The research and clinical literature on biofeedback treatment of temporomandibular joint (TMJ) dysfunction is devoid of normative or comparative electromyographic (EMG) studies examining muscle activity in either patient or normal samples. For the present study, resting EMG levels for each masseter and temporalis were obtained from three groups of subjects: asymptomatic (female, N = 24, mean age = 26.4); subclinical (female, N = 31, mean age = 28.6); and patient (N = 61, female 70%, mean age = 31.9). A Biocomp 2001 biofeedback system was used to gather the EMG data from each of the four sites during a six- to eight-minute resting baseline period. The patient group demonstrated significantly higher EMG activity than the asymptomatic or subclinical groups for all variables except the right masseter (F (8,220) = 6.65, p less than 0.001). The temporalis was found to be the site of greatest EMG activity more frequently than the masseter. These findings strengthen diagnostic and assessment procedures and criteria, as well as suggest alternate treatment and research protocols.