Characteristics of Visual Fixation in Chinese Children with Autism During Face-to-Face Conversations.

Zhao et al. (2023) filmed 40 Chinese children while they talked with an adult. Half had autism, half were typical.

The team used free software called OpenFace to map where each child looked on the adult’s face. They counted seconds spent on eyes, mouth, whole face, and background.

Kids sat 60 cm away and chatted for two minutes while a webcam tracked their eyes.

Children with autism looked at the mouth 30 % less and at the whole face 25 % less than peers.

They also spent more time staring at the wall or floor behind the speaker.

All differences were large enough to matter in daily teaching.

The result backs up Avni et al. (2020), who saw odd gaze paths in kids watching social videos. Both studies say “autistic eyes move differently,” one in live talk, one on film.

Waldron et al. (2023) show the same reduced face gaze is even stronger in non-syndromic autism than in fragile-X, so the signal is clear across diagnoses.

Two papers seem to clash but don’t. Wang et al. (2023) got more face gaze from autistic kids when emotions grew from weak to strong on video. The difference is task: edited video clips can pull gaze, real faces in free chat do not. Likewise, Lemons et al. (2015) found no eye avoidance in preschoolers using puppets; Zhong used older kids in live talk—age and setting explain the gap.

If you run social-skills groups, expect learners to miss mouth cues and look away. Put pictures or text near your mouth, slow your speech, and give clear “look here” prompts. Check understanding often, because reduced face gaze means they may gather less information from expressions.