Evaluation of Emotion Recognition in Individuals With Autism Spectrum Disorder: Recommended Designs for Future Studies

ABSTRACT

Background and Aim

Autism spectrum disorder (ASD) is a lifetime neurodevelopmental condition, which its prevalence is on the rise. Difficulties with emotion recognition and perspective‐taking are commonly observed in individuals with ASD, leading to a profound impact on their daily functions. The aim of this document was to offer a reflective perspective on the evolution of research in this area.

Methods

This document evaluates current research approaches. The analysis focused on identifying key trends, gaps, and areas for further investigations.

Results

Most studies that examined emotional face recognition in individuals with ASD yielded mixed findings regarding their ability in recognizing specific types of emotions. It may be considered that the heterogeneous population as well as the stimuli or experimental designs in previous studies play pivotal roles in inconsistencies in findings.

Conclusion

This commentary critically reflects on the previous studies, highlights trends in the current literature, and suggests how assessments can be implemented in future studies.

Emotion recognition is an essential ability for human social interaction. Recognizing the emotions of others enables individuals to understand intentions and anticipate responses. Challenges in interpreting facial expressions and comprehending others' emotional states can be observed in individuals with socio‐emotional difficulties such as autism spectrum disorder (ASD) [1]. Individuals with ASD may show delays and difficulties in developing emotion recognition and social skills [2, 3, 4, 5]. Considering developmental milestones, emotion recognition improves with age, and individuals with ASD have similar learning curves compared to their typically developing (TD) counterparts [6]. Although some studies indicated similar emotion recognition abilities of individuals with ASD compared to their TD peers, those with ASD may face challenges in refining these skills further as they continue into adulthood [7]. However, there are mixed findings and inconsistencies in the literature regarding the type of emotion that children with ASD have difficulty recognizing.

The majority of previous studies have indicated that individuals with ASD might perform slower and less accurate in recognizing facial expressions, particularly negative emotions compared with their TD peers [7, 8]. Some studies have demonstrated that the processing of sad, fearful, angry, and surprised faces is challenging in individuals with ASD [9, 10, 11]. In contrast, other studies showed no significant difference in the performance of emotion recognition in individuals with ASD compared with their TD peers [12, 13]. The primary question is why there is such a huge variation in the findings of previous studies and how future experiments can reliably evaluate emotion recognition abilities in individuals with ASD.

Differences in demographic variables (e.g., age, IQ, severity of ASD, and co‐occurring conditions) may contribute to the variability in results [14, 15]. Furthermore, differences in methodological approaches, sample sizes, and assessment tools (e.g., familiar vs. unfamiliar faces, dynamic vs. static faces, faces with similar vs. different ethnicities or age ranges to participants) across studies may help explain these inconsistencies. The choice of stimuli, which represent different levels of difficulty among individuals with ASD, may play a crucial role. Most of the previous studies examined performances of children with ASD in recognizing emotions rather than analyzing a threshold. The ceiling or floor effects observed in such experiments make it difficult to draw thorough conclusions about which emotions are the most challenging to process, and whether emotions are perceived differently by individuals with ASD compared with their TD counterparts.

It can be suggested that one of the best ways to resolve these ambiguities is to calculate the threshold of emotion recognition or detection using adaptive procedures. This method manipulates the task difficulty based on the subjects' performance via staircase morphed faces, instead of testing all subjects at the same fixed difficulty level [16]. Although this technique allows a sensitive, highly precise, and robust evaluation of the threshold in expression detection or expression recognition among individuals with and without ASD, it has been used in a few studies [13, 17]. The slow emotion recognition or atypical facial processing in individuals with ASD may not be necessarily due to the nature of social stimuli, but rather due to the complexity of the stimuli. Specifically, less‐complex stimuli may result in enhanced perception, whereas more complex stimuli such as faces will lead to reduced performance in individuals with ASD [18]. This may suggest that including nonsocial control stimuli in experiments might be beneficial. Furthermore, presenting stimuli within or out of context, or aligning visual stimuli with auditory cues, while considering their ecological validity, is crucial when interpreting findings.

Despite the growing evidence on socio‐emotional difficulties in individuals with ASD, to the best of our knowledge, very little information has been reported to show comprehensive data on the threshold aligned with the performance of individuals with ASD. Having various types of experiments including expression labeling, expression detection (the ability of detecting the emotional face from the neutral face), expression identification (the matching ability to find the target stimulus among other stimuli), and face‐to‐object identification will help draw a more robust conclusion [15, 19]. It should be noted that most of the previous studies have used a “labeling task,” compared with other types of tasks, as the main experiment. However, individuals with ASD often have semantic difficulties affecting written and oral comprehension [20], which is necessary for some experiments such as labeling tasks that have been frequently used. Although these experiments attempt to evaluate an individual's social and emotional perception, each has unique features that assist in analyzing the nature of socio‐emotional processing in individuals with ASD.

Exploring visual processing as a component of facial processing among individuals with ASD requires further attention. Different patterns of facial/emotion recognition in ASD may derive from their atypical visual processing [21]. Compared to TD individuals, those with ASD may display a unique visual scanning of faces. Some studies indicated shorter fixation and limited attention to the eyes or increased focus on the mouth in ASD [22, 23, 24]. Individuals with ASD often show a detailed or local‐based information processing bias in visual tasks. This bias may interfere with their ability to interpret emotional faces, a task that mostly requires a holistic and global processing of the face [25]. The reduced tendency of individuals with ASD to apply a general global processing approach may result in their poor performance [21]. Although faces and face‐like objects share similar characteristics, previous studies have shown that human faces are detected more quickly than face‐like objects in visual search tasks in TD individuals [26]. This may suggest that human faces are special stimuli, requiring a global processing approach. Thus, experiments may benefit from the combination of eye‐tracking methodology and various visual stimuli to better uncover the processing patterns [27].

Another assumption for such differences in ASD versus TD groups may be related to processing simple aspects of visual stimuli. It has been shown that low‐level visual processing, such as processing of orientation, luminance, or contrast, is not often affected in individuals with ASD, but the high‐level visual processing is affected at the level of integration and interpretation [28, 29]. However, relatively few studies have examined low‐level visual processing in individuals with ASD, suggesting that its further exploration is warranted. Undoubtedly, some of the difficulties in interpreting disparities in individuals with ASD compared to their TD peers lie in the heterogeneous nature of the disorder. ASD is on a spectrum with a diverse range of phenotypes, so atypical perceptual abilities may be associated with a portion of those with ASD. Although atypical visual processing may be due to the attentional or higher level cognitive processes, some individuals with ASD may display different neural activity in the early visual cortex and may have difficulties in subcortical regions involved in face processing [30]. Therefore, without accounting for these variations, interpretations of findings of studies in individuals with ASD compared to TD can be challenging.

To conclude, despite a growing body of literature investigating emotion recognition among individuals with ASD, consideration must be given to the methodological concern of selecting the type of task and difficulty level of the stimuli for such a diverse population. Future studies should take advantage of various experiments in the study design to evaluate both high and low levels of visual processing among individuals with ASD.

Author Contributions

Parisa Ghanouni: conceptualization, data curation, formal analysis, investigation, methodology, supervision, writing – original draft, writing – review and editing. Laura Eves: writing – review and editing.

Disclosure

All authors have read and approved the final version of the manuscript. The corresponding author had full access to all the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the findings.

Conflicts of Interest

The authors declare no conflicts of interest.

Transparency Statement

The lead author, Parisa Ghanouni, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Data Availability Statement

No data is associated with the article, and all data underlying the results is presented.