To determine whether nocturnal periodic breathing (PB) at altitude is due primarily to unstable control of ventilation or the inability to maintain stable sleep states, we performed visual and computer analyses of the electroencephalographic and respiratory records of healthy volunteers at simulated altitudes of 4572, 6100 and 7620 m. Transient arousals were associated with < 52% of the apneas identified; thus, the PB cycle was not always associated with transient arousal. Following the termination of oxygen breathing, the reinitiation of PB was not dependent on the occurrence of arousal as the primary event. The transition from apnea to breathing preceded the appearance of arousal by approximately 1 to 4 sec. Ventilatory drive in the breaths immediately following arousal was significantly larger than corresponding control breaths, matched for SaO2. Our findings suggest that altitude-induced PB is unlikely to result from primary fluctuations in state. Arousals promote the development of PB with apnea and help to sustain these episodes, but are not necessary for their initiation.