We hypothesized that: the balance of electrical activities between the upper airway and chest wall inspiratory muscles affects upper airway inspiratory caliber, and at low levels of central respiratory neural efferent activity, an imbalance between the electrical activities of these 2 inspiratory muscle groups exists that results in a decreased upper airway caliber. These hypotheses were tested during periodic breathing induced by mild hypoxemia in NREM sleep in 9 healthy male subjects. In 6 subjects during periodic breathing as central respiratory neural activity decreased, the tonic and phasic EMG activity of the upper airway inspiratory muscles decreased at a rate greater than that of the chest wall EMG activity. When the ratio of upper airway to chest wall EMG activity decreased below a critical level, which was reproducible across subjects, upper airway inspiratory resistance increased hyperbolically. Resistance at peak inspiratory flow increased from 4.10 +/- 0.97 (mean +/- SEM) to 48.70 +/- 21.00 cmH2O/L/s as tidal volume decreased from 0.79 +/- 0.12 to 0.20 +/- 0.02 L during periodic breathing in these subjects. In the 3 remaining subjects, the ratio of the upper airway to chest wall EMG activity did not decrease below the critical level as the activity of both muscle groups decreased during periodic breathing, and upper airway resistance did not increase. We conclude that within the confines of this study the nonlinear activation of upper airway and chest wall inspiratory muscles contributed to fluctuations in upper airway resistance observed during periodic breathing in sleep.(ABSTRACT TRUNCATED AT 250 WORDS)