Introduction

Autism spectrum disorder (ASD) is characterised by two major features: social communication and social interaction deficits, and restricted and repetitive behaviours, interests, or activities [1]. Without experiencing significant dysfunction, some people who do not receive a formal diagnosis of ASD may exhibit a range of autistic traits, such as pragmatic difficulties, communication difficulties, rigidity, stereotyped behaviours and impaired emotional recognition [2, 3]. Although being considered subclinical, these undiagnosed individuals with autistic traits can still encounter similar challenges in their daily lives, particularly in social activities.

Autistic individuals often develop coping strategies, such as ‘camouflaging’, to navigate the challenges they encounter in their everyday social interactions [4]. Although there is no clear definition of camouflaging, the term usually refers to the subjective experience of autistic individuals employing specific behavioural and cognitive strategies, with an attempt to appear non-autistic and socially competent during social situations [5,6,7]. Camouflaging strategies may involve adopting a non-autistic persona [8, 9] or compensating for social challenges by emulating others' behavioural expressions [10], allowing them to blend in and reduce risk of isolation.

Camouflaging has been reported to be associated with negative effects on mental health. A growing body of evidence suggests that camouflaging is linked to various mental health issues, including generalised anxiety, depression, social anxiety, and even thoughts of suicide and self-harming behaviours [11,12,13]. Furthermore, not only autistic individuals but also non-autistic individuals have reported similar experiences of camouflaging [14, 15]. However, research on camouflaging behaviours in non-autistic individuals is limited.

Doujin (どうじん) is a Japanese term that originally referred to a group or circle of individuals who share the same interests, hobbies, or tastes [16]. Nowadays, despite encompassing a wide range of shared interests, such as movies and idols, doujin is closely and primarily associated with the anime, comics, and games (ACG) subculture. This subculture has become widespread and influential among young adults and teenagers worldwide, including Taiwan [17]. For example, Taiwan's most significant ACG doujin event, held eight times a year, attracts an average of 80,000 participants each time, according to Taiwan Creative Content Agency (TAICCA) [18]. With a substantial number of fans who engage with media texts and contribute creatively on their own terms [19], a significant volume of creative works and related industries have emerged from this subculture. These self-published works, whether original or derivative, are also referred to as doujin and encompass various forms, including comics, literature, music, and cosplay (the practice of dressing up as existing or original fictional characters), with no limitations on specific formats [20]. ACG doujin creators often establish connections with one another and share their creations at doujin conventions, primarily through formats like doujinshi (printed fan comics or fan fiction) and cosplay. They also connect and interact through social networking systems such as Facebook, Twitter, and Plurk [17].

ACG doujin creators share some common features with people with autistic traits. Previous study has indicated that anime fans have a higher prevalence of autism spectrum disorder [21]. Doujin creators also require an exceptionally high level of passion and dedication, devoting themselves in their interested field. Furthermore, the heightened emotions and expressions depicted in anime and comics may resonate more with individuals with higher autistic traits [22], making it easier for them to relate and become engrossed in creating doujin works. However, the prevalence of autistic traits and its correlation with mental health issues were still unexplored in ACG doujin creators.

ACG doujin creators may also have similar challenges “fitting into” the mainstream society as people with autistic traits. Hull et al. demonstrated that the motivations for camouflaging included to feel safe, be viewed as “normal”, be known by their peers, and reduce the stress in social situation [6]. One study indicated that anime fans are susceptible when it comes to discrimination and stereotype, with non-fans referring to them as “socially awkward nerds who are detached from reality”, thus putting them at potential risk of isolation [23]. Although there is no direct evidence indicating that ACG doujin creators face a similar situation as autistic people and have the similar motivations to camouflage themselves, the possibility is high due to the overlapping features of two groups. Since the association between camouflaging and adverse mental health outcomes has been reported, it is possible that ACG doujin creators use camouflaging strategy to cope with their challenge in social situation, which affects their mental health.

Most of previous studies focused on the camouflaging and its mental health consequences in autistic people, but the evidence of association between camouflaging and mental health in subculture groups, such as ACG doujin creators, was limited. In the current study, we tried to make preliminary definition of ACG doujin creators, and investigate both autistic traits and camouflaging behaviours as well as their associations with mental health in this group. In our hypothesis, ACG doujin creators may have similar presentation of camouflaging behaviours as autistic people, instead of non-autistic people. The aims of our study included: (1) to serve as a preliminary investigation on autistic traits, camouflaging, and mental health (depression and anxiety) among Taiwanese ACG doujin creators, (2) to explore the associations between different domains of autistic traits and camouflaging, and (3) to investigate the associations between autistic traits, camouflaging, and mental health.”

Methods

Participant

Participants were recruited through social networking systems using convenience sampling. To reach the target population, our survey was posted and advertised on the public anonymous internet forums of Facebook and Plurk, since these social networking systems are frequently used as platforms for Taiwanese ACG doujin lovers and creators to share their works and communicate with others. The anonymous Facebook page, designed as a platform for people who like drawing to share their thoughts, was followed by about 77,000 people at the time we posted, while all Plurk users could see the post. Required information was collected through the online survey, which was administered via Google Forms. All the participants were well informed that the questionnaire filling was optional, without any form of enforcement. The inclusion criteria below were listed on the first page of the questionnaire for the participants to confirm whether they met the criteria or not before they answered: (1) living in Taiwan; (2) above 18 years old; (3) engaging in ACG doujin creation, including drawing, writing, and cosplaying; (4) having a social networking system (e.g., Facebook, Twitter, Plurk, Instagram, etc.) account or fan page to post their works and communicate with other creators. We included participants who have a social networking system account because of two reasons. First, one of the core activities of ACG doujin creators is sharing their works with their peers. Second, since camouflaging is a social coping strategy, this phenomenon should be found among ACG doujin creators who interact socially with others. Those who did not recognise themselves as ACG doujin creators or failed to meet the inclusion criteria according to their answers were excluded from the study, and their data were not included in the analysis.

Measurement

Socio-demographic variables, doujin activity, and past psychiatric history

Socio-demographic variables such as age, gender (including male, female, and others), and education level were collected. Participants were asked to describe the type of their ACG doujin activity (such as drawing, writing, or cosplaying), by choosing from a set of choices or by a narrative response, and the use of online doujin-related interaction. We also collected their past psychiatric history, and who gave them the diagnosis. All the data mentioned above was self-reported.

General Anxiety Disorder-7 (GAD-7)

GAD-7 is a 7-item self-report instrument commonly used as a screening tool for general anxiety disorder with good reliability [24, 25]. Each item of GAD-7 is scored on a 4-point Likert scale (from 0, not at all, to 4, nearly every day), leading to the total score of GAD-7 ranging from 0 to 28 where a higher score indicates higher anxiety. In a previous research, Chinese version of GAD-7 has showed good reliability and validity in Taiwanese patient with epilepsy [26].

Patient Health Questionnaire-9 (PHQ-9)

PHQ-9 is a 9-item questionnaire with a four-point Likert scale (from 0 “not at all” to 3 “nearly every day”) designed to detect depression in primary care or other medical settings [27,28,29]. According to Kroenke et al., mild, moderate, moderately severe, and severe depression are indicated by PHQ-9 scores of 5, 10, 15, and 20, respectively, with the total score ranging from 0 to 27. With its diagnostic validity and ease of scoring, it has been widely used for both clinical practice and research [28]. Traditional Chinese version of PHQ-9 has been validated and showed good reliability among primary care patients in Taiwan from a previous research [30].

Chinese version camouflaging autistic traits questionnaire (CAT-Q-Ch)

Camouflaging Autistic Traits Questionnaire (CAT-Q) is a questionnaire with a seven-point Likert scale (“strongly disagree” as 1 and “strongly agree” as 7) developed to evaluate autistic adults' subjective experiences of camouflaging, with good reliability and validity [6, 31]. The Chinese version of CAT-Q (CAT-Q-Ch) is an adapted 23-item version to measure camouflaging behaviour and was recently validated in Taiwanese adolescents with high reliability [32]. In our study sample, CAT-Q-Ch showed high internal reliability with a Cronbach’s alpha of 0.898.

The 35-item version of autism-spectrum quotient (AQ-35)

The AQ is a 50-item questionnaire designed to assess autistic traits and is widely used in the general population [33]. A higher score in AQ indicates higher autistic traits. The AQ-35 is a short form of the AQ in Chinese and has been validated in a Taiwanese population [34] with a 5-factor structure (socialness [e.g. I am good a social chit-chat], mindreading [e.g. I find it easy to ‘read between the lines’ when someone is talking to me], patterns [e.g. I am fascinated by numbers], attention to detail [e.g. I tend to notice details that others do no], attention switching [e.g. I frequently get so strongly absorbed in one thing that I lose sight of other things]).” In order to investigate the relationships between different aspects of autistic traits and camouflaging, the total score and subscales of AQ-35 were enrolled as variables. The AQ-35 employed a 4-point Likert scale (ranging from 1–4 in items with autistic features and inverted counting for the reversal items) instead of the dichotomy scoring in its original form [34], which leads to the total score ranging from 35 to 140.

Data analysis

For differences between genders, the Mann–Whitney U test was conducted for continuous variables such as age and scale scores because the hypothesis of the normal distribution was not assumed. Chi-Square tests were used for categorical variables. Pearson correlation was conducted to show the correlations between the variables. Linear regression was conducted for the association between CAT-Q-Ch, total score and subscales of AQ-35, and mental health scores such as anxiety and depression. The results were listed both without adjustment (model 1) and with adjustment for gender and age (model 2). A two-tailed p-value < 0.05 was defined as statistically significant. A 95% confidence interval was also calculated. Statistical analyses of this study were conducted in PASW (Predictive Analytics Suite Workstation) statistics, version 18 (SPSS Inc., Chicago, IL, USA).

Results

A total of 187 people responded to our invitation. Four of them were excluded because they were below 18 years old (n = 1), did not have an account on a social networking system (e.g., Facebook, Twitter, Plurk, Instagram, etc.) or a fan page to post their works and communicate with other creators (n = 2), or did not identify as an ACG doujin creator (n = 1). Among the 183 participants (ages ranging from 18 to 41 [23.9 ± 4.7]), 146 (79.8%) of them identified as female, and 37 (20.2%) identified as male (Table 1). The majority of participants (n = 159, 86.9%) had graduated from university or higher, and 155 (84.7%) had a job. Among all participants, 36 (19.7%) had a history of anxiety disorders, 39 (21.3%) had a history of depressive disorders, 8 (3.8%) had a history of bipolar disorder, and 5 (2.7%) had a diagnosis of autism (Table 2). The CAT-Q-Ch total scores of participants were 101 ± 22.6 (101.2 ± 22.6 for females and 100.4 ± 22.7 for males, respectively), with a score of 91.8 ± 11.0 (92.4 ± 11.3 for females and 89.8 ± 9.7 for males, respectively) for AQ-35, 13.3 ± 7 (13.7 ± 6.7 for females and 12.1 ± 8.1 for males, respectively) for PHQ-9, and 11.0 ± 5.8 (11.1 ± 5.6 for females and 10.5 ± 6.6 for males, respectively) for GAD-7. There was no significant gender difference in any of the scales.

Table 1 Gender differences in ACG doujin creators: age, CAT-Q-Ch, AQ-35, PHQ-9 and GAD-7
Full size table
Table 2 Gender differences in ACG doujin creators: education level, job and past history of psychiatric disorders
Full size table

The association between subscales of AQ-35 and CAT-Q-Ch

After controlling for age and gender, total score of AQ-35 was significantly associated with camouflaging (B = 1.372 [0.791 – 1.952], p < 0.001 in model 2 [adjusted with age and gender]) (Table 3). Among the subscales of AQ-35, socialness (B = 1.313 [0.855 – 1.771], p < 0.001), patterns (B = 0.992 [0.079 – 1.905], p = 0.033), attention to details (B = 2.046 [0.568 – 3.525], p = 0.007), and attention switching (B = 1.063 [0.038 – 2.087], p = 0.042) had a significantly positive association with CAT-Q-Ch total score, except the mindreading subscale (B = -0.347 [-1.113 – 0.418], p = 0.372) (Table 4). Age, but not gender, had a significantly negative association with CAT-Q-Ch total score (B = -1.184 [-1.790 – -0.578], p < 0.001). The correlations between total score and subscales of AQ-35, CAT-Q-Ch, and mental health outcomes were showed in Supplementary Table 1.

Table 3 The linear regression model of CAT-Q-Ch and AQ-35 total score
Full size table
Table 4 The linear regression model of CAT-Q-Ch and AQ-35 subscales
Full size table

The association between AQ-35, CAT-Q-Ch, and mental health

CAT-Q-Ch had a significantly positive association with PHQ-9 and GAD-7 (B = 0.114 [0.069 – 0.159], p < 0.001, and B = 0.080 [0.042 – 0.119], p < 0.001, for PHQ-9 and GAD-7, respectively) (Tables 5 and 6). AQ-35 also showed a significant positive association with both PHQ-9 and GAD-7 (B = 0.128 [0.039 – 0.217], p = 0.005, and B = 0.130 [0.054 – 0.207], p = 0.001, respectively) after controlling for age and gender. Age showed a significant negative association with PHQ-9 (B = -0.205 [-0.401 – -0.010], p = 0.040), but it was not significant for GAD-7.

Table 5 The associations between PHQ-9 and CAT-Q-Ch/AQ-35
Full size table
Table 6 The associations between GAD-7 and CAT-Q-Ch/AQ-35
Full size table

Discussion

Our study was the first to focus on the associations between autistic traits, social camouflaging behaviours, and mental health scores in an ACG doujin creator subculture group. To summarise, we found that mean AQ-35 score of our participants was descriptively higher than in a previous study on Taiwanese parents of autistic children [34]. In ACG doujin creators, CAT-Q-Ch exhibited a significantly positive association with total score and the most of subscales in AQ-35, except for the mindreading subscale. We also found that both CAT-Q-Ch and AQ-35 had a significant positive association with depression and anxiety after controlling for age and gender.

The ACG doujin creators had a higher mean AQ-35 score (91.8 ± 11.0) than in a previous study on Taiwanese parents of autistic children (71.87 ± 12.29 and 76.49 ± 11.70 for mothers and fathers, respectively) [34]. Dubourg and Baumard suggested in previous study that the preference for systemisation may explain why people are attracted to imaginary worlds [35]. Autistic people score higher on the systemising quotient [36, 37], and may be more attracted to imaginary world, which ACG doujin subculture is abundant with, leading to descriptively higher mean AQ-35 score in Taiwanese ACG doujin creators. Besides, our sample also had a descriptively higher level of CAT-Q-Ch (101 ± 22.6, average score 4.39) than both autistic and non-autistic (88.3 ± 23.8 and 77.2 ± 21.1, respectively) and male and female adolescents (81.1 ± 23.6 and 85.6 ± 21.8, respectively) in our previous study [32]. Although it is challenging to directly compare data between different studies, these findings suggest a possibility that the ACG doujin creator group may exhibit different features of autistic traits and camouflaging behaviours compared to the general population. Previous research suggested that members of stigmatised groups tend to hide their fan identity when facing stigma, which may lead them to deny experiencing personal discrimination [38]. We thus speculated that being in a stigmatised subculture [39], when interacting with others, ACG doujin creators may not express what they actually feel and think or what they are interested in on account of potential risk of isolation. They thus develop patterns that can be categorised as camouflaging, trying to avoid the situations which may render them or their fandoms vulnerable. In such a context, camouflaging behaviours may not only be used to hide their autistic traits, but may also to hide their ACG interests, which may subsequently cause more stress than the general population.

Our study found a significant association between autistic traits and camouflaging behaviours among Taiwanese ACG doujin creators. Most of the subscales of AQ-35 exhibited a significantly positive correlation with CAT-Q-Ch, except for the mindreading subscale. There is ample evidence illustrating the association between autistic traits and camouflaging. Hull et al. found that the total score of CAT-Q was significantly and positively correlated with self-rated autistic traits (assessed by the Broad Autism Phenotype Questionnaire) in both autistic and non-autistic adults [31]. Livingston et al. also revealed that self-rated higher autistic traits, which was assessed by ten-item autism spectrum quotient, was associated with a higher compensation strategy in a mixed sample of non-autistic and autistic adults [40]. In our study, the mindreading subscale of AQ-35 was the only subscale without a significant association with camouflaging. However, it is not new that the ability to employ theory of mind does not predictive camouflaging. Hull et al. had speculated that the inability to report actual success rate of camouflaging by CAT-Q, which may be elevated by good theory of mind abilities, is possible the main reason [41]. This explanation may also have a role in our scenario. Besides, we suspected that as the previous study suggested [42], autistic people may still be motivated to create and maintain friendships with others. For ACG doujin creators, the idea of “circle of friends” is closely bonded with doujin events or art works. Social behaviours among ACG doujin creators include selling and exchanging their works with other creators in doujin events, or attracting attention through posting their works on social networking systems. In the context of specialised social behaviours among doujin creators, the ability of mindreading may not influence the camouflaging behaviours as well as other subscales do.

After controlling for age and gender, camouflaging and autistic traits remained significantly associated with both depression and anxiety in this study. Our study aligned with previous research suggesting that both autistic traits and camouflaging are associated with an increased risk of anxiety and depression [43, 44]. However, unlike our study, previous research also showed that in the general population, men are more likely to exhibit more autistic traits, while anxiety and depression are more common in women [45,46,47]. We speculated that our finding of no significant gender differences in autistic traits, depression, or anxiety may be related to the features of the ACG doujin subculture, which creates a welcoming space for individuals with social communication difficulties to interact with ACG enthusiasts. This feature may attract females with higher autistic traits and provide them with a relatively more supportive environment, thus reducing the gender gap in depression and anxiety within this subculture, despite the rates of anxiety and depressive disorders were still high (26.0% and 22.6%) in the female group.

Limitations

This study had several limitations. First, this is an exploratory study with a small sample size and our sample may not represent the Taiwanese ACG doujin creator subculture group. Moreover, the study sample was female predominant. While the demographic data of the ACG doujin creator population was unknow, the representativeness of our study group should be carefully concerned. Caution regarding selection bias should be exercised when interpreting our study results, and it is important to note that our results cannot be generalised to the general population. Second, it is challenging to define a subculture group, and individuals may belong to more than one subculture simultaneously. We attempted to confirm qualifications through both objective and subjective means by asking participants to self-identify and assessing their participation in doujin-related activities. Nevertheless, this process was somewhat crude, and it may overlook the potential for participants to be influenced by another subculture apart from ACG doujin creators. Moreover, as being mainly based on self-rated measurements, it should be noted that the results in this study may not reflect the objective conditions of the participants, considering that rater bias may influence the scores. Finally, this was a cross-sectional study, so we were unable to establish a chronological order for these conditions. In other words, we cannot determine whether pre-existing mental health issues developed before or after someone engaged in ACG doujin creation.

Conclusions

This study has delved into the connections between autistic traits, camouflaging behaviours, and mental health scores among Taiwanese ACG doujin creators. Our findings indicated that within this particular subculture, ACG doujin creators exhibit relatively higher level of autistic traits, compared to a previous study on Taiwanese parents of autistic children [34]. Furthermore, camouflaging behaviours were significantly positively correlated with autistic traits except for the mindreading subscale, and had negative associations with mental health outcomes, particularly increasing the risk of depression and anxiety.

While our study had certain limitations, these findings still provided valuable insights into the experiences and needs of ACG doujin creators. This study may be a pioneer to raise the concern of the relationship between camouflaging and mental health of subculture groups, and future research, based on our study, can explore further details about challenges in social situations and its possible adverse mental health impact in subculture groups.