The present study evaluated the effect of co-activation of tongue protrusor and retractor on pharyngeal patency in patients with obstructive sleep apnea. The effect of genioglossus (GG), hyoglossus (HG) and co-activation of both on nasal pressure (Pn):flow relationships was evaluated in a sleep study (SlS, n=7) and during a Propofol anesthesia study (AnS, n=7). GG was stimulated with sublingual surface electrodes in SlS, and with intramuscular electrodes in AnS, while HG was stimulated with surface electrodes in both groups. In AnS, the cross-sectional area (CSA):Pn relationships was measured with a pharyngoscope to estimate velopharyngeal compliance. In SlS, surface stimulation of GG had no effect on the critical pressure (Pcrit), HG increased Pcrit from 2.8±1.7 to 3.7±1.6 cmH₂O, but co-activation lowered Pcrit to 0.2±1.9 cmH2O (p<0.01 for both). In AnS, intramuscular stimulation of GG lowered Pcrit from 2.6±1.3 to 1.0±2.8 cmH₂O, HG increased Pcrit to 6.2±2.5 cmH₂O (p<0.01), and co-activation had a similar effect to that of GG (Pcrit = 1.2±2.4 cmH₂O, p<0.05). None of the interventions affected significantly velopharyngeal compliance. We conclude that the beneficial effect of co-activation depends on the pattern of GG fiber recruitment: although surface stimulation of GG failed to protrude the tongue, it prevented the occlusive effect of the retractor, thereby improving pharyngeal patency during co-activation. Stimulation of deeper GG fibers with intra-muscular electrodes enlarged the pharynx, and co-activation had no additive effect.