The Effects of Social Training Games on Visual Gaze to Social Stimuli

Background: There is accumulating interest in the gamification of social-communication skill development for children with autism spectrum disorder (ASD). To date, many games have targeted older individuals with average to high-average cognitive/language abilities, or have limited experimental controls. Our previous research has found that a more inclusive game rewarding attention to faces over objects increased eye-gaze to faces in children after 15 minutes of gameplay. However, it is unknown whether extended gameplay can maintain this effect.

Objectives: This study tested the effects of social training games on visual attention in children with ASD.

Methods: Participants included 65 children (10 females) diagnosed with ASD (mean age=8.59, SD=2.16, range=5-12 years) randomly allocated to play either a training (n=33) or control (n=32) set of games. No exclusion criteria related to cognitive, language, or functional ability were applied. Children were asked to play the games for at least 15 minutes/day for seven days. In each game, children were rewarded for correctly sorting emotions (Emotion Recognition), following a character’s eyes (Joint Attention), or selecting faces over objects (Social Attention). Each game embedded these socially-oriented goals into a child-friendly format, with visual rewards and sounds to maintain engagement, and no written or verbal instructions to enable accessibility for children with minimal receptive language. A control version was designed to ensure equivalent exposure to the visual game components, but without active training contingencies programmed. Children completed eye-tracking pre- and post-gameplay, measuring attention to (1) eye regions of faces, (2) directed eye gaze, and (3) faces relative to objects. Qualitative data from children were collected on their game experiences.

Results: On average, children played an equivalent amount of time in either condition (t(63)=0.46, p=.65), however control games were passed significantly more frequently; suggesting that control games were easier, but time engaging in games were relatively comparable. Compared to before playing the game, children in the training group exhibited no significant changes in visual attention across all eye-tracking tasks after playing the game, relative to the control group (all p values>.05). Daily game usage indicated that 16 (24.6%) children did not engage in one or more of the games for more than one day after returning home; once these children were excluded pair-wise from eye-tracking analyses, there was a borderline significant interaction in gaze following (F(1, 47)=4.15, p=.05, partial eta²=.08); this indicated that children playing the Joint Attention training game significantly increased the speed at which they could follow gaze directed to an object post-gameplay. No other changes in eye-gaze were observed. Qualitative data indicated that the games were well received, with high acceptability ratings.

Conclusions: Games represent an acceptable format to engage children with ASD in social-communication development. Our results suggest that, while engagement with these games was acceptable, specific transfer to changes in visual gaze was limited. These results may be partially influenced by the heterogenous nature of the sample included and variability in gameplay. However, qualitative data supported overall acceptability and highlighted feasibility of designing games for a range of functional, cognitive, and language ability levels.