Longitudinal bidirectional relationships between maternal depressive/anxious symptoms and children’s tics

Background: Previous studies have revealed an association between maternal depressive/anxious symptoms and children’s tics. However, the longitudinal relationships between these symptoms remain unclear. We examined the longitudinal relationships between maternal depressive/anxious symptoms and children’s tics across early adolescence in a population-based sample. Methods: The participants consisted of 3,171 children and their mothers from the Tokyo Teen Cohort (TTC) study, a population-representative longitudinal study that was launched in Tokyo in 2012. Maternal depressive/anxious symptoms and children’s tics were examined using self-report questionnaires at the ages of 10 (time 1, T1) and 12 (time 2, T2). A cross-lagged model was used to explore the relationships between maternal depressive/anxious symptoms and children’s tics. Results: Higher levels of maternal depressive/anxious symptoms at T1 were related to an increased likelihood of children’s tics at T2 (β = .06, p = .001). Furthermore, the presence of children’s tics at T1 was positively related to maternal depressive/anxious symptoms at T2 (β = .06, p = .001). Conclusions: These ndings suggest a longitudinal bidirectional relationship between maternal depressive/anxious symptoms and children’s tics during early adolescence that may exacerbate each other over time and possibly create a vicious cycle. When an early adolescent has tics, it might be important to identify and treat related maternal depressive/anxious symptoms. depressive/anxious children’s tics inuence bidirectionally. bidirectional tic on the between depressive/anxious While there has been a consensus on the importance of family psychoeducation in the treatment of tics (7, 60), it is not known whether maternal psychiatric problems inuence the course of children’s tics. This study provides new insights for future research and practice.

these children, 10,234 were successfully contacted, and these children were invited to participate in the cohort study. Of these 10,234 children, 4,478 children participated in the baseline survey named the Tokyo Early Adolescence Survey (T-EAS). This baseline survey was conducted from October 2012 to January 2015, when the participants were approximately 10 years old (time 1, T1). Among the 4,478 participants in the T-EAS, candidates were chosen as participants for the second wave of the TTC study. For the sake of cohort management, the target number of participants to be included in the second wave of the TTC was 3,000 children. When choosing these TTC participants, an oversampling method was used instead of inclusion criteria, considering the low follow-up rate of families with low annual household incomes. Thus, 3,171 participants were extracted as targets for the second wave of the TTC study. This second wave of the TTC study was carried out from August 2014 to December 2016, at the time when the participants were approximately 12 years old (time 2, T2). Of the 3,171 children who were invited, 3,007 individuals participated in the second wave of the TTC study (follow-up rate 94.8%). In each wave of the data collection, trained interviewers visited the participants' homes. They distributed questionnaires to the children and primary parents (mostly mothers), and they conducted psychological tests on the children.

Ethical approval
Ethical approval for this study was obtained from the research ethics committees of the Tokyo Metropolitan Institute of Medical Science (Approval number: , The Graduate University for Advanced Studies, SOKENDAI (2012002), and the Graduate School of Medicine and Faculty of Medicine, The University of Tokyo (10057). We obtained informed assent from the children and written informed consent from their primary parents.

Measures Tics
We evaluated tics at T1 and T2. The participants' primary parents answered a questionnaire about the children's tics; this questionnaire has been used in a previous study (5). The questionnaire includes a section with the following five questions about specific motor and vocal tics in the past year: "Q1: Has your child had any repeated movements of parts of the face and head (e.g., eye blinking, grimacing, sticking tongue out, licking lips, spitting)?"; "Q2: Has your child had repeated movements of the neck, shoulder or trunk (e.g., twisting around, shoulder shrugging, bending over, nodding)?"; "Q3: Has your child had repeated movements of the arms, hands, legs, or feet?"; "Q4: Has your child had repeated noises and sounds (e.g., coughing, clearing throat, grunting, gurgling, hissing)?"; and "Q5: "Has your child had repeated words or phrases?" Each question is answered as either "definitely", "probably" or "not at all" present. Furthermore, we asked the following question about the frequency of these repetitive behaviors: "Q6: About how often does/did this happen in the last year?" This question was answered on the following 5-point Likert scale: "1: less than once a month, 2: 1-3 times a month, 3: about once a week, 4: more than once a week, 5: every day." We evaluated the presence of tics using binary variables consisting of either "with tics" or "without tics." We de ned the participants who responded "definitely" or "probably" to any of Q1, Q2, and Q4 as having tics. The participants who only endorsed repeated movements of the arms, hands, legs or feet (Q3) or repeated words or phrases (Q5) in the absence of a positive response to the other questions about the types of tics (Q1, Q2, Q4) were excluded from all case definitions to remove nontic movements such as stereotypy or isolated echolalia. We de ned as tics all responses of "definitely" or "probably" to questions concerning motor and/or vocal tics regardless of their frequency because there is no condition of frequency in the diagnostic criteria of tic disorders (1) and because we aimed to exhaustively nd tics in the general population.
In the post hoc analysis, we used a narrower de nition of tics that required answers of "definitely" or "probably" to questions concerning motor and/or vocal tics and frequencies of "daily" or "more than once a week." This narrower de nition of tics referred to the "broad de nition of TS and CT" used in a previous population-based study investigating the prevalence of TS/CT (5). The "broad de nition of TS and CT" used in this previous study differed from the narrower de nition of tics used in this study in that the previous study was carried out on participants aged 13 and excluded subjects with an intellectual disability or autism.

Maternal depressive/anxious symptoms
We employed the Kessler Psychological Distress Scale (K6) (27)(28)(29) for T1 and the General Health Questionnaire-28 (GHQ-28) (30, 31) for T2. The K6 and the GHQ-28 are both widely used self-report questionnaires that were developed to evaluate depressive/anxious symptoms. We used different scales between T1 and T2 in the current study because the TTC study also switched the scale used for maternal depressive/anxious symptoms from the K6 to the GHQ-28 starting at T2. The K6 is a short questionnaire consisting of 6 questions about the subjective mental distress of the respondent over the past 30 days that are answered on a 5-point scale, and the scores of the 6 items are added together (0-24 points). The GHQ-28 consists of 28 questions about the respondent's subjective physical and mental states over the past few weeks, with a total score being calculated for each item by giving 0 points each for the right two responses and 1 point each for the left two responses (0-28 points). Cutoff values are often used to screen for anxiety disorders and depression when assessing the K6 and the GHQ-28. However, in this study, we used raw values of the K6 and the GHQ-28 as continuous scales instead of screening scales, for the purpose of evaluating the severity of depressive/anxious symptoms, including the normal range in the general population. The Cronbach's alpha value was .84 for the K6 and .88 for the GHQ-28. We found that the distributions of the K6 and the GHQ-28 were similar based on the graphing cumulative distribution of their Z scores (Figure 1). If a primary parent other than a mother answered the K6 or the GHQ-28, we regarded those responses as missing values.

Other variables
Sex (5,7,32), age (33)(34)(35), maternal age (23,(36)(37)(38), socioeconomic status (39), and maternal alcohol use during pregnancy (40) were included in the analyses since previous studies have reported that these factors in uence the occurrence of TS/CT. The data for these variables were obtained from the responses to the questionnaires completed by caregivers. To assess socioeconomic status, family income was evaluated on a 10-point scale, which ranged from "0-990,000 yen" to "more than 10,000,000 yen." Information on maternal alcohol use during pregnancy was obtained from maternity record books that were provided for almost all mothers by local public organizations in Japan.

Statistical analysis
Longitudinal relationships between maternal depressive/anxious symptoms and children's tics were studied with structural equation modeling. We used SPSS® (Statistical Package for Social Science; IBM Corp., Armonk, N.Y. USA) version 21.0 for the characteristics of the study participants and Amos ver. 22.0 (IBM Corp, New York) for the structural equation modeling. We used the following three cross-lagged design models. The rst model analyzed the longitudinal relationships between maternal depressive/anxious symptoms and children's tics without adjusting for covariates (unadjusted model). The second model adjusted for sex, age in months, family income, maternal age, and maternal alcohol use during pregnancy (adjusted model). Finally, the third model took into account the nonnormal distribution of particular dependent variables (maternal depressive/anxious symptoms and children's tics at T2), logarithmically converted these measures and then conducted an analysis (logarithmically transformed model).
We also conducted a post hoc analysis by using the narrower de nition of tics, which referred to the "broad de nition of TS and CT" used in a previous study about the prevalence of TS/CT (5). The aims of the post hoc analysis were to compare the prevalence of tics in this study with that from the previous study and to examine the stability of the results on the relationships between maternal depressive/anxious symptoms and children's tics.
Missing values in the categories of tics, maternal depressive/anxious symptoms, and the covariates were accounted for by full information maximum likelihood procedures available in Amos. This method estimates model parameters and standard errors using all available data while adjusting for the uncertainty associated with missing data (41).
The threshold for statistical signi cance was set to p < .05 (two-sided) for all analyses. We evaluated the fit of our models by using the comparative fit index (CFI) and the root mean square error of approximation (RMSEA). A good model fit was indicated by an RMSEA value smaller than .05 and a CFI value larger than .95 (42,43).

Results
Characteristics of the study participants Table 1 shows the demographic characteristics of the 3,171 study participants. Of the 3,171 included children, 2,601 children (82.0%) had complete data about tics across both time points; 67 (2.1%) and 484 (15.3%) children had missing data about tics in either T1 or T2, respectively, and 19 (0.6%) children had missing data about tics in both T1 and T2. Across both time points, data about maternal depressive/anxious symptoms were complete for 2,683 mothers (84.6%); 24 (0.8%) and 309 (9.7%) mothers had missing scores in either T1 or T2, respectively; and 155 mothers (4.9%) had missing data about maternal depressive/anxious symptoms in both T1 and T2.
Of the participants, 23.9% children (744 of 3,109 available data) at T1 and 23.6% children (632 of 2,674 available data) at T2 had tics. These prevalence rates are consistent with previous studies that have estimated the point prevalence of tics in childhood to be approximately 20-29% (2-4).
Comparing the presence of tics between T1 and T2 using McNemar's test, we found no signi cant difference (p = .42). Of the 2,612 people whose data on the presence of tics were obtained in both T1 and T2, 343 participants endorsed tics at both T1 and T2, 290 participants endorsed tics only for T1, 270 participants endorsed tics for only T2, and 1,724 did not endorse tics at either time point.
Longitudinal relationships between maternal depressive/anxious symptoms and children's tics We investigated the relationships between maternal depressive/anxious symptoms and children's tics in a cross-lagged model analysis ( Figure 2, Table 2). There was a cross-sectional association between maternal depressive/anxious symptoms and children's tics at T1 and T2. Higher levels of maternal depressive/anxious symptoms at T1 signi cantly increased the presence of tics at T2 (adjusted model: β = .06, p = .001). In contrast, the presence of tics at T1 was related to higher levels of maternal depressive/anxious symptoms at T2 (adjusted model: β = .06, p = .001). All of these models indicated acceptable model t to the data (adjusted model: CFI = .951, RMSEA = .046). These results revealed that maternal depressive/anxious symptoms and children's tics had longitudinal, bidirectional relationships with each other.

Post hoc analysis: analysis with a narrower de nition of tics
We also analyzed the cross-lagged design models with the narrower de nition of tics which required the frequency of tics to be either "daily" or "more than once a week". This narrower de nition referred to the "broad de nition of TS and CT" used in a previous population-based study that investigated the prevalence of TS/CT (5). In this previous study, the prevalence rate of the "broad de nition of TS and CT" was reported to be 11.8% at the age of 13 (5). Using the narrower de nition of tics, 14.0% of the participants at the age of 10 (431 out of 3,085 effective data) and 12.1% of the participants at the age of 12 (323 out of 2,668 effective data) had tics among participants in this study. The presence of tics at T2 was signi cantly less than that in T1 (p = .003). Furthermore, we found longitudinal, bidirectional relationships between maternal depressive/anxious symptoms and children's tics (Table 3).

Discussion
This was the rst study to examine the longitudinal relationships between maternal depressive/anxious symptoms and children's tics in a population-based early adolescent sample. The following two ndings were obtained. First, when more severe maternal depressive/anxious symptoms are present, children are more likely to present tics two years later. Second, the severity of maternal depressive/anxious symptoms and the presence of children's tics are longitudinally associated with each other.
Several explanations may be possible for the signi cant magnitude-response relationship between maternal depressive/anxious symptoms and children's later tics. First, maternal depressive/anxious symptoms may affect the occurrence, persistence, and exacerbation of children's tics as an environmental factor because environmental factors such as psychosocial stresses are known to exacerbate tics (44)(45)(46)(47)(48). There are no reports of an association between maternal psychiatric symptoms and the course of children's tics. However, some studies have shown an association between maternal psychiatric symptoms and the course of children's psychiatric symptoms. For example, maternal depression has been shown to be associated with increased psychiatric diagnoses and emotional and behavioral problems in children, and when maternal depression is remitted, the children's problems are more likely to also be in remission (49). Furthermore, improvement in parental depression has a positive impact on the health, emotional, cognitive, academic and overall functioning of children (50). Thus, similar to these reports, maternal depressive/anxious symptoms might affect the course of children's tics as an environmental factor. Second, there might be genetic relationships between maternal depressive/anxious symptoms and the occurrence, persistence, and exacerbation of tics. Family studies of TS have suggested that TS has genetic correlations with depressive disorders and anxiety disorders (51,52), although these correlations are possibly mediated through ADHD and OCD (52). There has been no research examining the genetic relationships between maternal depressive/anxious symptoms and the persistence or exacerbation of tics, but the results of this research could not rule out these possibilities. Furthermore, there is also a possibility that an interaction of genetic and environmental factors is involved in the relationships between maternal depressive/anxious symptoms and children's tics. The results of this study could not distinguish genetic and environmental contributors to the relationships between maternal depressive/anxious symptoms and children's tics.
Our ndings on the association between the presence of tics and increased maternal depressive/anxious symptoms two years later can be explained by several potential mechanisms. One of the possible mechanisms is that the parenting stress associated with bringing up children with tics might in uence maternal depressive/anxious symptoms. Parents of children with TS experience increased levels of caregiver burdens and parenting stress compared to parents of children without TS (53,54). The visible nature of tics can have an impact on the parent-child relationship, with parents becoming overprotective of, worrying about, struggling to accept or trying to control children's tics, which can lead to family con icts, poor parent-child relationships and increased frustrations in parenting (7,9,10,55,56). Parenting stress in parents of children with tics could also occur due to children's comorbidities, such as ADHD, OCD, and behavioral problems (53,54). In a previous population-based study, 21.2% of children had tics, and children with tics were more affected by psychopathologies, including ADHD and OCD, than were children without tics (12). In addition to these environmental factors, both genetic factors and genetic/environmental interactions might have an effect on children's tics on maternal depressive/anxious symptoms.
In a post hoc analysis, the prevalence rates of tics found by using the narrower de nition of tics were 14.0% at the age of 10 and 12.1% at the age of 12. This narrower de nition of tics referred to the "broad de nition of TS and CT" used in a previous study, and the prevalence rate of "broad de nition of TS and CT" has been reported to be 11.8% at the age of 13 (5). The prevalence rates of tics in this study were concordant with those found in the previous study, considering that the previous study excluded subjects with an intellectual disability or autism (5) and that the prevalence of tics decreases slowly from early adolescence (7,13,33,(57)(58)(59)(60). Furthermore, when using the narrower de nition of tics, we also found longitudinal bidirectional relationships between maternal depressive/anxious symptoms and children's tics. Thus, we con rmed the stability of the results of the main analysis.
The implications of this study were that the longitudinal bidirectional relationships between maternal depressive/anxious symptoms and tics in early adolescence may suggest a vicious cycle in which maternal depressive/anxious symptoms make tics worse, and tics make maternal depressive/anxious symptoms worse. This study also suggested that not only intervention in children's tics but also intervention in maternal depressive/anxious symptoms might be important for the treatment of tics. However, the present study was unable to separate genetic and environmental factors in the association between tics and maternal depressive/anxious symptoms; therefore, further research is needed to determine the effect of intervention on maternal anxiety/depressive symptoms. While there has been a consensus on the importance of family psychoeducation in the treatment of tics (7, 60), it is not known whether maternal psychiatric problems in uence the course of children's tics. This study provides new insights for future research and practice.
The strength of this study was that, for the rst time, it was shown that higher levels of maternal depressive/anxious symptoms are related to an increased likelihood of tics two years later and that there are longitudinal relationships between maternal depressive/anxious symptoms and children's tics. Other strengths were the large sample size, the high follow-up rate of the study, and the inclusion of nonclinical tics. In contrast, this study also had several limitations. First, we used different measures of maternal depressive/anxious symptoms for T1 and T2. We found that the distributions of the K6 and the GHQ-28 were similar by graphing the cumulative distribution of the Z scores of the K6 and the GHQ-28 ( Figure 1). Additionally, this limitation did not in uence the course from maternal depressive/anxious symptoms to children's tics. Second, in this study, maternal depressive/anxious symptoms and children's tics were evaluated not by direct clinical assessments but by questionnaires to caregivers. However, we deliberately chose rigorous tic definitions and sought to exclude participants with nontic movement disorders (e.g., stereotypies associated with autism or an intellectual disability, repetitive arm/leg movements that could be better explained by tremor or motor restlessness) (5). In this study, the prevalence of tics was 23.9% at age 10 and 23.6% at age 12. These prevalence rates could be considered reasonable based on the following evidence. Point prevalence depends strongly on age; the highest rate is estimated to be approximately 20% at age 5-10, and the lifetime prevalence is much higher (3). In previous studies that have directly observed children, tics were found in 29.2% of fourth-grade children in an elementary school in Washington D.C. (2) and in 21.2% of children aged 9-17 years old (mean 13.1 years old) in Monroe County, Rochester, New York (4). The prevalence rates of tics in the present study were consistent with those found in these previous studies. Third, the data analysis in this study could not adjust for ADHD and OCD, which are frequently comorbid with tics. That may be because of the strong association of tics with ADHD and OCD. Future studies are needed to examine the effects of ADHD and OCD on the bidirectional relationships between maternal depressive/anxious symptoms and children's tics. The fourth limitation was that the research interval in this longitudinal study was relatively short. Typically, tics improve gradually during adolescence, with repeated periods of remission and exacerbation. Thus, it might be di cult to capture change in the short research period of two years. Longer-term follow-up periods are needed in the future. The fth limitation was that we did not collect information about the maternal history of tics. Given the low rate of medical consultation for tics (5,12,51) and the clinical outcome that tics often improve or disappear after adolescence (7,13,14), it is probably not possible to obtain accurate information on the maternal history of tics. Finally, there were also some limitations inherent to the cross-lagged model (61); i.e., there is a possibility that there are multiple potential additional factors (not included in the model) that in uence the bidirectional relationship over time.
The following two studies would be helpful in testing the viability of the relationships between maternal depressive/anxious symptoms and children's tics and in advancing research and practice. First, if the course of children's tics could be observed at three or more time points, it would be possible to con rm the vicious cycle that develops between maternal depressive/anxious symptoms and children's tics and to investigate the mediating factors. For example, children's anxiety/depression symptoms might mediate the relationships between maternal depressive/anxious symptoms and children's tics. In addition, poor quality of life or comorbidities might mediate the relationships between children's tics and maternal depressive/anxious symptoms. Second, intervention studies could examine whether improvements in maternal depressive/anxious symptoms improve children's tics.
This study was the rst to show the relationship between preceding maternal depressive/anxious symptoms and an increased risk of children's tics two years later in early adolescence. Furthermore, we found longitudinal bidirectional relationships between maternal depressive/anxious symptoms and children's tics. Although we could not separate environmental factors and genetic factors in this research, the ndings implied that it may be important to care not only for children with tics but also for their mothers' depressive/anxious symptoms when tics are treated in early adolescence.

Conclusion
Previous studies have revealed the association between maternal depressive/anxious symptoms and children's tics. However, few studies have examined this association longitudinally; thus, the effect of maternal depressive/anxious symptoms on the course of tics has remained unknown. We examined the longitudinal relationships between maternal depressive/anxious symptoms and children's tics across early adolescence in a population-based sample. Our results demonstrated that with more severe maternal depressive/anxious symptoms, children are more likely to present tics two years later. This study also implied a longitudinal bidirectional relationship between maternal depressive/anxious symptoms and children's tics during early adolescence, which may exacerbate each other over time and possibly create a vicious cycle. Although we could not separate environmental factors and genetic factors in this research, we can speculate that it may be important to care not only for children with tics but also for their mothers' depressive/anxious symptoms when tics are treated in early adolescence.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests. Authors' contributions TY conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript, being supervised by SA, AN, SY, and YK designed the study and reviewed and revised the manuscript. KK critically reviewed the manuscript for important intellectual content. SU supervised the statistical analysis. All authors contributed to and have approved the final manuscript. Abbreviations: SD, standard deviation.
cNemar's test was used to compare the presence of tics between T1 and T2. amily income was evaluated on the 10-point scale described in the Method section and categorized into the four groups in this table. e obtained total 3,109 data for the presence of tics at T1 in the main analysis and total 3,085 data for that in the post hoc analysis. This screpancy in the total numbers arose because, as described in the methods, the presence of tics in the main analysis did not require the datum bout the frequency of tics while the presence of tics in the post hoc analysis required the datum about the frequency of tics. ** p < .01     Figure 1 Cumulative distribution of the Z scores of the K6 and the GHQ-28 Cross-lagged models of relationships between maternal depressive/anxious symptoms and children's tics. Abbreviations: e, error variable. T1 = 10 years of age; T2 = 12 years of age. Cross-lagged model of the relationships between maternal depressive/anxious symptoms and children's tics in the populationbased adolescent sample. This gure shows the results of the adjusted model in Table 2. Paths from covariates are omitted from the gure. **p < .01, ***p < .001