Associations between media exposure and mental health among children and parents after the Great East Japan Earthquake

ABSTRACT

Background: Exposure to natural disaster media coverage is associated with mental health problems, but its long-term impacts are still unclear. Also, no study has analysed the psychological impact of exposure to natural disaster media coverage among children who are generally sensitive to threatening events.

Objective: We aimed to examine how television images of victims after the 2011 Great East Japan Earthquake were associated with mental health among children and their parents.

Methods: In 2012, questionnaires for sociodemographic factors were distributed to 2053 families. Parents who provided written consent were contacted in 2013 and invited to provide information on mental health problems (outcome) and retrospectively provide information on television watching at the time of the earthquake (exposure). We used data from 159 parents who completed the survey as the final sample. We used a dichotomous variable to evaluate exposure to media coverage. Multivariable regression was used to examine the association between exposure to television images of victims and mental health, adjusting for potential confounders. Bias-corrected and accelerated bootstrap confidence intervals (CIs) were used.

Results: Exposure to television images of victims was significantly associated with worse psychopathology among children (β, 1.51; 95% CI, 0.07–2.96) and greater psychological distress among their parents (β, 1.49; 95% CI, 0.28–2.70). Child psychopathology and parental psychological distress were significantly correlated (r = 0.36, p < .001).

Conclusions: Exposure to television images of disaster victims may produce long-term impacts on mental health among children and their parents. To reduce the likelihood of mental health problems associated with disasters, clinicians may recommend reducing exposure to television images of victims.

HIGHLIGHTS

• Television images of victims after the 2011 Great East Japan Earthquake were associated with worse psychopathology among children in 2013, adjusted for potential confounders.

• Similar associations were found in psychological distress among their parents.

• Child psychopathology and parental psychological distress were correlated with each other.

Antecedentes: La exposición a la cobertura en los medios de los desastres naturales se asocia a problemas de salud mental, pero sus impactos en el largo plazo todavía no están claros. Además, ningún estudio ha analizado el impacto psicológico de la exposición a la cobertura en los medios de desastres naturales en los niños, quienes son generalmente sensibles a eventos amenazantes.

Objetivo: Buscamos examinar cómo las imágenes en televisión de las víctimas luego del Gran Terremoto del este de Japón en 2011 se asociaron a la salud mental en los niños y sus padres.

Métodos: En 2012, se distribuyeron cuestionarios de factores sociodemográficos a 2.053 familias. Los padres que proveyeron consentimiento escrito fueron contactados en 2013 e invitados a proveer información de problemas de salud mental (resultado) y retrospectivamente proveer información sobre la observación de televisión al momento del terremoto (exposición). Utilizamos los datos de 159 padres que completaron la encuesta como la muestra final. Usamos una variable dicotómica para evaluar la exposición a la cobertura mediática. Se utilizó regresión multivariable para examinar la asociación entre la exposición a imágenes de víctimas en televisión y salud mental, ajustando por potenciales confundentes. Se usaron intervalos de confianza de arranque corregidos por sesgos y acelerados.

Resultados: La exposición a imágenes de víctimas en televisión se asoció significativamente con peor psicopatología entre los niños (β,  1.51; IC del 95% 0.07 a 2.96) y mayor sufrimiento psicológico entre sus padres (β, 1.49; IC del 95% 0.28 a 2.70). La psicopatología de los niños y el sufrimiento psicológico de los padres estaban correlacionados de forma significativa (r = 0.36, p < .001).

Conclusiones: La exposición a imágenes de víctimas de desastres en televisión puede producir impactos a largo plazo en la salud mental de los niños y sus padres. Para reducir la probabilidad de problemas de salud mental asociados con desastres, los clínicos pueden recomendar disminuir la exposición a imágenes de víctimas en televisión.

背景：暴露于自然灾害媒体报道与心理健康问题有关，但其长期影响仍不清楚。 此外，还没有研究分析过自然灾害媒体报道对于总体对威胁事件敏感的儿童的心理影响。

目的：我们旨在研究 2011 年东日本大地震后受害者电视图像如何与儿童及其父母的心理健康相关联。

方法：2012年向2053户家庭发放社会人口因素问卷。 2013年联系了提供书面同意的父母，并邀请他们提供有关心理健康问题（结果）的信息，并回顾性地提供地震发生时看电视的信息（暴露）。 我们使用完成调查的 159 位家长的数据作为最终样本。 我们使用二分变量来评估媒体报道的暴露率。使用多变量回归考查受害者电视图像暴露与心理健康之间的关联，控制了潜在的混杂因素。 使用偏差校正和加速自举置信区间 (CI)。

结果：接触受害人电视图像与儿童较差的心理病理（β，1.51；95% CI，0.07 至 2.96）和父母较大的心理困扰（β，1.49；95% CI，0.28 至 2.70）显著相关。 儿童精神病和父母心理困扰显著相关（r = 0.36，p < .001）。

结论：接触灾难受害者电视图像可能对儿童及其父母的心理健康产生长期影响。 为了减少与灾难相关的心理健康问题的可能性，临床医生可能会建议减少接触受害者电视图像。

KEYWORDS:

• Television exposure
• natural disasters
• psychopathology
• psychological distress
• Japan

PALABRAS CLAVE:

• Exposición a Televisión
• Desastres naturales
• Psicopatología
• Sufrimiento psicológico
• Japón

关键词:

• 电视暴露
• 自然灾害
• 心理病理学
• 心理困扰
• 日本

1. Introduction

Disasters cause not only physical harm but also mental health problems (Newnham et al., 2022; Norris et al., 2002). Notably, an elevated risk of posttraumatic stress disorder (PTSD) and depressive disorder were reported across natural, technological, and human-made disasters (Bonde et al., 2016; Neria et al., 2008). The risk of subclinical symptoms related to mental disorders is also elevated in disasters (Beaglehole et al., 2018; Fergusson et al., 2014). The risk factors may include traumatic exposure to frightening and life-threatening conditions (Perkonigg et al., 2000) and consequential adverse life events such as family death or loss of employment (Avison & Gotlib, 1994). The role of indirect exposure to disaster through media has been considered (Goodwin et al., 2013; Lau et al., 2006; McLeish & Del Ben, 2008; Weems et al., 2012), but its long-term impact is unclear.

Exposure to disaster media coverage is considered as deleterious as direct trauma exposure (Lubens & Holman, 2017), depending upon the type of disaster and individuals’ attributes (Pfefferbaum et al., 2019). This issue has been most frequently studied in terrorist incidents (Ahern et al., 2002, 2004; Busso et al., 2014; Pfefferbaum et al., 2001, 2019; Stickley et al., 2020). Television images of the September 11 Terrorist Attacks are suggested to be linked to PTSD (Ahern et al., 2002, 2004), although non-work-related television exposure is no longer included in the current diagnostic criteria of PTSD in the DSM-5 (American Psychiatric Association, 2013). The issue was also studied in natural disasters. For example, positive associations between posttraumatic stress and exposure to media coverage were found in studies after Hurricane Katrina (McLeish & Del Ben, 2008), Hurricane Gustav (Weems et al., 2012), Hurricane Sandy (Goodwin et al., 2013), the 2004 tsunami in Hong Kong (Lau et al., 2006). These observations suggest the negative impacts of indirect exposure to natural disasters on mental health.

Similar relationships have been reported after the 2011 Great East Japan Earthquake (Bui et al., 2012; Nishi et al., 2012; Rodgers et al., 2012). However, the previous studies have limitations. Firstly, their follow-up periods are relatively short, namely, four months at the longest (Nishi et al., 2012). No long-term follow-up data has been available despite accumulated concerns about mental health problems lasting over the years (Ando et al., 2017; Fujiwara et al., 2014; Kino et al., 2020). Another limitation is that most previous studies focused on adult participants except for one enrolling urban-school youths (Weems et al., 2012). As children's media use is reportedly linked to threat perception (Comer et al., 2008), it is crucial to understand how exposure to natural disaster media coverage affects mental health in this population. Here, we analysed data from 159 parents in the two years following the 2011 Great East Japan Earthquake to address these issues. We hypothesised that watching specific television images would be associated with mental health problems.

2. Methods

2.1. Participants

In this cross-sectional study, we used data from the Tama Children’s Survey (TCS) targeting a community-based sample of the Tokyo metropolitan area. Details of the TCS are shown in our past reports (Haraguchi et al., 2019; Saito et al., 2017). The Tokyo metropolitan area is located approximately 400 km from the epicentre of the Great East Japan Earthquake. In March 2012, questionnaires for sociodemographic factors and child negative affectivity were distributed to 2053 families. One thousand four hundred six of them provided written consent to participate. Four hundred twenty-six parents with a four- or five-year-old child at the time of the earthquake were contacted and invited to participate in a mail survey from February to March 2013. Parents provided information about themselves and their children’s mental health problems in 2013 (outcome). Also, they retrospectively provided information on television watching after the earthquake (exposure) and potential confounders. The survey was approved by the National Center of Neurology and Psychiatry Institutional Review Board (A2012-056). The response rate was 42.5%, i.e. 181 parents completed the survey. Note that we used data for only those who completed the survey. All child-parent pairs had television exposure after the earthquake. We excluded 22 pairs who were out of the Tokyo metropolitan area on 11 March 2011. Thus, data from 159 parents were analysed.

2.2. Measures

2.2.1. Outcome: mental health problems in 2013

The Japanese version of the Strengths and Difficulties Questionnaire (SDQ) (Matsuishi et al., 2008; Moriwaki & Kamio, 2014) was completed by parents from 6 February to 11 March 2013. The SDQ is a screening questionnaire for children aged from four to 16 years comprising 25 items. Its subscales cover emotional symptoms, conduct problems, hyperactivity/inattention, peer relationship problems, and prosocial behaviour. Each item was rated on three-point response options (0 = ‘Not True,’ 1 = ‘Somewhat True,’ or 2 = ‘Certainly True’), yielding a maximum total score of 10 on each subscale. Some items were reverse-scored so that all items were interpreted in the same direction. The scores on the first four subscales were summed to provide a total difficulties score ranging from 0 to 40. A greater total score suggests worse psychopathology.

The Japanese version of the Six-item Kessler Psychological Distress Scale (K6) (Furukawa et al., 2008) was used to assess parental distress. Each item was self-reported on five-point response options ranging from 0 (no distress) to 4 (maximum stress), with possible total scores ranging from 0 to 24. Our data showed good internal consistency (Cronbach’s α = 0.86). A greater score suggests greater psychological distress.

2.2.2. Exposure to television images of victims after the earthquake

Exposure to television images after the earthquake was evaluated with a series of questions for several specific images broadcast repeatedly. Parents were asked whether they saw these images on television with their children after the earthquake. Child-parent pairs were considered to watch the same television images since it was considered unlikely for children to watch television alone at the time of the disaster. The images included ‘people being rescued,’ ‘people on the roof escaping from the tsunami,’ ‘people running from debris,’ and ‘people who have difficulty returning home.’ Watching at least one of these images constituted the television images variable. The image of ‘people being rescued’ was included because it reportedly has detrimental effects in the aftermath of a natural disaster (McLeish & Del Ben, 2008). In the present study, we focused on television images of victims rather than non-human images because 98.7% of the participants (157 child-parent pairs) watched at least one non-human image, such as the earthquake or tsunami, which should not be included in the analyses. We created a dichotomous variable for exposure to television images of victims.

2.2.3. Covariates

Covariates before watching television images were adjusted for. For children, age, sex (male or female), parental employment status (full-time, part-time, or no formal job) (Paavonen et al., 2006; Suzuki et al., 2014), evacuation (yes or no) (Kunii et al., 2016; Sawa et al., 2013; Yabe et al., 2014), date of first television exposure, and preexisting negative affectivity were included. We evaluated whether participants evacuated because the earthquake caused fatalities and property damage even in Tokyo, and a sizable number of people evacuated. The date of the first television exposure was included because television images changed from day to day at the time of the earthquake. Preexisting negative affectivity evaluated by the Early Childhood Behavior Questionnaire (ECBQ) (Putnam et al., 2006) was included since preexisting conditions may have an essential role in the association between mental health and television exposure (Weems et al., 2012). We used an Interim version of the Very Short Form ECBQ that included a 12-item Negative Affectivity dimension (Kamio et al., 2018). Each item was rated on seven-point response options ranging from 1 (never) to 7 (always). Some items were reverse-scored so that all items were interpreted in the same direction. Our data showed acceptable internal consistency (Cronbach's α = 0.68). An average score of 12 items was used in the analyses. A greater total score suggests worse negative affectivity.

For parents, sex (male or female), employment status (full-time, part-time, or no formal job) (Paavonen et al., 2006; Suzuki et al., 2014), evacuation (yes or no) (Kunii et al., 2016; Sawa et al., 2013; Yabe et al., 2014), and date of first television exposure were included. Parents self-reported if they had been diagnosed with any mental disorder in the 12 months preceding the earthquake to account for preexisting conditions. However, no parent reported preexisting mental disorders, and this condition was not included in the analyses. We used information for potential confounders before watching television images to ensure that we did not improperly adjust for mediators (Schisterman et al., 2009). The complete study design is displayed in Figure 1.

Figure 1. Directed acyclic graph of the hypothesised associations between exposure, outcome, and potential confounders.

2.3. Statistical analysis

Statistical analysis was conducted by using R version 4.1.0 and STATA 15. Multivariable regression was used to examine the associations between mental health problems (SDQ scores among children and K6 scores among parents) and exposure to television images of victims, adjusting for potential confounders. Both unadjusted and adjusted analyses were conducted. Bias-corrected and accelerated bootstrap confidence intervals (CIs) were used to obtain accurate information (Cumming, 2014; Felsenstein, 1985; Lumley et al., 2002). In executing bootstrap CIs, the size of the bootstrap sample was set at 1000 with 95% CIs (Carpenter & Bithell, 2000). To evaluate the relationship between child SDQ scores and parental K6 scores, we used Spearman's rank correlation coefficient because Shapiro-Wilk tests revealed that these data were not normally distributed. The strength of the correlation coefficient was interpreted based on Dancey and Reidy's definition (weak = 0.1–0.3; moderate = 0.4–0.6; strong = 0.7–0.9; perfect = 1) (Akoglu, 2018; Dancey & Reidy, 2004).

3. Results

3.1. Descriptive statistics

Table 1 provides the characteristics of the study participants in 2011. Approximately half of the child-parent pairs (n = 79, 49.7%) watched television images of victims after the earthquake. The number of child-parent pairs who watched each television image was as follows: ‘people being rescued,’ 37 (23.3%); ‘people on the roof escaping from the tsunami,’ 55 (34.6%); ‘people running from debris,’ 25 (15.7%); ‘people who have difficulty returning home,’ 45 (28.3%). Overall, these characteristics were balanced in both children and parents. Most participants experienced their first television exposure on 11 March 2011, i.e. the day of the earthquake. Participants who did not watch television images of victims but reported the date of first television exposure are considered to be watching non-human images such as the tsunami.

Table 1. Demographics of the study population in 2011.

Characteristic	Overall (n = 159)	Watching television images of victims
		Yes (n = 79)	No (n = 80)
Children
Age, mean (SD), y	4.35 (0.30)	4.33 (0.31)	4.37 (0.29)
Sex, no. (%)
Female	73 (45.9)	37 (46.8)	36 (45.0)
Male	86 (54.1)	42 (53.2)	44 (55.0)
Parental employment status, no. (%)
Full-time	40 (25.2)	22 (27.8)	18 (22.5)
Part-time	21 (13.2)	10 (12.7)	11 (13.8)
No formal job	98 (61.6)	47 (59.4)	51 (63.8)
Evacuation, no. (%)
No	148 (93.1)	72 (91.1)	76 (95.0)
Yes	11 (6.9)	7 (8.9)	4 (5.0)
Date of first television exposure^a, no. (%)
11 March 2011	136 (86.6)	63 (81.8)	73 (91.3)
12 or 13 March 2011	21 (13.4)	14 (15.6)	7 (8.8)
Preexisting negative affectivity, mean (SD)	3.13 (0.75)	3.26 (0.82)	3.00 (0.65)
Parents
Sex, no. (%)
Female	154 (96.9)	77 (97.5)	77 (96.3)
Male	5(3.1)	2 (2.5)	3 (3.8)

Abbreviation: SD, standard deviation.

^a Missing data for two child-parent pairs.

3.2. Associations between media exposure and mental health

Table 2 summarises unadjusted and adjusted associations between media exposure and SDQ total difficulty scores among children. In the unadjusted model, children who watched television images of victims presented higher SDQ scores than those who did not, suggesting worse psychopathology (β, 1.64; 95% CI, 0.17–3.11). This association was slightly attenuated but remained significant in the adjusted model, including age, sex, parental employment status, evacuation, date of first television exposure, and preexisting negative affectivity (β, 1.51; 95% CI, 0.07–2.96).

Table 2. Unadjusted and adjusted associations between television images of victims in 2011 and child SDQ total difficulty scores.

	Child SDQ total difficulty scores
Variable	β (Unadjusted) [95% CI]	β (Adjusted) [95% CI]	Z value (Adjusted)	P value (Adjusted)
Television images of victims	1.64 [0.17 to 3.11]	1.51 [0.07–2.96]	2.06	0.04
Age	NA	−2.46 [−4.68 to −0.23]	−2.17	0.03
Sex
Male	NA	Reference
Female	NA	−0.09 [−1.54 to 1.37]	−0.12	0.91
Parental employment status
No formal job	NA	Reference
Part-time	NA	−0.34 [−2.80 to 2.13]	−0.27	0.79
Full-time	NA	0.78 [−1.02 to 2.58]	0.85	0.39
Evacuation
No	NA	Reference
Yes	NA	0.36 [−2.49 to 3.21]	0.25	0.81
Date of first television exposure
11 March 2011	NA	Reference
12 or 13 March 2011	NA	−2.92 [−5.03 to −0.82]	−2.72	< 0.01
Preexisting negative affectivity	NA	1.20 [−0.06 to 2.46]	1.86	0.06

Abbreviations: SDQ, Strength and Difficulty Questionnaire; Scale; NA, not applicable.

Table 3 summarises unadjusted and adjusted associations between media exposure and K6 scores among parents. In the unadjusted model, parents who watched television images of victims showed higher K6 scores than those who did not, suggesting greater psychological distress (β, 1.28; 95% CI, 0.10–2.45). A numerically larger coefficient was found in the adjusted model, including sex, employment status, evacuation, and date of first television exposure (β, 1.49; 95% CI, 0.28–2.70).

Table 3. Unadjusted and adjusted associations between television images of victims in 2011 and parental K6 scores.

	Parental K6 scores
Variable	β (Unadjusted) [95% CI]	β (Adjusted) [95% CI]	Z value (Adjusted)	P value (Adjusted)
Television images of victims	1.28 [0.10 to 2.45]	1.49 [0.28–2.70]	2.41	0.02
Sex
Male	NA	Reference
Female	NA	0.70 [−0.86 to 2.26]	0.88	0.38
Employment status
No formal job	NA	Reference
Part-time	NA	1.78 [−0.10 to 3.66]	1.85	0.06
Full-time	NA	−0.62 [−1.81 to 0.57]	−1.02	0.31
Evacuation
No	NA	Reference
Yes	NA	0.06 [−2.20 to 2.31]	0.05	0.96
Date of first television exposure
11 March 2011	NA	Reference
12 or 13 March 2011	NA	−1.51 [−2.60 to −0.42]	−2.71	< 0.01

Abbreviations: K6, Six-item Kessler Psychological Distress Scale; NA, not applicable.

3.3. Correlation between child psychopathology and parental psychological distress

Finally, we calculated Spearman’s rank correlation coefficient to examine how SDQ and K6 scores correlated. The analysis showed that these outcomes were significantly correlated (r = 0.36, p < .001). The strength of correlation coefficient was weak.

4. Discussion

This study provided new information on the link between exposure to natural disaster media coverage and mental health problems. As hypothesised, we found that watching television images of victims after the earthquake was associated with mental health problems among children and their parents. These associations remained significant after adjusting for potential confounders. These results are consistent with previous reports on the short-term psychological impacts of exposure to natural disaster media coverage (Bui et al., 2012; Goodwin et al., 2013; Lau et al., 2006; McLeish & Del Ben, 2008; Nishi et al., 2012; Rodgers et al., 2012; Weems et al., 2012). The present study is the first to suggest the long-term impact of exposure to natural disaster media coverage on mental health, which is an important public health topic (Ando et al., 2017; Fujiwara et al., 2014).

Television exposure may have detrimental impacts comparable to direct trauma exposure (Lubens & Holman, 2017). So far, it is controversial whether mental health problems associated with the Great East Japan Earthquake diminish (Goodwin et al., 2020) or persist (Ando et al., 2017; Fujiwara et al., 2014; Kino et al., 2020) in a long-term period. The mental health problems observed two years after the television exposure may have been more pronounced at earlier time points. To clarify this issue, assessments of the time course of mental conditions would provide relevant information.

Exposure to natural disaster media coverage was associated with children and their parents’ mental health problems. Similar findings have been reported among urban-school youths (Weems et al., 2012). However, no research has studied preschool children who are generally sensitive to threatening events (Comer et al., 2008). These findings are pertinent to a post-disaster intervention targeting media literacy among parents of children (Comer et al., 2008). Promoting media literacy may nullify the link between exposure to natural disaster media coverage and threat perceptions (Comer & Kendall, 2007), preventing subsequent mental health problems.

In the present study, images of victims facing adverse events associated with the earthquake were implicated in producing persistent psychological impacts. A previous study showed that an image of ‘people falling or jumping from the towers of the WTC’ after the September 11 Terrorist Attacks was most consistently associated with PTSD and depression, compared with those of airplane crashes or building collapses (Ahern et al., 2002). In the present study, human images were also shown to have significant and persistent psychological impacts, although we did not compare them to the impacts of non-human images. The DSM-5 Criterion A4 for PTSD excludes non-work-related indirect exposure, including television watching (American Psychiatric Association, 2013), and the outcomes measured in the present study were not PTSD symptoms. Still, it should be noted that television images showing affected victims might disturb a vulnerable population, such as young children and parents who are sensitive to threatening events to their children, which may result in strong psychological impacts.

Psychopathology among children (measured by the SDQ) and psychological distress among their parents (measured by the K6) were significantly correlated in the present study. Children of parents with depression are reportedly more likely to experience mental health problems (Hanington et al., 2010–2; Olfson et al., 2003), which may result in increased utilisation of mental health services and related expenditures (Olfson et al., 2003). Conversely, psychopathology among children may lead to mental health problems among their parents (Barroso et al., 2018–4; Garber et al., 2011–1). Other studies demonstrated a bidirectional relationship between parental stress and child behaviour problems (Mackler et al., 2015–4, p. 10; Neece et al., 2012–1). The significant correlation between psychopathology among children and psychological distress among their parents may be congruent with these previous reports.

5. Limitations

The present study includes some points that we should cautiously consider. First, we had a relatively low response rate and a small sample. Since there may be cause to doubt the assumptions underlying parametric CIs, we used bootstrap CIs to obtain accurate information (Carpenter & Bithell, 2000). Still, external validity may not be high. Also, we could not examine and contrast each of the four television images of victims due to a small sample. Second, a categorical variable was used to evaluate whether child-parent pairs watched television images of victims. Continuous variables, such as frequency or hours, may have enabled dose-response assessments of the effect of exposure to television images. Third, all information was parent-reported retrospectively, which may cause recall bias and may not allow firm conclusions. For example, the date of the first television exposure might have been inaccurate. Parents who remember watching television images of victims might have reported more problems. Fourth, our data did not allow analyses adjusting for some potential confounders, e.g. parental age. No parent reported preexisting mental disorders, which social desirability biases might cause. It is also possible that parents with mental disorders were less likely to respond to our survey. We did not have information on parental psychological distress in 2011, which might be associated with the accuracy of reporting media exposure. Fifth, reverse causation is conceivable, i.e. vulnerable people are more likely to watch disaster-related news. Stronger prospective studies are warranted to confirm the causal effect of media exposure on mental health. Finally, the present findings do not suggest which covariates are important for mental health problems because we chose the covariates to adjust for confounding between exposure to television images of victims and mental health problems.

6. Conclusion

In conclusion, the present study provides novel information on the potential long-term impacts of exposure to natural disaster media coverage on mental health among children and parents. To reduce the likelihood of mental health problems associated with disasters, clinicians may recommend reducing watching television images of victims. Future prospective studies should examine a larger sample and the dose-response effects of such exposure.

Availability of data and materials

The National Center of Neurology and Psychiatry Institutional Review Board imposes strict rules on sharing the data as these are classified according to ethical restrictions due to privacy concerns. Anonymized data are available to researchers and institutions upon request.

Author contributions

AO and YK devised initial research questions. ZN conducted analyses. The manuscript was written by ZN and finalised by AO, HT, AS, TS and YK and with substantial text contributions from all authors.

Ethics approval and consent to participate

The survey was approved by the National Center of Neurology and Psychiatry Institutional Review Board (A2012–056).

Consent for publication

Written informed consent was obtained from all participants.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Health and Labour Sciences Research Grants for Comprehensive Research on Persons with Disabilities of AMED [grant numbers 20dk0307099 and 21he2202007], Intramural Research Grants for Neurological and Psychiatric Disorders of NCNP (3–1), JH Grants [grant number 2020-B-08], Research Grant for Nervous and Mental Disorders [grant number H20-KOKORO-004] and Health and Labor Science Research Grant for Research on Psychiatric and Neurological Diseases and Mental Health (H24-Seishin-Shitei-002(Fukkou) and H27-Seishin-Shitei-003) of the Ministry of Health, Labour and Welfare of Japan.