The physiology of eye contact in children with autism
Appropriate eye contact is an important aspect of our social behavior. Individuals with autism have shown deficiencies in making and maintaining eye contact with others, which contribute to the social difficulties that define the disorder.
A common assumption has been that individuals with autism are indifferent to others, which has also been a motivating factor in reinforcement-based therapies. However, a growing number of research projects have shown that the behavioral lack of eye contact is often accompanied by a physiologically aroused autonomic nervous system (ANS), which points to an oversensitivity, rather than disinterest, in eye contact.
Guided by the hypothesis of an excitatory/inhibitory imbalance in the socio-affective processing in autism (Lasalle et al., 2017), two objectives motivate present research. The first is to explore pupil mimicry (Fawcett et al., 2018; Kret, Fischer & De Dreu, 2015) as an index of autonomic function through physiological responses such as skin conductance and heart rate. Pupil mimicry has been found to be the mechanism by which an observer matches the arousal level with the observed individual. The second objective is to examine how pupil mimicry effect correlates with autistic traits, an individual's sensory profile, and degree of anxiety.
The present project holds theoretical, clinical and social implications for individuals who have difficulty with eye contact and for whom traditional therapies have proven to be inconsequential.
Final report
1. Project Purpose and Development
The project investigated mechanisms underlying children’s social attention and autonomic responses, focusing on pupillary contagion, eye contact, and physiological reactivity in relation to autistic traits. A total of 67 children were tested across two studies, including participants who received an autism diagnosis, those who indicated elevated autistic traits but have not been formally diagnosed, and neurotypical controls. Analytical samples varied by modality due to strict signal quality requirements. The project progressed in accordance with the planned methodological and scientific aims and no major deviations from initial protocol were performed.
2. Implementation
Data collection was conducted in a controlled laboratory environment (Gillberg Neuropsychiatry Center in Göteborg, Sweden) using two complementary paradigms:
• Study 1 – Pupillary Contagion Task: Children viewed photographs of peers with manipulated pupil sizes (large vs. small) under free-viewing and gaze-constrained conditions. Measures included pupil diameter, heart rate, and skin conductance.
• Study 2 – Social Orientation Task: Children viewed emotional faces preceded by fixation cues directing gaze to the eyes or mouth region of the presented faces, enabling measurement of gaze orientation, avoidance, and latency. Measures included gaze behavior, heart rate variability (HRV), and skin conductance level (SCL).
Methodological advances included the development of custom Python pipelines and scripts for automated areas of interest extraction, multimodal signal synchronization, pupil preprocessing, electrodermal activity filtering, and heart rate extraction.
3. Results
Across the studies we found, that:
(1) Pupillary contagion aligns with changes in the physiological arousal, co-occurring with heart rate differences. This supports the interpretation of pupillary contagion as an arousal-based mechanism rather than a perceptual artifact.
(2) Social anxiety did not account for observed attentional or physiological patterns, indicating mechanisms specific to autistic traits rather than general anxiety. The dimensional analysis further clarified this separation by modeling trait contributions on continuous scales.
(3) Across the two reports, we found that in treating autistic traits dimensionally rather than as categorical diagnostic categories allowed us to detect subtle, graded associations between social attention and physiological responses.
Overall conclusion: Reduced eye contact in autism appears linked to heightened social arousal rather than diminished social motivation.
4. Planned future research questions
Currently we aim to explore in more detail the temporal changes that occur in pupil reactions. To be more exact, we hope to further extract metrics such as peak dilation, latency to peak dilation and area under baseline-corrected curve (AUC) and correlate those metrics with reported sensory sensitivity scores. Examining these aspects will help to reveal physiological patterns that may help identify meaningful subgroups of children with various levels of autistic traits.
5. Dissemination and Collaboration
Findings from these projects were presented at academic seminars, national meetings, and interdisciplinary groups. Specifically, as a talk at ESSENCE conference, South Africa April 2023 and European Conference on psychological theory and research on Intellectual and Developmental Disabilities (ECIDD) in Göteborg June 2025. In addition, this data was presented at various seminars both at Gillberg Neuropsychiatry Center and Department of Applied IT at Göteborg University. Major collaborators included the University of Gothenburg, Stockholm University, and Massachusetts General Hospital / Harvard Medical School.
The project investigated mechanisms underlying children’s social attention and autonomic responses, focusing on pupillary contagion, eye contact, and physiological reactivity in relation to autistic traits. A total of 67 children were tested across two studies, including participants who received an autism diagnosis, those who indicated elevated autistic traits but have not been formally diagnosed, and neurotypical controls. Analytical samples varied by modality due to strict signal quality requirements. The project progressed in accordance with the planned methodological and scientific aims and no major deviations from initial protocol were performed.
2. Implementation
Data collection was conducted in a controlled laboratory environment (Gillberg Neuropsychiatry Center in Göteborg, Sweden) using two complementary paradigms:
• Study 1 – Pupillary Contagion Task: Children viewed photographs of peers with manipulated pupil sizes (large vs. small) under free-viewing and gaze-constrained conditions. Measures included pupil diameter, heart rate, and skin conductance.
• Study 2 – Social Orientation Task: Children viewed emotional faces preceded by fixation cues directing gaze to the eyes or mouth region of the presented faces, enabling measurement of gaze orientation, avoidance, and latency. Measures included gaze behavior, heart rate variability (HRV), and skin conductance level (SCL).
Methodological advances included the development of custom Python pipelines and scripts for automated areas of interest extraction, multimodal signal synchronization, pupil preprocessing, electrodermal activity filtering, and heart rate extraction.
3. Results
Across the studies we found, that:
(1) Pupillary contagion aligns with changes in the physiological arousal, co-occurring with heart rate differences. This supports the interpretation of pupillary contagion as an arousal-based mechanism rather than a perceptual artifact.
(2) Social anxiety did not account for observed attentional or physiological patterns, indicating mechanisms specific to autistic traits rather than general anxiety. The dimensional analysis further clarified this separation by modeling trait contributions on continuous scales.
(3) Across the two reports, we found that in treating autistic traits dimensionally rather than as categorical diagnostic categories allowed us to detect subtle, graded associations between social attention and physiological responses.
Overall conclusion: Reduced eye contact in autism appears linked to heightened social arousal rather than diminished social motivation.
4. Planned future research questions
Currently we aim to explore in more detail the temporal changes that occur in pupil reactions. To be more exact, we hope to further extract metrics such as peak dilation, latency to peak dilation and area under baseline-corrected curve (AUC) and correlate those metrics with reported sensory sensitivity scores. Examining these aspects will help to reveal physiological patterns that may help identify meaningful subgroups of children with various levels of autistic traits.
5. Dissemination and Collaboration
Findings from these projects were presented at academic seminars, national meetings, and interdisciplinary groups. Specifically, as a talk at ESSENCE conference, South Africa April 2023 and European Conference on psychological theory and research on Intellectual and Developmental Disabilities (ECIDD) in Göteborg June 2025. In addition, this data was presented at various seminars both at Gillberg Neuropsychiatry Center and Department of Applied IT at Göteborg University. Major collaborators included the University of Gothenburg, Stockholm University, and Massachusetts General Hospital / Harvard Medical School.