The velopharynx is the most collapsible segment of the upper airway in patients with obstructive sleep apnea. However, we do not know if velopharyngeal compliance is uniform throughout its length, or if compliance is modified by contraction of upper airway muscles. We tested the hypothesis that rostral and caudal velopharyngeal (VP) compliance differs, and that tongue muscle contraction reduces compliance. High-resolution MR images of the VP were made at nasopharyngeal pressures ranging from -9 to 9 cmH(2)O in anesthetized rats. Images were obtained twice at each pressure, once with and once without bilateral hypoglossal nerve stimulation. The volume of the caudal and rostral VP was computed at each pressure. The caudal VP was significantly (P = 0.0058) more compliant than the rostral VP, but electrical stimulation of the tongue muscles did not change compliance. VP critical pressure (Pcrit; pressure at zero airway volume) averaged -25.2 and -12.1 cmH(2)O in the rostral and caudal VP, respectively (P < 0.0001). Coactivation of tongue protrudor and retractor muscles or contraction of protrudor muscles alone dilated the VP and made Pcrit more negative (P < 0.0001), but only in the caudal VP. In the rat, the caudal VP is more collapsible than the rostral VP, and either coactivation of tongue protrudor and retractor muscles or contraction of protrudor muscles alone makes this region more difficult to close. Thus, tongue muscle contraction protects the caudal VP, which appears to be a particularly vulnerable segment of the nasopharyngeal airway. With suitable modification, the methods described here, including tongue muscle stimulation at different pharyngeal pressures, may be appropriate for experiments in human subjects.