Objectives: The current study aims to validate a new joint attention eye tracking paradigm on a population of adults by evaluating difference in attention and pupil reactivity during congruent and incongruent videos.
Methods: 35 neurotypical adults aged 18-50 (M=23.41, SD=7.28; 65% male) completed a background questionnaire, medical questionnaire, and the Broader Autism Phenotype Questionnaire, which measures subtle traits that resemble the characteristic features of autism, including pragmatic language (Hurley et al., 2007). Participants that scored above cutoff did not differ from participants that scored below cutoff on the following descriptive variables: age, gender, or education. Each participant’s eye gaze behavior and pupil size were recorded using a Tobii T-60 eye tracker. The stimuli presented are dynamic, social videos where an adult model shifts her gaze to different corners of the screen. In the congruent condition, the model’s gaze follows a Pokémon character. In the incongruent condition, the model gazes to a non-Pokémon corner. Mean pupil diameter was calculated for five 500-ms intervals time locked to the presentation of the Pokémon. In the first interval the Pokémon appears, but the model does not switch her gaze. Relative change in pupil size was calculated and standardized for each additional time interval. To avoid the pupillary response to changes in luminance (appearance of Pokémon), data analysis included intervals 3 and 4.
Results: RM-ANOVA reveals that adults spend a significantly longer time viewing the Pokémon during congruent videos when compared to incongruent videos, F(1,34)=30.87, p<.01. This finding indicates that adults were more likely to follow the model’s gaze and focus their attention to the Pokémon character during the congruent videos when compared to incongruent videos. A separate RM-ANOVA reveals that change in pupil size is larger in the congruent condition during intervals 3 and 4 when compared to the incongruent condition (F(1,19)=4.11, p=.057; F(1,19)=6.32, p<.05), indicating that cognitive load or emotional response was greater during the congruent condition (which is designed to be social in nature) when compared to the incongruent condition. Finally, participants that scored above cutoff on the pragmatic language subscale of the BAPQ demonstrated a larger change in pupil size during intervals 3, 4 when compared to individuals that scored below cutoff (t=2.06, p=.051; t=2.08, p<.05) during incongruent videos. Meaning, individuals with more BAP traits demonstrated greater pupil reactivity during incongruent clips when compared to individuals showing fewer BAP traits.
Conclusions: Results indicated that congruent and incongruent videos elicit different visual attention and pupil reactivity. We are currently extending this work to typically developing children and children with autism.
See more of: Core Deficits and Symptoms
See more of: Symptoms, Diagnosis & Phenotype