Background and aims: Improvements in physical disability after chronic pain treatment have been associated with normalization of primary motor cortex (M1) thickness. Primary target for rTMS therapy is M1, while alternative targets related to pain processing, such as premotor cortex and secondary somatosensory cortex (S2) are also being studied. Here, we examined whether cortical thickness predicts the response to rTMS treatment. 
Methods: 24 patients with severe unilateral (15 left, 9 right side) persistent facial pain (aged 35 to 75 years) were enrolled. Patients underwent two five-day high-frequency rTMS treatment courses separated by six weeks targeted to M1 area of lower face. Responders were defined as the ones who continued with maintenance treatment. Anatomic MRI obtained at baseline was analysed using Freesurfer (v6.0.0) for whole brain cortical thickness between responders and non-responders. Age was entered as a nuisance covariate to the GLM model, and uncorrected independent t-test was done. 
Results:  In the left hemisphere, responders had thicker cortex in face M1 whereas non-responders had thicker cortex in face S1. In the right hemisphere, responders had thicker cortex in parietal operculum while non-responders had thicker cortex in the posterior insula. Additionally, non-responders had thicker cortex in prefrontal area. 
Conclusions: Bilateral differences in cortical thickness were found between responders and non-responders. Structural differences may explain why only some patients benefit from rTMS treatment. Further studies are warranted to see if cortical thickness changes are similarly predictive in other chronic pain conditions and if they can be used as guidance for rTMS target selection.

CORTICAL THICKNESS AS A PREDICTOR FOR RTMS TREATMENT OUTCOME IN ATYPICAL FACIAL PAIN