A coordinated activation of upper airway and chest wall muscles may be crucial in maintaining airway patency and ventilation. The alae nasi (AN) and diaphragm (DIA) electromyograms (EMG) were recorded with surface electrodes in 17 unsedated healthy preterm infants during both active (AS) and quiet sleep (QS). Airflow was measured via a nasal mask pneumotachograph and integrated to obtain tidal volume. Studies were performed during inhalation of room air and mixtures of 2 and 4% CO2 in air. In room air, phasic AN EMG accompanied 45 +/- 7% of breaths during AS compared with 14 +/- 5% of breaths during QS (P less than 0.001); however, with inhalation of 4% CO2 the incidence of AN EMG increased to comparable levels in both sleep states. During room air breathing onset of AN EMG preceded that of the DIA EMG and inspiratory airflow by 41 +/- 8 ms (P less than 0.01) and 114 +/- 29 ms (P less than 0.05), respectively. Peak AN activity preceded peak DIA activity by 191 +/- 36 ms (P less than 0.01). Alteration in sleep state or increasing chemical drive did not significantly alter these temporal relationships. Nevertheless, with each increase in end-tidal CO2, peak DIA EMG and tidal volume increased while peak AN EMG only showed a consistent increase during 4% CO2 inhalation. We conclude that although there exists a mechanism that temporally coordinates AN and DIA activation, the amount of AN EMG activity with each breath is not clearly correlated with DIA activation, which may contribute to the high incidence of respiratory dysrhythmias in preterm neonates.