Distinct Attention–Behavior Pathways in Neurodiverse Users: An Eye-Tracking Study of Visual Attention to Mobile AR Interfaces

Augmented reality (AR) is increasingly adopted in digital environments to support multimodal experiences. However, limited empirical evidence exists regarding how users with diverse neurocognitive profiles allocate attention within AR interfaces. This study investigates attention and interaction behavior in autistic and typically developing (TD) children (N = 49) using a mobile AR interface. Eye-tracking metrics, including percentage time fixated (PTF) and fixation transitions (FT), were combined with video-based behavioral analysis. TD children demonstrated greater attention to textual instructions and more frequent transitions between interface elements, whereas autistic children exhibited more localized visual attention and fewer transitions. The overall group difference in attention allocation remained significant after demographic and cognitive adjustment, although some AOI-specific contrasts were attenuated. No significant group difference in task performance was detected. These findings suggest distinct attention–behavior pathways, although statistical equivalence in task performance was not established, highlighting the importance of neuroinclusive AR interfaces accommodating diverse attentional strategies.