부산대학교 도서관

The first chapter presents a systematic review on the literature quantifying patterns of disfluency in autistic populations. Disfluencies are hypothesised to be an interruption of a speech plan with approximately 6% of spontaneous speech being disfluent. Around half of the autistic population display quantitatively and qualitatively different speech patterns compared with the neurotypical population. Therefore, these different speech patterns could represent a behavioural marker for autism. Twelve studies were screened and met the following inclusion criteria: empirical studies quantifying types of disfluencies in autism, with a sample size greater than five, and the inclusion of a neurotypical comparison group. The results indicated that filled pauses were produced significantly less frequently by autistic individuals relative to neurotypical controls. Autistic individuals produced significantly higher amounts of unfilled pauses, repetitions, and repairs than neurotypical groups. The pattern of results suggest that there may be meaningful differences in the production of disfluencies between autistic and neurotypical individuals. However, there are a number of methodological and conceptual issues (unaccounted moderating and mediating variables, varied methodologies, heterogeneity of autism) to overcome before meaningful conclusions can be drawn and generalised. The purpose of the empirical study was to address the methodological issues that fewer pauses found previously in autistic groups may simply reflect sex differences rather than a phenotypic expression of autism due to sex-imbalanced groups. We included nearly equal numbers of males and females, to compare the use of filled pauses (Um, Uh) between autistic and neurotypical participants, and between sex, as well as the interaction between the two. In addition, filled pause use was also examined to investigate the linguistic camouflaging hypothesis. Eighty-four participants aged 8-13 from mainstream schools completed interviews (structured, unstructured) as well as tests measuring their verbal and nonverbal IQ. Neither significant main effects nor interactions were found for all indices for the structured task. For the unstructured task, no interactions were found. However, autistic participants produced a significantly higher proportion of Um to total words than neurotypical participants. Males produced a significantly higher proportion of Uhs than females, and females produced a higher Um ratio than males. Differences in filled pause production was affected by the type of interviews. The linguistic camouflaging hypothesis was disconfirmed. Future research should include larger samples (e.g., more females in the autistic group, children with high traits), different task types, and listeners’ perceptions in order to further understand filled pauses in autism.