In temporomandibular disorder (TMD), the effects of standard interventions such as using an occlusal splint and its impact on pain relief and pain catastrophizing are poorly understood. Earlier work pointed to a crucial role of insula activation with changes in pain relief by occlusal splint treatment. We performed a functional imaging study using specially developed splint systems to allow for a placebo-controlled longitudinal design. Using functional MRI we examined 20 TMD patients during repetitive occlusal movements at baseline and over the course of splint therapy and also collected self-reported pain catastrophizing. For balancing performance between baseline and after intervention we used occlusion force measures in an individualized fMRI-splint system. Splint therapy lasted for approximately 7 weeks with one group selected by randomization wearing a palatine placebo splint over the first 3 weeks (delayed start; 11 individuals). As expected, fMRI activation in areas involved in pain processing (insula, primary and secondary somatosensory cortex) decreased with intervention. At baseline a positive correlation between activation of the left anterior insula and pain catastrophizing was present. Both parameters decreased over intervention while associations were primarily observable for patients with rather mild TMD.