Adults with autism are less proficient in identifying biological motion actions portrayed with point-light displays.

E-Shire et al. (2019) asked adults with and without autism to name simple actions shown only as moving dots. The dots formed point-light displays—like a walking person made of Christmas lights. The adults pressed keys to say what action they saw while speed and accuracy were timed.

Adults with autism were slower and made more errors than neurotypical adults. The gap was small but steady across every action tested. Even simple moves like waving were harder to catch.

Krüger et al. (2018) ran a near-copy task and got the same result: adults with autism rated happy point-light dances as less intense and felt less sure of their answers. Together, the two papers show motion-reading trouble is real for both plain actions and emotions.

de Korte et al. (2021) flips the picture. They found autistic adults can match neurotypicals on gesture recognition accuracy, but they move and scan the gestures differently. So accuracy can look fine while the reading style stays odd—extending the story beyond simple error counts.

Shire et al. (2019) seems to disagree: autistic adults recognized emotions from film clips just as well as controls. The trick is the clips gave faces, voices, and context, not just motion dots. When only sparse motion cues exist, autistic adults slip; when rich scenes exist, they keep up.

When you show video models or read body language in sessions, know that adults with autism may need more time or clearer cues. Slow the playback, add narration, or freeze key frames. Skip tiny point-light demos unless you check understanding—because sparse motion alone can fail.