Masticatory muscle activity during sleep has been considered independent of variations in the structural characteristics of the dental occlusion. However, scientific evidence contradicting an occlusal causal role is missing. The purpose of this study was to test the null hypothesis that sleep bruxism (SB) is independent of the presence of occlusal interferences. A total of 17 healthy female subjects (mean [SD] age 24.9 [4.1] years) presenting with SB, and randomly divided into two groups, were evaluated after receiving either elimination of occlusal interferences (Test group) or elimination of sharp margins without change in occlusal contacts (Control group). Audio-video polysomnography (PSG) recordings were undertaken before and after treatment (mean [SD] duration 14.4 [3.8] months). Two subjects in each group (< 2 rhythmic masticatory muscle activity [RMMA]/h) were excluded for statistical analysis. During the total sleep time (TST), the Test group exhibited a higher reduction in frequency of episodes per hour than the Control group (p < 0.05). The reduction in duration of episodes was also higher in the Test group during the TST. The Test group presented an increase in sleep stage N3 (p < 0.05) at the final PSG when compared with the initial PSG, and a higher percentage of N3 (p < 0.05) at the final PSG when compared to the Control group. Elimination of occlusal interferences resulted in a significant reduction of masseter and temporal muscle activity during sleep in females presenting with SB. Based on these results, the null hypothesis is rejected. The hypothesis of occlusal interferences as a risk factor for SB is still patent.