To test the hypothesis that the pain-producing effect of parafunctional clenching is mediated by oral contraceptive use and estrogen levels, eight premenopausal women participated in daily (five days/week) 20-minute-long EMG biofeedback training sessions (on the left and right temporalis and masseter muscles) structured as a two-phase cross-over study. Four subjects used oral contraceptives, and four did not. Subjects were instructed to maintain temporalis and masseter activity below 2 microV during decrease training and above 10 microV during increase training. All subjects began their participation at the start of menses. The initial week of training was followed by a week of rest and then a second week of training at mid-cycle. Preliminary screening examinations showed that none of the subjects had TMD. One subject was diagnosed with TMD pain during increase training, and no subjects were diagnosed with TMD during decrease training. Self-reported pain following training increased significantly during increase training. No effects on pain were observed for oral contraceptive status. We conclude that chronic, low-level parafunctional clenching may be a factor producing temporomandibular disorder pain and that oral contraceptive status does not play a role in the TMD pain produced by the experimental protocol.