In this study, the chase detection paradigm was used to compare the way that individuals with and without ASD processed gaze cues in dynamic interactive situations, and the eye-tracking method was used to explore the impact of gaze cues on their visual and cognitive processing during detection.
Our behavioral data show that in the low (15°) and moderate (45°) chasing subtleties, the influence of directional gaze cues did not appear to have obvious effects on accuracy of detection of either group. Under the high (75°) chasing subtleties, on the other hand, participants were influenced more by gaze cues. This suggests that when the task is of moderate difficulty, individuals will rely more on agents themselves to detect the chase, such as the way they move and how far away they are. When the chase subtlety cues provide less helpful information to accurately detect a chase, the role of gaze cues will be revealed (that is also the meaning of setting the chasing subtleties). Further analysis shows that when the chasing subtlety is high, the accuracy of the TD group under the oriented condition is significantly higher than that under the reverse or perpendicular cue conditions. These participants showed a detection advantage under oriented conditions, which was similar to previous studies (Gao, McCarthy, & Scholl, 2010; Scholl & Gao, 2013). In contrast, ASD group showed no detection advantage under the oriented condition. When the chasing subtlety was high, there was no significant difference in their accuracy between the oriented condition and the other two gaze cue conditions, results consistent with our hypothesis.
Defects of executive functions, such as response inhibition and poor cognitive flexibility, will affect keyboard responses of ASD group, thereby preventing behavioral data from revealing their processing pattern. To overcome this bias, we collected eye movement data during chase detection trials. These data were incorporated into the working model of chase detection described by Roux and his colleagues (2015) to further comprehend detection processing of ASD participants.
According to this model, the process of chase detection includes two aspects. The first is attention to the stimulus, which is mainly represented by the stray-looking rate describing the effects of low-level eye movement factors. The second is exploration, which includes the visual response stage and the behavioral response stage. Visual responding is measured by ocular sensitivity which represents the extent to which the API reveals the implicit detection of chasing. The behavioral response stage is measured by cognitive sensitivity which shows the extent to which explicit chase responses reflect the implicit detection of chasing. In addition, an individual’s exploration strategies are mainly determined by their agent-looking rates. Accordingly, we will discuss the influence of gaze cues on the two stages of chase detection processing.
Previous studies have provided evidence of how eye movement patterns reflect global processing strategies (Bombari, Mast, & Lobmaier, 2009; Kucharský et al., 2020; Lemieux, Collin, & Nelson, 2015). In our eye-tracking data analysis, the proportions of each participant’s eye gaze located within different areas of interest were calculated (agent-looking rate and stray-looking rates). We noticed that when TD participants detected the chase among four moving balls, they focused more on the central area of the balls rather than on individual balls. This suggested a processing strategy whereby TD participants tended to integrate multiple stimuli by paying close attention to their displacement relationships with each other. Under the oriented condition, the agent-looking rates of the TD group were significantly higher than these rates in other conditions. This suggested that the oriented gaze cues helped the TD group detect chasing by enabling the participants to direct their eye gazes more to each agent than on the center of the agents.
For the ASD group, the agent-looking rate was approximately 0.63, significantly higher than that of the TD group, whether under the oriented, perpendicular or reverse cue conditions. This suggests that young adults with ASD are less likely to adopt the integral detecting strategy and more inclined to locate and process agents separately, regardless of the nature of the cue. These participants were insensitive to the gaze cues and were not good at adjusting detection strategies according to the different directions of the gaze cues. Our finding is consistent with a weak central cognitive processing style of ASD (Happé & Frith, 2006; Skorich et al., 2016).
This viewpoint was also supported by the behavioral results showing that the chase detection accuracy of the two groups decreased as the difficulty of the chase subtlety increased, and the TD group had a greater reduction than the ASD group. This difference indicates that the chase detection performance of ASD group was less dependent on chasing subtlety than that of TD. These results were similar to those of Vanmarcke et al. (2017). They suggested that the processing of the motion relationship between agents has lost its saliency with higher chasing subtlety and a more attention-focused processing of each of the moving balls gradually became a better search strategy.
Previous studies in the field of social attention have shown that individuals with ASD process social information differently from TD individuals. Vlamings et al. (2005) examined reflexive visual orienting following eye direction or symbolic (arrow) cues, either congruent or incongruent with target presentation, in persons with and without ASD. These investigators found that both the incongruent eye direction cues and the arrow cues enabled individuals with ASD to detect the direction of a target more quickly. This indicates that instead of a specific ‘eye direction detector’, persons with ASD might have a general ‘symbol direction detector’. Other researchers found that eye gaze cues attracted attention more effectively than the arrow in TD children, while children with ASD shifted their attention equally in response to eye gaze and arrow direction, suggesting they failed to show preferential sensitivity to the social cue (Senju, Tojo, Dairoku, & Hasegawa, 2004). Combining behavioral and eye movement data, we guessed that the TD group socialized the gaze cues by focusing more on where the gaze is pointing. In the oriented condition, there was the ball being chased in the range indicated by the line of sight, which helped TD participants detect a chase. However, under the reverse and perpendicular conditions, the range of the line of sight was not followed by the chased ball, and the inconsistency between the line of sight and the direction of motion violated their social perception (Lawson, Aylward, Roiser, & Rees, 2018; Schultz & Bülthoff, 2013; Tremoulet & Feldman, 2000). Therefore, their detection accuracy was lower under these two conditions. But individuals with ASD are different from them, and there may be no difference in processing pattern between social cues and non-social cues.
Moreover, Šimkovic and Trauble (2015) explored the role of eye movements in the detection of chasing. They argued that subjects do not compare the movement of the pursued pair to a singular template that describes a chasing motion. Rather, subjects bring certain hypotheses about what features of motion may qualify as chase and then, through feedback, they learn to look for a motion pattern that maximizes their performance. However, we did not observe this gradually optimized processing strategy in the eye movement data of the ASD group whose agent-looking rates were comparable in all gaze cue conditions. This indicated that their visual detection strategy was relatively stable not affected by gaze cues. This result also confirmed the EM hypothesis that individuals with ASD have difficulty changing their strategies to adapt to a change in the environment (Colombetti, 2012; Klin, Jones, Schultz, & Volkmar, 2003).
The results of sensitivity analysis showed that the ocular sensitivity of participants with ASD was generally higher than that of the TD group, while the cognitive sensitivity was generally lower. It revealed that the implicit, early and online detection of chasing was intact in participants with ASD. Their eye movements were more related to the presence of a chase, suggesting that they may have more often produced ocular detection of chasing and the chasing information had been correctly processed at the visual level. Furthermore, their preferred local looking strategy partly explained the increased ocular sensitivity. The decreased cognitive sensitivity revealed diffculties deciding whether a chase was present or not and/or producing the appropriate response, even when their eye movement (Ramus, Passerieux, & Roux, 2015). This result may have some implications for the intervention of the later explicit cognitive stages in the cognitive processing of individuals with ASD.
Although this study supports the hypothesis that the processing of gaze cues in individuals with ASD is atypical, there remain limitations in this study that should be addressed in future research. Participants selected for this study were younger adults with ASD because the chase detection paradigm requires higher attention and comprehension of participants. Children with ASD may be less likely to understand task requirements or able to maintain a longer period of concentration (Bröring et al., 2018; Dawson et al., 2004). Therefore, any conclusions based on this study may be limited to related phenomena in adults with ASD. Studies have shown that in TD individuals, the ability to estimate whether an object is social by motor cues appears early in infancy, but this ability is impaired in individuals with ASD (Klin & Jones, 2006; Träuble, Pauen, & Poulin-Dubois, 2014; Weisberg et al., 2014). The abnormal performance of their social cognition may result from a severe lack of social information input at an early stage, leading to insufficient social information processing needed to promote the development of corresponding systems within the nervous system (Fischer, Koldewyn, Jiang, & Kanwisher, 2014; Unruh et al., 2016).
Future studies should be designed with more accessible and understandable experimental tasks to explore the causes of social cognitive abnormalities in children and adolescents with ASD, if possible, utilizing a developmental perspective. In addition, in this study, eye-tracker technology was used to record the subject's eye movements during the behavioral experiment. This method enabled us to distinguish their processing strategies at different phases of the task. Our findings show the value of eye movement technology in exploring dynamic interactive social processing-related problems of individuals with ASD. Using neurophysiological recording methods such as ERP and fMRI to further understand the neural mechanism and physiological activation of individuals with ASD may also be good ways to explore their processing of social information.
In conclusion, this study supports the hypothesis that the role of a variety of gaze cues in the chase detection process of individuals with ASD was significantly different from that of TD individuals. In contrast to TD participants, individuals with ASD utilize an atypical processing pattern, which makes it difficult for them to use social information contained in oriented gaze cues. This finding may provide a new perspective on theoretical hypotheses and improve our understanding of the social information processing in individuals with ASD in real life.