This general question has been intensely debated for years. The answer is complex, dependent upon many parameters, and will certainly vary with the specific disorder, i.e. diagnosis, being investigated. Below are the results of one study attempting to understand and clarify this question.
Wexler GB, McKinney MW, Temporomandibular Treatment Outcomes within Five Diagnostic Categories, Jour of Craniomandib Practice, (17)1: 30-37, 1999
This study of temporomandibular disorder (TMD) treatment outcomes examines 274 consecutive patients in five diagnostic categories and a 25 patient comparison group to determine relative levels of symptom improvement. Employing a psychometric outcome measure, the TMJ Scale, it was found that patients receiving active TMD treatments manifest statistically significant symptom improvements. Untreated patients reported minor and statistically insignificant symptom variations. Patients with intracapsular TM joint dysfunctions exhibited higher levels of improvement in pain and other TMD symptoms than patients presenting with primarily muscle symptoms. This research supports the hypothesis that TMDs are not self-limiting and require active treatment interventions. It is suggested that some studies cited to show that TMDs are self-limiting have major methodological limitations, relying upon unvalidated and subjective assessments of symptom levels. This research also outlines a procedure for TMD practitioners to measure treatment efficacy and the relative effectiveness of differing treatment modalities in a valid, consistent and unbiased manner.
More research studies can be found under Scientific Publications at www.tmjscale.com