Research on patient–ventilator interactions has largely focused on inspiratory events, with little attention paid to expiration. We sought to determine the importance of the timing and magnitude of expiratory muscle activity in causing patient–ventilator dyssynchrony. Our study was done with healthy subjects receiving pressure support in whom we induced airflow limitation with a Starling resistor. The timing and magnitude of expiratory muscle activity were obtained by wire electromyographic recording of the activity of the transversus abdominis muscle, and were compared with the cycling of the ventilator and inspiratory muscle activity as determined from a flow tracing and diaphragmatic electromyogram (EMG), respectively. Induction of airflow limitation produced significant phase differences in the cycling of the subjects' expiratory muscle group and that of the machine. Some inspiratory efforts failed to trigger the ventilator, owing in part to an increase in elastic recoil consequent to the commencement of expiratory efforts before the termination of mechanical inflation. A delay in relaxation of the expiratory muscles did not interfere with the success of subsequent inspiratory efforts to trigger the ventilator. We also investigated the accuracy of two approaches for distinguishing between the contributions of expiratory muscle activity and elastic recoil to intrinsic positive end-expiratory pressure (PEEPi): the expiratory increase in gastric pressure (Pga) correlated better with transversus abdominis electromyographic activity (r = 0.7 to 0.95) than did the early inspiratory decrease in Pga (r = 0.04 to 0.53). In conclusion, the continuation of mechanical inflation into neural expiration was associated with failure of the subsequent inspiratory attempt to trigger the ventilator.