The children in this study were from six kindergartens in Jinan and Dezhou, the province of Shandong in eastern China. The study was carried out over the course of 3 months, from June to August 2022. After giving mothers through the kindergartens informed consent forms and inviting them to attend lectures on family education and child development, the mothers filled out questionnaires. Each mother received a gift worth ¥30 in return for completing the questionnaire. All study procedures were authorized by Shandong Normal University’s Institutional Review Board.

Using the cluster sampling method, the original sample for this study consisted of 665 mothers and their TD children (4–6 years old). After recall, invalid questionnaires such as too many omissions, random responses and regular responses were excluded, and the final sample included 648 mother–child dyads with a questionnaire validity rate of 97.44%. Mothers’ ages ranged from 25 to 49 years (M = 36.53; SD = 7.22). According to data on socioeconomic status, 61.8% of mothers had a college degree or higher, while just 17.1% of mothers had completed high school. Regarding employment, 54.11% of the mothers were employed in working-class jobs (e.g., factory workers), while 45.89% of the mothers held a professional, technical, or managerial position (e.g., teachers). Mothers in the present study do not have ASD children. All the children were from public kindergartens. The mean age of the children (352 boys and 280 girls) was 5.03, SD = 0.82. 203 were 4-year-olds, 208 were 5-year-olds, and 221 were 6-year-olds. 28.3% of the children were the only child.

This study used SPSS 24.0 and its PROCESS software package for preliminary analysis of the data, which was used to check the completeness of the data and to calculate correlations. Preliminary analysis of the data was performed, which included descriptive statistics and correlation analysis of all data. Mediated effects and tests of mediated effects with moderation were analyzed using the SPSS process plug-in. First, the mediation model was tested to confirm the mechanism of influence of mothers’ negative emotion expressions between mothers’ ATs and children’s anxiety. We calculated the 95% CI based on 5,000 bootstrap samples. Second, the mediating effect with moderation was tested. If the regression coefficients of the interaction term between child gender and mothers’ negative emotion expressions were significant, the mediating model with child gender as the moderating variable was valid. Child gender was assigned separate values, boys = 1 and girls = 2.

Table 1 shows the relationship between the key factors and demographic variables of this study. There were significant positive correlations between mothers’ levels of ATs and their expressions of negative emotion, mothers’ ATs and children’s anxiety, and mothers’ expressions of negative emotion and children’s anxiety, as shown in Table 1. A multivariate analysis of variance (MANOVA) of 2 (child gender: boy and girl) × 3(child age: 4-, 5-, and 6-years old) was conducted for children’s anxiety. The results revealed that the main effect of child gender, the main effect of child age, and the interaction effect of child gender and age (ps > 0.05) for children’s anxiety were not significant.

The PROCESS macro procedure proposed by Hayes (2012) was used to test the hypothesized model. A simple mediation model was first tested on the data using Model 4 to explore the mediating effect of mothers’ negative emotion expressions between mothers’ ATs and children’s anxiety. Table 2 shows that mothers’ ATs are directedly and positively predictive of children’s anxiety (β = 0.275, p<0. 001) and are positively associated with mothers’ negative emotional expressions (β = 0.166, p<0. 001), which is positively related to children’s anxiety (β = 0.227, p<0. 001).We calculated the 95% CI based on 5,000 bootstrap samples. It confirmed the significant indirect effects of mothers’ negative emotional expressions in the relationship between their ATs and children’s anxiety [β = 0.038,95% CI (0.018, 0.061)]. These results indicated that mothers’ negative emotional expressions partially mediates the relationship between their ATs and children’s anxiety.

Using Model 14 of PROCESS macro (Hayes, 2012), we conducted the moderated mediation analysis. As shown in Table 3, the interaction of mothers’ negative emotional expressions and child gender has a significant effect on children’s anxiety (β = 0.062, p<0.001). Furthermore, a simple slope analysis was conducted and we plotted the effect of mothers’ negative emotional expressions on children’s anxiety separately for boys and girls (Figure 2). The simple slope test showed that mothers’ negative emotional expressions were a stronger predictor of anxiety in girls (β = 0.111, p<0.001) compared to boys (β = 0.049, p<0.05). The conditional indirect effect of mothers’ ATs on children’s anxiety via mothers’ negative emotional expressions was also calculated, as a function of child gender.

In addition, the bootstrapping results showed that the conditional indirect effect of mothers’ ATs on children’s anxiety via mothers’ negative emotional expressions was significant by child gender, with the effect being 0.062, SE = 0.030, 95% CI [0.005, 0.120]. Specifically, the conditional indirect effect in girls [β = 0.123, 95% CI (0.058, 0.206)] was stronger than that in boys [β = 0.054, 95% CI (0.007, 0.112)].

This study examined, for the first time among the general population, the direct predictive impact of maternal ATs on children’s anxiety. This might be explained by the cognitive similarities between individuals with high level of ATs and those with ASD. Studies of individuals from the general population who scored highly on ATs have shown a pattern of cognitive characteristics that is similar to but less severe than that commonly seen in ASD. For example, Saure and colleagues found the improved cognitive rigidity among individuals with high ATs (Saure et al., 2022). That means, as levels of maternal ATs increased, these mothers came to develop negative interpretations when confronted with ambiguity and uncertainty in their surroundings in the present study. Thus, these mothers might warn their children about potentially dangerous stimuli and circumstances or tell them that the surroundings are dangerous. Consequently, children’s interpretations of ambiguous circumstances might be adversely biased when parents convey to children their negative impressions or assessments of those events, such as by overemphasizing risk factors in the surroundings and by vocally conveying dangerous information regarding unclear stimuli and circumstances. Additionally, due to observational learning, children who repeatedly see their mothers’ cognitive rigidity in reaction to new and unclear environmental stimuli are more likely to experience anxiety and terror of these stimuli (Eley et al., 2015; Aktar et al., 2017). As a result, children’s level of anxiety may increase.

The second hypothesis of this study was validated by showing that mothers’ negative emotion expressions played a mediating role between their ATs and children’s anxiety. Consistent with existing research, the current investigation discovered a correlation between increased expressions of negative emotions and higher levels of maternal ATs. This may be explained by the similarity of emotional characteristics between high levels of ATs and ASD. As reviewed above, mood problems and ASD frequently coexist (White et al., 2018). The majority of studies have discovered that individuals with ASD frequently feel and display unpleasant emotions including anxiety and depression (Joshi, 2017; Taylor et al., 2022). Additionally, prior research has indicated that people with high levels of ATs exhibit emotional regulation deficiencies comparable to those in ASD, with a propensity to be unable to effectively moderate their negative emotions (Cai et al., 2019; Zhao et al., 2020).

Moreover, the present study found that maternal negative emotion expressions significantly and positively predicted children’s anxiety, which is consistent with our expectations. Research has shown that mothers’ emotional expression and regulation significantly affect the physical and mental health of family members (Morris et al., 2017; Risley et al., 2021). Mothers’ sadness and anxiety, in particular, have been shown to have intergenerational transmission consequences. On one hand, the intergenerational transmission model of depression and anxiety (Goodman, 2020; Åhlén et al., 2022) suggests that mothers with higher levels of depression and anxiety may cause themselves to neglect their children’s needs, especially by being less sensitive to their children’s emotional responses and less feedback and supportive, while generating more resentment, derogation, and blame for their children, which leads to higher levels of anxiety in their children. On the other hand, based on Social Learning Theory (Bandura, 1977; Fu et al., 2022), anxiety is acquired by observing mothers’ fearful, apprehensive reactions to environmental stimuli. For example, if children witness their moms reacting nervously to ambiguous or somewhat threatening stimuli, they will react anxiously when they are in the situation themselves. In addition, the model of emotional socialization suggests that parental emotional expression is a reference paradigm for children (Denham, 1998; Denham and Burton, 2003; Fu et al., 2022). For instance, if mothers express more negative emotions, children learn this pattern of expressions and tend to express negative emotions as well. Consequently, more expressions of negative emotions by mothers tend to create a depressing and negative emotional climate in families (Fredrick et al., 2019; Khoshroo and Seyed Mousavi, 2022), which in turn increases children’s negative emotional experiences and increases the incidence of internalizing problems (Hu et al., 2017), leading to increased levels of anxiety in children. These findings imply that mothers’ negative emotional expressions may be one mechanism by which maternal ATs predict children’s anxiety.

Consistent with the hypothesis, the present study discovered the moderating role of child gender. Specifically, girls experience more anxiety when their mothers expressed more negative emotions toward them compared to boys. There are three possible explanations as to the finding of the moderated effect of child gender. First, guided by the gender-specific social learning model, children view same-gender parents as role models and imitate their behavior patterns, so girls express anxiety with their mothers as targets of social learning. Second, girls are more sensitive than boys (Milfont and Sibley, 2016) and have stronger internal tension and anxiety when confronted with mothers’ anxiety. Finally, mothers may have more emotional interactions and connections with girls in Chinses family (Ming et al., 2020), which lead girls to be more vulnerable than boys to high levels of mothers’ negative emotions, resulting in experiencing more anxiety.

It is important to acknowledge that this study has several limitations, even if it offers a deeper understanding of the impact of maternal ATs on children’s anxiety in Chinese families for the first time.

First, this study used maternal ATs as an independent variable and did not consider whether and how paternal ATs affect children’s emotional problems. It has been suggested that there are gender differences in ATs, as evidenced by generally higher levels of ATs in males than in females (Baron-Cohen et al., 2001; Pisula et al., 2013). Therefore, future studies could select fathers as subjects to compare with the results of this study to fully explore the relationship between the level of parental ATs and children’s emotional problems. Second, it is challenging to make conclusions from the cross-sectional data on the associations between mothers’ ATs, children’s anxiety, and mothers’ negative emotional expression. To clearly establishing causal and temporal conclusions, longitudinal investigations are required in the future. Third, because preschoolers were too young to complete the questionnaires, this study only used mother-report data, which could artificially inflate some of the associations across variables due to shared method variance. To solve this issue, more research evaluating children’s anxiety using a variety of techniques will be required. Finally, the Family Self-Expression Inventory was used in the present study, which includes a variety of negative emotions but not concentrating on a certain feeling such as anxiety and depression. Future studies could select maternal anxiety and depression as mediating variables, respectively, to further explore the mechanisms by which maternal ATs affect children’s anxiety in detail. Moreover, it has been demonstrated that, like ASD, individuals with more ATs are insensitive to emotional information from others, have difficulty recognizing emotions (Cai et al., 2019), and lack certain empathic abilities (Vargas et al., 2005), which prevent mothers with high levels of ATs from detecting their children’s anxiety in a timely manner and from helping their children relieve their anxiety through acceptance, empathy, and understanding, thus leading to increased anxiety in their children. Future research needs to specifically analyze the mediating pathways through which mothers’ ATs affect their children’s anxiety in terms of emotion recognition and empathy.

Despite these drawbacks, the present study offers some insightful data and significant practical applications. First, this is the first study, to our knowledge, to explore the impact of ATs in general population in Chinese families. The findings of this study will expand the scope of research in the field of ATs and provide a research base for further exploration the impact of adult ASD individuals on family members, especially for their children. Second, the present study examined the association between maternal ATs and children’s anxiety in the general population. The results of this investigation provide a significant addition and expansion to earlier analyses of the mechanisms underlying children’s anxiety in family systems. Finally, the present study found that maternal negative emotional expressions mediated their ATs and children’s anxiety and the moderated effect of child gender. This conclusion emphasizes, especially for girls, that effective preventative measures aimed at lowering the level of children’s anxiety, particularly for mothers in the family who reported more ATs, should target not just the children themselves (e.g., children’s negative interpretation bias), dyadic factors (e.g., parent–child relationship), but also the function of maternal personality traits.