21507
Eye Movement during Reading and Answering Inferential Questions in High-Functioning Autism: Strategies and Cognitive Components
Reading comprehension problems in Autism Spectrum Disorder (ASD) are well documented in the literature, however, it is still unclear which processes contribute to this deficit (Brown, Oram-Cardy & Johnson, 2013). The ability to make inferences has been shown to be essential for discourse (Snyder & Caccamise, 2010) and reading comprehension (Cain, Oakhill, Barnes & Bryant, 2001). The results in literature regarding inference generation across ASD individuals are not entirely consistent. Studies have shown individuals with ASD interpret expressions literally (Loukusa, et al., 2007) and have difficulty drawing inferences (Arciuli et al. 2013; Huemer and Mann 2010; Ricketts et al. 2013). However, others have reported that inference generation was intact in ASD when primed for implicit inferences (Saldaña & Frith, 2007). Hence, the reading strategies underlying inference generation in ASD remain unclear.
Objectives:
First, the present research aimed to investigate whether ASD were as accurate as typically developing (TD) individuals in answering inferential questions. Second, eye movement strategies of ASD and TD during the reading of texts that required answering text-based or inferential questions were explored. Finally, it aimed to explore cognitive components related to gaze behavior during inference generation in ASD.
Methods:
Twenty-two high-functioning children and adolescents with ASD and twenty-two TD participants, matched for chronological age, intelligence scores, language skills, and reading ability, read five stories, each divided into six paragraphs. Following each paragraph participants answered an inferential or text-based question. Global and target word (word defined a priori as fundamental to question responding) eye movement data was recorded continuously during text reading. Cognitive abilities were measured using the Wechsler Intelligence Scale for Children or Wechsler Adult Intelligence Scale, and the Spanish Peabody Picture Vocabulary Test.
Results:
Results showed that ASD were as accurate as TD in responding to inferential and text-based questions. Analyses of eye movement data revealed longer reading time in the inferential condition only for ASD with respect to the target word. No group differences were found with respect to global paragraph reading and responding time or eye movement around distractor words. With regard to cognition, in the ASD group only, results revealed a positive relationship between vocabulary size and inferential accuracy and a negative relationship between perceptual reasoning and re-reading time of the target word in the inferential condition. Finally, analyses of correct answers revealed that ASD spontaneously fixated on the question prior to the text more often in inferential condition and that a positive relationship existed between the frequency of the question-before-text strategy and score of total intelligence, perceptual reasoning, and working memory.
Conclusions:
Despite similar accuracy scores and global paragraph reading, ASD exhibited delayed processing of the target words in the inference condition. In addition, higher cognitive functions appeared to be related to inferencing ability, target word eye movement behavior, and self-initiated question reading strategy.