RECOGNITION OF RESPIRATORY PATTERNS AND S LEEP STAGES WITH NEURAL NETWORKS IN INFANTS

During polysomnographic examinations the physiological state of a pati ent is monitored. Different signals regarding breathing and blood gase s as well as some electrophysiological signals are recorded continuous ly up to ten hours. The time required for a manual analysis by physici ans is considerable, but can be reduced by computerised analysis proce dures. This work describes the automatic recognition of breathing dist urbances, especially the cessation of breathing and the distinguishing of the course of sleep in different sleep states with neural networks .