Validation of a Quantitative Approach of the Application of Autism Diagnostic Observation Schedule (ADOS): A Preliminary Eye-Tracking Study

Background: Autism Diagnostic Observation Schedule (ADOS) is a semi-structured assessment of communication, social interaction and imagination. It allows to assess and diagnose autism spectrum disorder (ASD) across ages, developmental levels and language skills. However, its use is dependent on the examiner’s clinical expertise in neurodevelopmental field and experience with the instrument, which may potentially lead to errors in the diagnostic procedure. Eye-tracking is a powerful tool for studying visual attention patterns, revealing differences in the way individuals with typical development (TD) and ASD explore images.  

Objectives: This study aims to assess the psychometric validity of eye-tracking measures to be used as diagnostic tool, by developing and testing a quantitative approach that may provide a more accurate diagnosis and quantify ASD severity.

Methods: The sample consisted of eighteen school-aged children: eleven with high-functioning ASD (mean age=13 years 6 months) and seven chronological-aged matched TD (mean age=12 years 10 months). ASD patients had ADI-R and ADOS positive results and met ASD DSM-5 criteria. The parents of TD participants completed Social Communication Questionnaire and Social Responsiveness Scale to exclude ASD symptomatology. Associated medical conditions were excluded by experienced neurodevelopmental pediatricians in ASD group. Eye-tracking data were collected with RED eye-tracking system (SMI-SensoMotoric Instruments,Germany) for both ASD and TD participants during the administration of digitalized activities from ADOS: “Description of a Picture”,“Cartoons”/“Telling a Story from a Book”. Eye movement data was recorded with iViewX™ and analysed offline with BeGaze™software where different areas of interest (AOI’s) where defined: “faces”/“objects”/“animals”. We compared (IBM-SPSS-21) the total fixation count (TFC) and percentage of total fixation duration time (PTFDT) in the defined AOI’s within the two groups with Mann-Whitney-U test. Significance level (α)=0.05(p<0.05). 

Results: TFC of “objects” in “Description of a Picture” task significantly differed between the two groups (p=0.015), while for “faces” and “animals” it was not significantly different (p>0.05).  ASD group had higher number of TFC in “objects” while describing the given picture than TD group. Although not reaching significant statistical difference, the inverse pattern appears in AOI “faces”, with ASD group having lower TFC than TD. Concerning PTFDT there were no significant statistical differences between groups (p>0.05). In “Cartoons” task both groups did not differ in TFC in the different AOI’s (p>0.05). Total PTFDT of “objects” AOI significantly differed between the two groups (p=0.027), while in “faces” AOI was not significantly different (p>0.05).  

Conclusions: Eye-tracking measures of visual scanning, while describing activities from ADOS: “Description of a Picture”/“Cartoons”, can discriminate between ASD and TD groups concerning to attention to objects in the pictures. ASD individuals are also showing a trend to decreased attention towards faces. In our analysis, we focused on attention to faces, objects and animals in the pictures from tasks of ADOS thereby providing a quantitative diagnostic measure that is not examiner’s dependent. With this strategy, we hope to improve the accuracy of ASD diagnostic evaluation and obtain a quantitative measure of severity, ideally obtaining a spectrum of severity from normal to pathological behaviour. These preliminary findings should be replicated in a larger sample.