Evaluating virtual reality therapy for phantom limb pain: functional connectivity and pain outcomes

Poster Session F - Tuesday, March 10, 2026, 8:00 – 10:00 am PDT, Fairview/Kitsilano Ballrooms

Catherine Scanlon¹ (cs3925@drexel.edu), Apoorva Kelkar¹, Elisabetta Ambron^2,3, Harrison Stoll¹, Laurel Buxbaum², Branch Coslett³, John D. Medaglia^1,3; ¹Drexel University, Philadelphia, PA, ²Thomas Jefferson University, Philadelphia, PA, ³University of Pennsylvania, Philadelphia, PA

Phantom limb pain (PLP) is a common disorder among amputees yet remains poorly understood. Existing theories emphasize somatomotor cortical plasticity without considering broader pain-related networks which may influence treatment outcomes. Virtual reality (VR) is a promising treatment for PLP, but it remains unclear whether its analgesic effects are related to PLP-specific neural mechanisms. We collected 7T fMRI data from 11 lower-limb amputees undergoing either active or distractor VR treatment. Pain was assessed pre- and post-treatment using the McGill Visual Analog Scale Pain Questionnaire, and functional connectivity was calculated as the median correlation among sensorimotor, insular, cingulate, and thalamic regions. We hypothesized that functional connectivity would positively predict pain across timepoints and that pain would decrease more in the active VR group than in distractor VR. A bivariate linear regression model showed that functional connectivity did not significantly predict pain score at both timepoints, β=-0.303,t(94)=-1.344,p=.182. To test whether active VR reduced pain relative to the distractor condition, we fit a linear regression including treatment, timepoint, and their interaction (R²=.32,F(3,92)=14.23,p< .001). Pain significantly decreased over time in both groups, β=-2.25,t(92)=-4.641,p<.001, with limited evidence for a difference between treatment groups, β=1.125,t(92)=1.895,p=.061. The interaction effect was also near significant, β=1.5,t(92)=1.786,p=.077. Overall, PLP decreased following VR intervention, but active treatment was not significantly different from the distractor. While pain was not significantly related to functional connectivity, this preliminary sample is underpowered for neuroimaging analyses. Future work in larger samples will better characterize the relationship between pain, functional connectivity, and VR treatment effects with PLP.

Topic Area: NEUROANATOMY