Conducting head-mounted eye-tracking research with young children with autism and children with increased likelihood of later autism diagnosis

**Background:** Head-mounted eye-tracking allows researchers to study millisecond-level gaze behaviors during naturalistic social interactions, offering insights into early attentional development. While this method has been used extensively with typically-developing children, its application to children with autism has been limited due to sensory sensitivities, social-communication difficulties, and challenging behaviors. This paper aims to share practical strategies for successfully collecting head-mounted eye-tracking data from young children with autism and those at increased likelihood of later autism diagnosis, including minimally verbal or nonspeaking children and those with intellectual disabilities.

**Methods:** The study included 41 children with autism (aged 1.3–8.9 years) and 17 children at increased likelihood of later autism diagnosis (aged 0.5–4.5 years). All children with autism met diagnostic criteria on the ADOS or ADOS-2. Participants completed a 5-minute parent-child play session with a standardized set of eight toys while wearing a head-mounted eye-tracker (Positive Sciences, Inc., 51 g). The procedure included pre-visit preparation (e.g., providing pictures, mock devices, and hat acclimation), a child-friendly waiting room, individualized breaks, and calibration using nine targets. Researchers used behavioral strategies such as modeling, positive reinforcement, and offering choices. Families could split visits and were encouraged to bring preferred reinforcers. Challenging behaviors were recorded, and success rates were calculated.

**Key Results:** The overall success rate for head-mounted eye-tracking data collection was 92.68% among children with autism (86.36% for 1–4-year-olds, 100.00% for 5–8-year-olds) and 94.11% among children at increased likelihood of later autism diagnosis. Three of 41 autistic children (all under 5 years) and 1 of 17 at-risk children failed to complete the session. Challenging behaviors were common: 88.8% of younger autistic children (1–4 years) and 29.4% of older autistic children (5–8 years) exhibited at least one challenging behavior, including taking off the device, elopement, and distress. On average, each child contributed 8,757 frames (SD = 689) per camera, approximately 90.37% of the play session. No substantial demographic or clinical differences were found between completers and non-completers.

**Clinical Implications:** The high success rates demonstrate that head-mounted eye-tracking is feasible with young children across the autism spectrum, including those with significant communication and behavioral challenges. The detailed strategies—such as desensitization, parent collaboration, flexible scheduling, and individualized reinforcers—can be adapted for other wearable technologies (e.g., EEG, heart-rate monitors) and other developmental disabilities. This methodology can reveal early gaze behaviors relevant for autism screening, diagnosis, and intervention, potentially improving early identification and reducing disparities in access to services. The authors emphasize the importance of "flexibility within fidelity" to accommodate individual needs while maintaining protocol integrity.