Configuration does not affect the allocation of visual attention to foreground and background information differently in autism on a change detection task

The visual processing of autistic people has been connected to a tendency to process local, or detailed, information more readily than the overall global, or configural, structure of visual information. This style has been highlighted in evidence of unique and efficient attention processing on certain tasks, such as the detection of changes. However, research on change detection in autism is mixed, with discrepancies seemingly attributed to the diversity of stimuli and the specific components of perception and attention involved in the task. We assessed rudimentary differences in change detection between autistic and non-autistic participants using a simple geometric task manipulating local/global processing and foreground/background attention. Thirty-two autistic adolescents and adults ( M age = 20.21, SD = 6.10) and 32 mental-age matched non-autistic adolescents and adults ( M age = 19.79, SD = 4.92) completed a change-detection task involving displays of rectangles presented in configural (global) and non-configural (local) arrangements. The participants were asked to indicate whether changes to the colours of these displays occurred. Three mixed-effect analyses of variance comparing accuracy, detection sensitivity, and response bias revealed a similar pattern of visual prioritization for both groups, with more accurate, sensitive and less biased change detection for foreground changes, especially non-configural ones. While accuracy levels were similar across groups, non-configural changes produced greater detection sensitivity for non-autistic than autistic participants. The findings reflect similar attentional patterns between autistic and non-autistic adolescents and young adults in change detection, even when issues of both configural/non-configural arrangement and foreground/background elements were considered. • All participants showed similar accuracy when detecting changes at the local and global level of static visual displays. • All participants were better at detecting changes in the foreground than the background of displays. • Participants showed a local processing advantage when detecting foreground changes in simple displays void of social content.