Friend matters: sex differences in social language during autism diagnostic interviews

A cross-dataset study on automatic detection of autism spectrum disorder from text data

OBJECTIVE

The goals of this article are as follows. First, to investigate the possibility of detecting autism spectrum disorder (ASD) from text data using the latest generation of machine learning tools. Second, to compare model performance on two datasets of transcribed statements, collected using two different diagnostic tools. Third, to investigate the feasibility of knowledge transfer between models trained on both datasets and check if data augmentation can help alleviate the problem of a small number of observations.

METHOD

We explore two techniques to detect ASD. The first one is based on fine-tuning HerBERT, a BERT-based, monolingual deep transformer neural network. The second one uses the newest, multipurpose text embeddings from OpenAI and a classifier. We apply the methods to two separate datasets of transcribed statements, collected using two different diagnostic tools: thought, language, and communication (TLC) and autism diagnosis observation schedule-2 (ADOS-2). We conducted several cross-dataset experiments in both a zero-shot setting and a setting where models are pretrained on one dataset and then training continues on another to test the possibility of knowledge transfer.

RESULTS

Unlike previous studies, the models we tested obtained average results on ADOS-2 data but reached very good performance of the models in TLC. We did not observe any benefits from knowledge transfer between datasets. We observed relatively poor performance of models trained on augmented data and hypothesize that data augmentation by back translation obfuscates autism-specific signals.

CONCLUSION

The quality of machine learning models that detect ASD from text data is improving, but model results are dependent on the type of input data or diagnostic tool.

Developmental Language Differences in Children with Autism Spectrum Disorders and Possible Sex Difference

Developmental difference is a common characteristic of autism spectrum disorder (ASD) with unclear sex differences. The current study included 610 children with ASD, aged between 2 and 7 years, with completed language profiles. We used a nonparametric item response theory model called Mokken scale analysis to examine the order of acquisition of developmental language milestones in children with ASD. Our results demonstrated the developmental language differences in the expressive and receptive language dimensions in children with ASD compared with typical developmental sequences. Furthermore, The acquisition of gestures and pragmatics was more impaired in the female subgroup than in the male subgroup. The identified developmental language sequence could help provide a more comprehensive ASD developmental profile.

Is There a Bias Towards Males in the Diagnosis of Autism? A Systematic Review and Meta-Analysis

Autism is more frequently diagnosed in males, with evidence suggesting that females are more likely to be misdiagnosed or underdiagnosed. Possibly, the male/female ratio imbalance relates to phenotypic and camouflaging differences between genders. Here, we performed a comprehensive approach to phenotypic and camouflaging research in autism addressed in two studies. First (Study 1 - Phenotypic Differences in Autism), we conducted a systematic review and meta-analysis of gender differences in autism phenotype. The electronic datasets Pubmed, Scopus, Web of Science, and PsychInfo were searched. We included 67 articles that compared females and males in autism core symptoms, and in cognitive, socioemotional, and behavioural phenotypes. Autistic males exhibited more severe symptoms and social interaction difficulties on standard clinical measures than females, who, in turn, exhibited more cognitive and behavioural difficulties. Considering the hypothesis of camouflaging possibly underlying these differences, we then conducted a meta-analysis of gender differences in camouflaging (Study 2 - Camouflaging Differences in Autism). The same datasets as the first study were searched. Ten studies were included. Females used more compensation and masking camouflage strategies than males. The results support the argument of a bias in clinical procedures towards males and the importance of considering a 'female autism phenotype'-potentially involving camouflaging-in the diagnostic process.

Examining gender effects in autistic written language skills: A small sample exploratory study

Background and aims

Gender differences in the written language of autistic individuals are an overlooked but important area of research. We contend that the gender differences in spoken language of autistic individuals may extend to written language, mirroring the gender differences of writing in the general population and reflecting the shared dimensionality of oral and written language. Our research question was: Do autistic adolescent females demonstrate written language characteristics, across persuasive, expository, and narrative genres, that are distinct from those of autistic adolescent males and non-autistic (NA) adolescent females?

Methods

We performed a secondary, exploratory analysis on writing samples collected from 18 participants (11 autistic males, three autistic females, and four NA females) from a larger investigation of autistic adolescents' writing skills. Each participant completed three writing samples-one persuasive, one expository, and one narrative (for a total of 54 writing samples). We compared sample length (total number of words), writing productivity (words written per minute), syntactic length (mean length of T-unit in words), vocabulary diversity (type-token ratio), and macrostructure of autistic females' samples to autistic males' and NA females' samples.

Results

Based on non-parametric analyses using variable medians, autistic males, but not autistic females, wrote significantly shorter expository samples than NA females. Autistic males' writing productivity was significantly lower in the persuasive and expository genres than both autistic females and NA females. Several other comparisons of sample length, productivity, vocabulary diversity, and persuasive and narrative macrostructure yielded large effect sizes but were not statistically significant.

Conclusions

Though our small sample sizes prevent us from drawing generalizable conclusions, we observed that some gender-specific findings of the current study differ from previous findings based on a single autistic group (females and males combined). Combining data of autistic females with autistic males may cloud the distinct written language characteristics of each group.

Implications

Our findings, especially when situated in the context of relevant literature, suggest that larger-scale investigation of gender differences in written language is essential in order to more fully describe the unique characteristics of autistic females. Clinicians should be prepared to support autistic writers' needs for producing written language to meet their developmental, academic, social, and employment-related goals.

Sex/Gender Differences in the Language Profiles of Italian Children with Autism Spectrum Disorder: A Retrospective Study

Sex/gender (S/G) differences in ASD language profiles have been poorly investigated. The present study aims to explore whether male (M) and female (F) children with ASD and with normal non-verbal cognitive abilities differ in their linguistic profiles. A sample of 76 Italian children with ASD (range: 4.9-8 years), including 50 Ms and 26 Fs, was retrospectively recruited. Language profiles were analyzed using standardized tests for the evaluation of receptive and expressive vocabulary as well as grammar. Grammatical comprehension was the most impaired domain compared to the other language measures in both M and F children. Comparing language profiles between S/G, Fs showed significantly better scores than Ms in grammatical production (p = 0.002), and Ms showed better active negative sentence comprehension (p = 0.035). Moreover, comparing the language profiles between Ms and Fs with a receptive disorder, Fs had significantly worse grammatical comprehension and better grammatical production than Ms. Even among children without a receptive disorder, Fs had significantly higher grammatical production scores. The S/G differences in language profile, particularly better expressive language in Fs than Ms, can partially contribute to the delayed ASD diagnosis or underdiagnosis of Fs without intellectual disability. Finally, the results document the importance of accurately investigating both expressive and receptive abilities in children with ASD.

Sex differences in the temporal dynamics of autistic children's natural conversations

BACKGROUND

Autistic girls are underdiagnosed compared to autistic boys, even when they experience similar clinical impact. Research suggests that girls present with distinct symptom profiles across a variety of domains, such as language, which may contribute to their underdiagnosis. In this study, we examine sex differences in the temporal dynamics of natural conversations between naïve adult confederates and school-aged children with or without autism, with the goal of improving our understanding of conversational behavior in autistic girls and ultimately improving identification.

METHODS

Forty-five school-aged children with autism (29 boys and 16 girls) and 47 non-autistic/neurotypical (NT) children (23 boys and 24 girls) engaged in a 5-min "get-to-know-you" conversation with a young adult confederate that was unaware of children's diagnostic status. Groups were matched on IQ estimates. Recordings were time-aligned and orthographically transcribed by trained annotators. Several speech and pause measures were calculated. Groups were compared using analysis of covariance models, controlling for age.

RESULTS

Autistic girls used significantly more words than autistic boys, and produced longer speech segments than all other groups. Autistic boys spoke more slowly than NT children, whereas autistic girls did not differ from NT children in total word counts or speaking rate. Autistic boys interrupted confederates' speech less often and produced longer between-turn pauses (i.e., responded more slowly when it was their turn) compared to other children. Within-turn pause duration did not differ by group.

LIMITATIONS

Our sample included verbally fluent children and adolescents aged 6-15 years, so our study results may not replicate in samples of younger children, adults, and individuals who are not verbally fluent. The results of this relatively small study, while compelling, should be interpreted with caution and replicated in a larger sample.

CONCLUSION

This study investigated the temporal dynamics of everyday conversations and demonstrated that autistic girls and boys have distinct natural language profiles. Specifying differences in verbal communication lays the groundwork for the development of sensitive screening and diagnostic tools to more accurately identify autistic girls, and could inform future personalized interventions that improve short- and long-term social communication outcomes for all autistic children.

Sex differences in friendships and loneliness in autistic and non-autistic children across development

BACKGROUND

Autistic children have been shown to have less complete definitions of friendships and higher levels of loneliness than their non-autistic peers. However, no known studies have explored sex differences in autistic children's understanding of friendships and reported loneliness across development. Autistic girls demonstrate higher levels of social motivation than autistic boys and appear to "fit in" with their peers, but they often have difficulty recognizing reciprocal friendships during middle childhood. As autistic girls develop a more complex understanding of friendship during adolescence, they may begin to redefine their friendships and experience heightened loneliness. Here, we explored how autistic and non-autistic boys and girls define the meaning of friendship and report feelings of loneliness across development. We also examined their perceptions of friendships and loneliness.

METHODS

This mixed-methods study analyzed the transcribed clinical evaluations of 58 autistic children (29 girls) matched to 42 non-autistic children (21 girls) on age and IQ. Transcripts were coded for four categories that children used to define friendships-personality, companionship, dependability, and intimacy-and for reported loneliness. We then compared these codes across diagnosis, sex, and age. Content analyses were further implemented to gain a more holistic understanding of children's perceptions of friendships and loneliness.

RESULTS

Girls, regardless of diagnosis, were more likely than boys to refer to personality when defining the meaning of friendship, and the likelihood of referring to dependability and intimacy increased with age. Most children reported having at least one friend, though some autistic adolescents reported not having friends or were uncertain whether they had friends. While autistic and non-autistic boys and girls were equally likely to report feeling lonely at times, several autistic girls and boys reported being frequently lonely.

LIMITATIONS

This study was a secondary data analysis. The standardized set of questions on the ADOS limited the amount of information that children provided about their friendships and perceptions of loneliness.

CONCLUSION

As with non-autistic children, autistic children acquire a more complex understanding of friendship throughout development. However, as children begin to prioritize dependability and intimacy in friendships, autistic adolescents may have difficulty developing friendships characterized by these constructs. Furthermore, the quantity and/or quality of autistic children's friendships may not be sufficient to alleviate loneliness.

Autism in Women

Relative to males, women with autism spectrum disorder (ASD) have neurobiological and clinical presentation differences. Recent research suggests that the male/female ASD prevalence gap is smaller than previously reported. Sex differences in symptom presentation as well as the male bias of ASD account for delayed/missed diagnosis among women. Investigating ASD and providing psychological evaluation referrals for women who are struggling socially and present with complex mental health conditions (e.g., ADHD, depression), even when they do not show typical autistic characteristics, is important. Accurate diagnosis facilitates understanding of challenges, increases access to treatments, and alleviates the burden of ASD.

A comparison of methods for measuring camouflaging in autism

Interest in social camouflaging has led to a multiplicity of measurement methods of uncertain validity. This two-part investigation first used a systematic review ("Study 1") to identify and appraise methods used to quantify camouflaging of autistic traits, using the Consensus-based Standards for the Selection of Health Status Measurement Instruments checklist. A total of 16 distinct measurement tools were identified; all are in the preliminary phases of psychometric evaluation. The systematic review highlighted: (1) the need for parent-report tools which specifically measure camouflaging; and (2) a lack of studies looking at associations between different methods of camouflaging, which limits understanding of their validity. "Study 2" aimed to begin to address these gaps in knowledge. We created a parent-report version of the Camouflaging Autistic Traits Questionnaire (CAT-Q) and evaluated its concurrent validity in autistic young people by examining associations with the self-report CAT-Q and a discrepancy measure. Discriminant validity was investigated by comparing all three methods of measuring camouflaging to a measure of social skills, to test whether they assess a construct distinct from social ability. The self- and parent-report CAT-Q were significantly related (r = 0.47, 95% CI = 0.24-0.65), and were related weakly (r = 0.20, 95% CI = -0.06 to 0.43) and strongly (r = 0.46, 95% CI = 0.23-0.64), respectively, to the discrepancy approach. No measure was associated with social skills. Improving the psychometric properties of these methods, and introducing a novel parent-report measure, may help selection of appropriate methods in future research and integration into clinical practice.

Intellectual disabilities moderate sex/gender differences in autism spectrum disorder: a systematic review and meta-analysis

BACKGROUND

Girls/women with autism spectrum disorder (ASD) are suggested to exhibit different symptom profiles than boys/men with ASD. Accumulating evidence suggests that intellectual disability (ID) may affect sex/gender differences in ASD. However, a systematic review and meta-analysis on this topic is missing.

METHODS

Two databases (MEDLINE and PsycINFO) were used to search for studies reporting sex/gender differences (girls/women versus boys/men) in social communication and interaction, restrictive and repetitive behaviour and interests (RRBIs), sensory processing, and linguistic and motor abilities in ASD. The final sample consisted of 79 studies. The meta-analysis was performed with Review Manager using a random-effects model. Participants with ASD without and with ID were analysed as separate subgroups, and the effects in these two subgroups were also compared with each other.

RESULTS

Girls/women with ASD without ID displayed fewer RRBIs, more sensory symptoms and less problems in linguistic abilities than their boys/men counterparts. In contrast, girls/women with ASD with ID displayed more social difficulties and RRBIs, poorer linguistic abilities and more motor problems than boys/men with ASD with ID. Comparisons of groups of participants with ASD without ID versus participants with ASD with ID confirmed differences in sex/gender effects on social difficulties, sensory processing, linguistic abilities and motor abilities.

CONCLUSIONS

Our results clearly suggest that the female phenotype of ASD is moderated by ID. Among individuals with ASD with ID, girls/women seem to be more severely affected than boys/men, whereas among individuals with ASD without ID, girls/women with ASD may have less symptoms than boys/men. Such phenotypic differences could be a potential cause of underrecognition of girls/women with ASD, and it is also possible that observed phenotypic differences may reflect underdiagnosing of girls/women with ASD.

Personal victimization experiences of autistic and non-autistic children

BACKGROUND

Autistic children report higher levels of bullying victimization than their non-autistic peers. However, autistic children with fewer social difficulties, as measured on the Autism Diagnostic Observation Schedule (ADOS), are more likely to report being bullied. Autistic children with stronger social skills may not only be more likely to identify and report incidents of bullying, but they may also be more likely to interact with their non-autistic peers, increasing their likelihood of being victimized. Autistic girls may be especially at-risk of experiencing bullying victimization, as a growing body of research suggests that autistic girls demonstrate fewer social difficulties and are more socially motivated than autistic boys. Here, we explored reported problems with peers and bullying victimization among a carefully matched sample of autistic and non-autistic boys and girls. Qualitative methods were further implemented to gain a more holistic understanding of the social experiences of autistic boys and girls.

METHODS

This mixed-methods study analyzed the transcribed clinical evaluations of 58 autistic children (29 girls) matched to 42 non-autistic children (21 girls) on age and IQ. Within each diagnostic group, boys and girls were matched on ADOS severity score. We compared reported problems with peers and bullying victimization across sex and diagnosis. Among autistic children, we further examined whether ADOS social affect (SA), restricted repetitive behaviors, and severity scores predicted problems with peers and bullying victimization. We then identified themes related to personal experiences of victimization.

RESULTS

Autistic children were more likely than non-autistic children to have experienced bullying victimization, and autistic children with lower ADOS severity and SA scores were more likely to report having been bullied. While autistic boys and girls reported similar levels of bullying victimization, qualitative analyses revealed sex differences in the underlying causes of peer conflict.

LIMITATIONS

This study was a secondary data analysis. The standardized set of questions on the ADOS limited the amount of information that children provided about their peer relationships, and variations in follow-up questions may have influenced children's responses.

CONCLUSIONS

Although autism symptomatology places autistic children at greater risk for bullying victimization compared to their non-autistic peers, greater social challenges among autistic children are associated with lower rates of victimization. This study further highlights the importance of using mixed-methods approaches to discover nuances in the social experiences of autistic girls and boys that may become opportunities for support.

Exclusion of females in autism research: Empirical evidence for a "leaky" recruitment-to-research pipeline

Autism spectrum disorder (ASD) is characterized by challenges in social communication and the presence of repetitive behaviors or restricted interests. Notably, males are four times as likely as females to be diagnosed with autism. Despite efforts to increase representation and characterization of autistic females, research studies consistently enroll small samples of females, or exclude females altogether. Importantly, researchers often rely on standardized measures to confirm diagnosis prior to enrollment in research studies. We retrospectively analyzed the effects of one such measure (Autism Diagnostic Observation Schedule, ADOS) on research inclusion/exclusion rates by sex in autistic adults, all of whom had a preexisting community diagnosis of autism (n = 145, 95 male, 50 female). Using the ADOS as a confirmatory diagnostic measure resulted in the exclusion of autistic females at a rate over 2.5 times higher than that of autistic males. We compared sex ratios in our sample to those in other large, publically available datasets that rely either on community diagnosis (6 datasets, total n = 42,209) or standardized assessments (2 datasets, total n = 214) to determine eligibility of participants for research. Reliance on community diagnosis rather than confirmatory diagnostic assessments resulted in significantly more equal sex ratios. These results provide evidence for a "leaky" recruitment-to-research pipeline for females in autism research. LAY SUMMARY: Despite efforts to increase the representation of autistic females in research, studies consistently enroll small samples of females or exclude females altogether. We find that despite making up almost 50% of the initially recruited sample based upon self-report of community diagnosis, autistic females are disproportonately excluded from research participation as a result of commonly used autism diagnostic measures. In our sample, and several other publically available datasets, reliance on community diagnosis resulted in significantly more equal sex ratios.