The breathing patterns from sleeping persons suffering from sleep apnea hav
e been measured. A method based on the neural network-like O-algorithm has
been applied to capture the onset of sleep apnea. This method is suggested
as an indicator for early on-line detection of obstructions in the upper ai
rway. Results from the system tested with airflow signals recorded from fiv
e patients during sleep indicate acceptable performance and treatment for d
eveloping apnea is possible. (C) 2000 Elsevier Science B.V. All rights rese
rved.