental health, most likely due to mismatched confounders and overlooked growth environments. Here, using 2,305 pairs of individuals with and without siblings well matched in confounders, we comprehensively examined the impact of GWS on changes in adult brain structure, function, and behaviors, as well as pathways from GWS, growth environments to brain and behaviors. Our findings uncovered novel associations, including GWS being linked to higher language-fiber integrity, lower motor-fiber integrity, larger regional cerebellum volume, and lower frontotemporal spontaneous brain activity. Contrary to stereotypical associations between GWS and problem behaviors, we found positive correlations of GWS with neurocognition and mental health. Despite direct effects, GWS affects most adult brain and behavioral outcomes through modifiable environments, such as socioeconomic status, maternal care, and family support, suggesting targets for interventions to enhance children\u0026rsquo;s healthy growth.\u003c/p\u003e","manuscriptTitle":"How growth without siblings affects adult brain and behavior","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-05 18:10:05","doi":"10.21203/rs.3.rs-3707132/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-human-behaviour","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nathumbehav","sideBox":"Learn more about [Nature Human Behaviour](http://www.nature.com/nathumbehav/)","snPcode":"","submissionUrl":"","title":"Nature Human Behaviour","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"11ff3e29-8076-408f-972a-ba16b8a6459d","owner":[],"postedDate":"January 5th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":27217961,"name":"Biological sciences/Neuroscience/Social neuroscience"},{"id":27217962,"name":"Scientific community and society/Social sciences/Psychology/Human behaviour"},{"id":27217963,"name":"Earth and environmental sciences/Environmental sciences/Environmental impact"}],"tags":[],"updatedAt":"2025-04-01T07:13:31+00:00","versionOfRecord":{"articleIdentity":"rs-3707132","link":"https://doi.org/10.1038/s41562-025-02142-4","journal":{"identity":"nature-human-behaviour","isVorOnly":false,"title":"Nature Human Behaviour"},"publishedOn":"2025-03-31 04:00:00","publishedOnDateReadable":"March 31st, 2025"},"versionCreatedAt":"2024-01-05 18:10:05","video":"","vorDoi":"10.1038/s41562-025-02142-4","vorDoiUrl":"https://doi.org/10.1038/s41562-025-02142-4","workflowStages":[]},"version":"v1","identity":"rs-3707132","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3707132","identity":"rs-3707132","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}