Changes over Age in Eye-Gaze Pattern in ASD from Childhood to Adolescence: A Cross-Sectional Eye-Tracking Study

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
A. Vincon-Leite1, E. Rechtman1, E. Douard1, A. Philippe2, N. Chabane3, H. Lemaître4, J. M. Tacchella1, F. Brunelle1, N. Boddaert1, A. Saitovitch1 and M. Zilbovicius1, (1)INSERM U1000, Institut Imagine, Paris, France, (2)UMR 1163, Institut Imagine, Paris, France, (3)INSERM U1000, Paris, France, (4)INSERM U1000, Institut Imagine, Université Paris Sud, Paris, France
Background:  Deficits in social interaction, notably difficulties to establish a direct eye-contact, are a core characteristic in ASD. Thanks to eye-tracking methodology, eye-gaze behavior, crucial for human interaction, can be measured objectively. It is now well documented that persons with ASD, adults and children, met a specific eye-tracking pattern when they passively view a dynamic social scene. Indeed, compared to individuals with typical development (TD), persons with ASD present less interest for facial cues conveying social indices useful to understand the main running social scene and they present more interest to background non-social details. To our knowledge, no study has demonstrated changes linked to age of this eye-gaze pattern in ASD and TD children using the same stimuli and across a large and representative cohort extending from infancy to adolescence.

Objectives:  In this context, the current study sought to investigate (i) whether individuals with ASD exhibit differences in eye-tracking pattern compared to TD children; and (ii) how this eye-gaze pattern could change with age during the childhood and adolescence period in ASD and TD children.

Methods: Forty-four children with ASD (age = 8.7 ± 3.7, range: 2.3-16) and forty-six TD children (age = 9.5 ± 3.5 range: 2.6 -17.9) participated in this study. ASD diagnosis was based on DSM IV-R and ADI-R criteria. Tobii-T120 eye-tracker was used to measure eye-gaze processing during passive visualization of social movies displaying characters engaged in peer to peer social interactions (Saitovitch et al., 2016). Viewing time was measured in areas with strong social contents (face-region) and in non-social areas (background-region). Firstly, viewing time was compared between ASD and TD children. Subsequently, a correlation analysis was performed between age and viewing time to the face or background-region. For both analyses we used linear regression model.

Results: Compared with TD children, children with ASD had significantly reduced viewing time to the face-region (p<0.001) and significantly increased viewing time to the background non-social region (p<0.001). Moreover, for children with ASD, we described a significant positive correlation between viewing time to face-region and age (p<0.001) as well as a significant negative correlation between viewing time to the background non-social region and age (p<0.001). For TD children, viewing time toward facial and background non-social regions remains stable and invariant across the period studied (interaction group: age p<0.05). Actually, quasi all TD children, independently of their age and since the early infancy, fixed the face-region more than 80% of the social movie duration and the background-region less than 20%.

Conclusions: The present cross-sectional study confirmed differences in viewing time to face and background non-social region between ASD and TD children. Furthermore, we present for the first time a correlation between age and viewing time to face and background-region in children with ASD that could reflect an adaptation and learning process. We can suppose that fixation to face is innate in TD children whereas for children with ASD it constitutes a learned behavior that change with age. Longitudinal studies are needed to confirm these results.