University of East Anglia researchers have found that small changes in pupil size and blink patterns can indicate mental workload during virtual reality tasks. Combining AI with eye tracking in VR may eventually support personalised treatment for chronic pain.
Lead researcher Dr Jordan Tsigarides said chronic pain affects an estimated 1.5 billion people worldwide and is a leading cause of disability. The work forms part of his PhD exploring AI and sensor technology to make VR treatments more responsive, aiming to balance challenge and comfort.
“Chronic pain affects an estimated 1.5 billion people worldwide and is one of the leading causes of disability. Virtual reality has enormous therapeutic potential – whether that’s guiding patients through a peaceful woodland, supporting relaxation exercises or engaging them with games and puzzles that help manage pain. It can help reduce chronic pain by immersing users in engaging, interactive environments that capture the brain’s attention and draw it away from pain, reduce pain-related stress and anxiety, and may even help retrain how the brain processes pain over time.
“But most virtual reality programs offer the same experience to everyone, even though people’s pain and mental effort can fluctuate from moment to moment. Our work suggests future systems could read mental effort through the eyes and adapt therapy in real time to each individual. If a headset can detect when someone is becoming mentally overloaded, it could potentially adapt the experience automatically to keep them engaged and maximise therapeutic benefit.”
In a study of 84 volunteers (42 healthy controls and 42 with chronic musculoskeletal pain such as arthritis or fibromyalgia), participants completed escalating memory (N-back) challenges while wearing a VR headset with eye tracking. As cognitive demand rose, pupils widened and blink rates increased.
AI models analysing temporal patterns distinguished low versus high mental workload with up to 88% accuracy in healthy volunteers. Tasks were well tolerated, with very low reported motion sickness or discomfort. The findings support feasibility of VR-embedded eye tracking for cognitive-load estimation; reliable individual-level inference in chronic pain will require participant-calibrated models and further validation. The researchers hope this could enable wearable systems that adapt therapy in real time.
Dr Tsigarides said: “These results suggest a future in which a headset could continuously monitor the brain’s workload through the eyes, automatically fine-tuning treatment to each individual user.
“Our earlier research showed promising pain reductions in people with fibromyalgia using therapeutic VR. Now we’re exploring how to make those experiences smarter, more responsive, and tailored to each person.
“Until now, most VR has been one-size-fits-all. Our findings could help make treatments more personalised and responsive. This could enhance future VR treatments, not just for pain but potentially for conditions like anxiety or depression.”
“In future, the same principles could be explored in education and professional training, where virtual learning experiences might adapt to the learner’s mental workload.”
However, the study also highlighted a challenge. While AI models performed strongly in healthy volunteers, accuracy was significantly lower among people living with chronic pain, suggesting that standardised approaches may not work equally well for all patients.
The researchers believe future systems may need to be individually calibrated, learning how each person’s eyes respond to mental effort before making treatment decisions.
And they say that larger studies are needed to determine whether adaptive VR therapies can improve real-world outcomes for patients.
Source: University of East Anglia
