Automated detection of respiratory phenotypes during sleep

Abstract

We hypothesized that normal individuals can be distinguished on the basis of specific respiratory phenotypes during sleep that determine one's susceptibility to the development of obstructive sleep apnea. To test this hypothesis, methods have been developed to 1) rapidly assess baseline upper airway and respiratory timing characteristics and compensatory responses in these parameters to experimentally induced upper airway obstruction during sleep in normal individuals, and 2) automate the detection of the respiratory pattern and the presence of flow limitation during these experimentally induced upper airway obstructions. We demonstrated that: (1) marked variability in the responses to upper airway obstruction exist among normal individuals, and that (2) upper airway collapsibility (critical pressure) is modulated dynamically by a number of reflex neuromuscular responses triggered by breathing through an obstructed upper airway. Moreover, (3) our methods allowed us to assess the relative strength of compensatory neuromuscular responses to upper airway obstruction in normal individuals. Our findings indicate that the normal individual's susceptibility to the development of upper airway obstruction during sleep is determined by the passive upper airway properties and/or by a lack of vigorous compensatory neuromuscular responses.