Summary: Social challenges associated with autism may reflect differences that emerge mainly in pressured situations or specific interaction types. New evidence calls into question the widespread idea that autistic adults cannot accurately read others’ facial emotions.
Source: Flinders University
There is a widespread belief that autistic people struggle to recognise others’ emotions and lack awareness of their own performance in this area.
New Australian research suggests a more nuanced picture: autistic adults are, on average, only slightly less accurate at identifying facial emotional expressions than non-autistic adults, and their insight into their own performance is similar.
Two recent papers published in the international journal Autism Research report findings from a comprehensive study that re-examines common assumptions about emotion recognition and metacognitive awareness in autism.
The study tested 63 adults diagnosed with autism and 67 non-autistic adults (IQs ranged from 85 to 143). Participants took part in extended testing sessions—each lasting between three and five hours—during which they judged 12 different facial emotion expressions, including anger and sadness, presented in various formats.
Dr Marie Georgopoulos collected a large dataset during her PhD research. Subsequent reanalyses by the team produced a series of articles that refine our understanding of how autistic and non-autistic adults recognise and evaluate facial emotions.
Rather than showing large, global deficits, the results indicate that social difficulties linked to autism may reflect differences that appear under particular interaction conditions or when tasks are demanding—rather than a general inability to read facial emotions.
Co-author Matthew Flinders Distinguished Emeritus Professor of Psychology, Neil Brewer, explains that by using a wide range of emotions, presented in multiple ways, the study found autistic individuals were, on average, only slightly less accurate and somewhat slower in classifying emotions.
“These results challenge the idea that autistic adults are disproportionately overwhelmed by dynamic or complex emotional cues and unable to recognise specific emotions,” Professor Brewer said.
Performance overlapped substantially between the two groups; only a small subgroup of autistic participants scored below the range of non-autistic participants. Group differences remained consistent across how the emotions were shown, the type of response required, and the specific emotions tested.
The researchers also measured metacognitive awareness—how well people understood whether their emotion judgments were correct. Although individuals varied widely in their self-awareness, there was no evidence of systematic differences between autistic and non-autistic participants.
“The rigorous methods in these studies refine our understanding of emotion processing in autism and highlight capabilities that have often been overlooked,” the authors note.
They add that future progress will likely depend on capturing emotion recognition and reactions in realistic situations, such as live social interactions or immersive virtual reality, where subtle differences may be more detectable.
About this autism research news
Author: Yaz Dedovic
Source: Flinders University
Contact: Yaz Dedovic – Flinders University
Image: The image is in the public domain
Original Research: Open access.
“Facing up to others’ emotions: No evidence of autism-related deficits in metacognitive awareness of emotion recognition” by Marie Georgopoulos et al. Autism Research
Abstract
Facing up to others’ emotions: No evidence of autism-related deficits in metacognitive awareness of emotion recognition
Difficulties recognising emotions in others are often thought to contribute to social and communicative challenges experienced by autistic people. Awareness of these difficulties—metacognitive insight—could be important for identifying and using strategies that reduce their impact.
This study assessed metacognitive awareness of facial emotion recognition in autistic (N = 63) and non-autistic (N = 67) adults across multiple conditions: static, dynamic and social face stimuli; free- and forced-report response formats; and four different sets combining six “basic” and six “complex” emotions.
Analyses linking individual recognition accuracy to post-response confidence revealed no indication that autistic participants were worse at discriminating correct from incorrect responses than non-autistic participants. Both groups showed substantial variability between individuals.
While the autistic group was, on average, slightly less accurate and slower at recognising emotions, confidence–accuracy calibration showed no reduced sensitivity to fluctuations in performance. Increases in accuracy were associated with higher confidence for both groups, producing similar calibration curves across stimulus type, response format and emotion.
Both groups’ calibration curves did show overconfidence at the highest confidence levels (overall accuracy lower than mean confidence), with the non-autistic group contributing more responses rated at 90%–100% confidence.
Comparisons of faster and slower responders found no evidence of a “hard–easy” effect (overconfidence on difficult tasks and underconfidence on easy tasks). This suggests that the slower responses observed in autistic participants may reflect different strategic choices rather than a simple processing speed limitation.