Positive-pressure ventilation (PPV) is widely used for treatment of acute cardiorespiratory failure, occasionally at the expense of compromised cardiac function and arterial blood pressure. The explanation why has largely rested on interpretation of intracardiac pressure changes. We evaluated the effect of PPV on the central circulation by studying cardiac chamber volumes with cardiac magnetic resonance imaging (CMR). We hypothesized that PPV lowers cardiac output (CO) mainly via the Frank-Starling relationship. In 18 healthy volunteers, cardiac chamber volumes and flow in aorta and the pulmonary artery were measured by CMR during PPV levels of 0, 10, and 20 cmH2O applied via a respirator and a face mask. All cardiac chamber volumes decreased in proportion to the level of PPV. Following 20-cmH2O PPV, the total diastolic and systolic cardiac volumes (±SE) decreased from 605 (±29) ml to 446 (±29) ml (P < 0.001) and from 265 (±17) ml to 212 (±16) ml (P < 0.001). Left ventricular stroke volume decreased by 27 (±4) ml/beat; heart rate increased by 7 (±2) beats/min; and CO decreased by 1.0 (±0.4) l/min (P < 0.001). From 0 to 20 cmH2O, right and left ventricular peak filling rates decreased by −146 (±32) and −187 (±64) ml/s (P < 0.05) but maximal emptying rates were unchanged. Cardiac filling and output decrease with increasing PPV in healthy volunteers. The decrease is seen even at low levels of PPV and should be taken into account when submitting patients to mechanical ventilation with positive pressures. The decrease in CO is fully explained by the Frank-Starling mechanism.