A physiologically informed virtual reality based social communication system for individuals with autism.

Lahiri et al. (2015) built a VR conversation game for autistic teens. The virtual human talked back. Two versions were tested in the same kids on different days.

One version watched only right or wrong answers. The other also watched the teen’s eyes and pupil size. The game got harder when pupils showed boredom and easier when eyes stayed on the speaker.

Kids gave more correct answers and looked at the virtual face longer when the game used both eye and pupil data. Performance-only feedback helped a little, but the bio-smart version helped more.

Moon et al. (2024) took the idea further. They kept the real-time tuning but swapped the body signal. Instead of watching pupils, their VR watched facial emotion and popped up hints when the child looked confused. The newer study found gains too, so the field is moving from bio signals to emotion signals.

Tao et al. (2025) pooled 26 VR work-skill studies and found a medium boost across the board. Uttama’s teen conversation game is one of the early bricks in that wall.

van der Miesen et al. (2024) also mixed VR with body read-outs, but used cortisol drops to judge if VR calmed kids at the dentist. Same tools, different goals—showing physiology plus VR is now a wide road, not a one-off trail.

You already shape tasks based on performance. This paper says add quick body clues—eye gaze, pupil change, even heart rate if you have the tech. When the learner’s eyes drift or pupils shrink, simplify or add novelty on the spot. Most VR social labs do this now; you can copy the logic in Zoom or in-person tasks with simple gaze tracking apps. Try it during conversation drills and watch engagement rise.