Illusory resizing of the painful knee is analgesic in symptomatic knee osteoarthritis

Key points

1. 01Watching your own knee appear to stretch or shrink while a matching gentle touch was applied (a visuotactile illusion) lowered pain, but touch alone or vision alone did not.
2. 02Both visual and touch input together were needed for relief, pointing to multisensory brain processing as the likely mechanism.
3. 03Holding the illusion for 3 minutes kept the relief going but did not deepen it; repeating it 10 times in a session increased the effect.
4. 04The effect size in the knee (25%) was smaller than reported for the hand (45%) in earlier work.

How it was conducted

Design

Exploratory proof-of-concept study using a randomised-order within-participant comparison of illusions versus control conditions, with sustained and repeated illusion phases across three sessions

Participants

People with current knee pain and a clinical diagnosis of knee osteoarthritis (ACR criteria); average age 67.3 years, 9 of 12 female

Intervention

MIRAGE mediated-reality system showing a live video of the knee that visually stretched or shrank with a matching gentle calf traction or compression

Conditions

Eight conditions: congruent visuotactile stretch/shrink, vision-only, tactile-only, and incongruent visuotactile controls

Primary outcome

Knee pain intensity on a 0 to 100 numerical rating scale, measured before and after each condition

What they found

• Congruent visuotactile illusions reduced pain by an average of 7.8 points (95% CI [2.0 to 13.5]), a 25% reduction (within-condition t1,11 = 2.96, p = 0.013).
• Tactile-only (t1,11 = 1.45, p = 0.17) and vision-only (t1,11 = -0.71, p = 0.95) controls produced no significant pain change.
• Condition x Time interaction confirmed the illusion outperformed controls (F2,22 = 4.2, p = 0.028); no main effect of Condition (F2,22 = 0.93, p = 0.41) or Time (F1,11 = 4.7, p = 0.053).
• Congruent visuotactile did not differ from incongruent illusions when vision was matched (F1,11 = 0.34, p = 0.57) but did outperform them when only touch was matched (F1,11 = 5.29, p = 0.042).
• Repeated illusions in Session 3 produced an average pain decrease of 20 points (95% CI [6.9 to 33.1]), a 40% reduction (pre-illusion t1,6 = 3.5, p = 0.013; post-illusion t1,6 = 3.9, p = 0.008).
• Sustained 3-minute illusions held relief steady but did not add to it (Session 1 t1,10 = 0.52, p = 0.61; Session 3 t1,7 = -0.697, p = 0.51).

Limitations

• Very small sample (12 completed Session 1, with only 6 and 7 completing Sessions 2 and 3), powered only for the primary analysis.
• Proof-of-concept design used a single short (~30 second) illusion dose for the primary comparison.
• No formal ratings of body ownership or illusion vividness were collected during the Session 1 conditions.
• Daily pain scores were largely unchanged, so it is unclear whether the in-lab relief translates to everyday symptoms.

Why it matters

For patients

It suggests a drug-free, non-invasive illusion-based approach might ease knee osteoarthritis pain, though it is still experimental and not yet a treatment you can access.

For clinicians

It offers early support for multisensory body-illusion techniques as a pain-modulation tool in knee osteoarthritis, but the small sample means it should inform research interest rather than practice.

For readers

It adds to growing evidence that altering how the brain represents a painful body part can change the pain itself.