Magnetic resonance imaging (MRI) is the method of choice for imaging disorders of the temporomandibular joint (TMJ) in patients with temporomandibular disorders (TMD) and pain. But the method is expensive, and wait times for MRI exams of the TMJ are long. So an evaluation of the diagnostic and therapeutic efficacy of MRI of the TMJ is valuable when estimating the imaging method’s use in treatment decisions and prognosis. To reduce cost to society, individual selection criteria for MRI of the TMJ need to be defined. The aim must be to minimize unnecessary examinations and to restrict use of MRI to patients who will benefit from the examination. To evaluate evidence for the efficacy of MRI in patients with TMD and pain, a systematic review of the literature was done (article I). The search yielded 494 titles, of which 22 were relevant. No publication had a high level of evidence; 12 and 10 publications had moderate and low levels of evidence, respectively. The evidence grade for diagnostic efficacy expressed as sensitivity, specificity, and predictive values was insufficient. No publication on diagnostic thinking efficacy or therapeutic efficacy met the inclusion criteria. The TMJs of 60 TMD patients were examined with MRI (article II). The patients were placed in one of two clinical diagnostic groups—(1) myofascial pain or (2) arthralgia/osteoarthritis—per the Research Diagnostic Criteria for TMD (RDC/TMD), a diagnostic classification system. This classification is widely used internationally, but further validation and development is needed. The most common MRI findings were various kinds of disc displacements; structural bone changes; and joint fluid, which occurred in both pain groups. The MRI findings did not support the RDC/TMD clinical diagnoses. In a randomized controlled trial of 80 patients suffering from TMD pain (article III), the short-term efficacy of a resilient appliance was compared with a control appliance consisting of a non-occlusal hard acrylic palatal appliance. After 10 weeks of treatment, 61% in the treatment group and 46% in the control group had at least a 30% reduction in TMD pain. Differences between groups were nonsignificant. Changes in condyle and disc position could be factors that are affected by appliance therapy and influence treatment outcome of TMD patients. MRI of the TMJ was performed in 48 of the 80 patients (article IV). The results showed that treatment outcome was not related to MRI-determined changes in condyle and disc position.