Abstract:

Language impairment is a salient and one of the earliest symptoms of Autism Spectrum Disorder (ASD). Over the past decades, researchers have employed various neuroimaging techniques (e.g., functional magnetic resonance imaging, fMRI) to explore the neural markers of language impairments in individuals with ASD. Individual investigations have converged on finding differences in neural responses of language processing (e.g., word, sentence, prosody) between autistic participants and typically developing (TD) controls. Results from meta-analytic studies further indicated that the differential neural activities were stable across studies. However, the meta-analysis that focused on semantic processing in the ASD group revealed somewhat different results from the meta-analysis on language processing. In addition, the results of sub-analyses that divided language processing tasks into subtypes reported that the ASD-TD differential activation patterns were moderated by task type. The previous meta-analysis of semantic processing in ASD populations included both sentence-level and word-level semantic processing studies. Considering the differentiated underlying mechanisms of word-meaning and sentence-meaning processing, it is essential to distinguish between these two levels of semantic processing. Lexical-semantic processing ability is at the forefront of language acquisition and a known area of impairment in individuals with ASD. Previous review studies have adequately summarized the behavioral performance of individuals with ASD for semantic processing at the single word level. However, a meta-analysis to investigate the cross-study brain activation patterns of individuals with ASD during lexical-semantics processing is still missing. The current study identified 11 published journal articles that used fMRI to investigate lexical-semantic processing in individuals with ASD. Activation likelihood estimation analysis was adopted to investigate whether atypical brain activation patterns in individuals with ASD were stable across studies and how the atypical performance manifested. The results revealed that the TD group activated the left inferior frontal gyrus, the left superior frontal gyrus, the left middle temporal gyrus, and the left middle frontal gyrus, which are brain regions responsible for completing lexical-semantics processing tasks. Like the TD group, the ASD group also showed activation in typical lexical-semantics processing brain regions, such as the left inferior frontal gyrus, the left middle temporal gyrus, and the left medial frontal gyrus. At the same time, differential brain activation patterns were steadily present between the ASD and TD groups. To be more specific, no brain region was found in the ASD group-TD group subtraction analysis, while the left superior frontal gyrus was present in the TD group -ASD group subtraction analysis. This finding indicated the atypical brain activation patterns of lexical-semantics processing in individuals with ASD manifested as hypoactivation in the left superior frontal gyrus. In addition, exploratory sub-analyses suggested that the atypical brain activation patterns of adults with ASD may differ from those of children and adolescents with ASD. Adults with ASD were more likely to show enhanced activation in the visual processing areas, while children and adolescents with ASD showed reduced activation in the left middle temporal gyrus. In summary, the current meta-analysis provided evidence for the atypical brain responses to lexical-semantics processing in verbal individuals with ASD from a cross-study perspective. Their atypical brain activation activities when processing meanings of words were consistent with the "decreased frontal lobe activation" hypothesis. The current study revealed novel findings which highlighted that distinguishing between different levels of language processing can help researchers identify the corresponding neural markers of certain types of language impairments in autistic populations. Due to the limited number of included studies, the current study had some limitations, such as only including fMRI studies and not considering the task, behavior and general language skills related factors. Future studies should take age, language background, general language skills, intervention experiences, and types of semantic processing into consideration.

Key words: Autism Spectrum Disorder, lexical-semantic, fMRI, meta-analysis