This prospective study correlated the results of magnetic resonance (MR) imaging, arthrography, and arthroscopy in 27 patients with clinically suspected temporomandibular joint (TMJ) dysfunction. Open surgical confirmation was available in 12 TMJs. The variables assessed with each modality were disk position and morphology, disk perforation, joint adhesions, and degenerative joint disease. MR imaging was superior to dual joint arthrography with arthrotomography in delineation of the disk position. MR imaging accurately demonstrated the disk position in 11 joints (92%); arthrography was accurate in nine of the 12 joints (75%) with surgically confirmed dysfunction. Disk perforations and joint adhesions were not demonstrated with MR imaging. Arthroscopy was superior to arthrography and MR imaging in the detection of disk perforations, joint adhesions, and early degenerative changes along the articular eminence and glenoid fossa. The authors believe MR imaging should constitute the first line of imaging because it provides accurate information about disk position in a noninvasive manner. In most cases, information obtained with MR imaging is adequate in deciding whether to institute conservative or surgical management; in difficult cases, triple correlation may be needed.