Biological Motion Perception in Autism Spectrum Disorder: A Meta-Analysis

Background: The motion of others – biological motion, conveys to the observer information that allows the identification of both affective states and intentions. As biological motion plays an important part in social functioning, its investigation in autism spectrum disorder (ASD) has received more attention in recent years. However, the findings tend to be mixed as some studies suggest no difference between neurotypically developing (NT) individuals and ASD individuals, whilst others show worse performance in the ASD group. A few questions stem from the existing literature: Are individuals with ASD able to perceive and interpret biological motion and what factors can explain the variability in the data?

Objectives: We aimed to create a quantitative summary of previous findings and investigate potential factors, which could lead to the variable and often contradictory results in the field.

Methods: We included papers, which compared ASD and NT individuals, and focused on biological motion perception (detection, action and emotion perception). An electronic search from Dissertations & Theses A&I, Dissertation & Theses: UK & Ireland, Web of Science, PsycINFO and MEDLINE yielded 37 eligible papers. We included behavioural, eye-tracking, EEG and fMRI studies. The overall sample size across categories was 730 autistic individuals and 932 neurotypical individuals across three age groups (children, adolescents and adults). In all included studies, the face of the stimuli was not visible. We used a three-level random effects meta-analytic approach to analyse the 34 papers eligible for a quantitative summarisation.

Results: Overall results suggest a general difficulty for individuals with ASD in perceiving and interpretuing biological motion (g=0.7287, 95% CIs: 0.4906, 0.9669). However, results suggest that greater differences occur when higher order information, such as emotion is required (g=1.2840, 95% CIs: 0.8388, 1.7293). Additionally, age also influences the size of the difference between the two populations. The difference between ASD and NT children was significantly greater than the ASD-NT difference in adults in action recognition paradigms (g=0.8362, 95% CIs: 0.1545, 1.5179). When emotion understanding is tested, the ASD-NT difference in children was not significantly different from the difference in adults (g=-0.7404, 95% CIs: -1.6191, 0.1382). No effect of either gender or IQ was found on the results.

Conclusions: To summarise, there appears to be a general difficulty in ASD in perceiving and interpreting biological motion. However, the size of the effect appears clearly modulated by both age and the type of paradigm. The results suggest a delayed development in the skills required to interpret biological motion and the largest effects are found when emotional understanding is required. Most importantly, clear distinctions need to be made between the age groups and the paradigms utilised, when trying to interpret differences between these two populations.