Apparatus for cpap using a neural network

A and treating for detecting during the sleeping respiration disorder methods, comprising for artificial condition networks, analysis the related nasal air flow the data. Is parallel with and a gas supply straw pressure (CPAP) system to the respiration loop for the patient is, the air flow of the return circuit by the cyclical sampling, and at calculates one or more of different a mode parameters; The parameter the cycle section is trained of the identification sleeping respiration disorder a mode the ANN, and analysis of the net; the CPAP and a gas supply in the through responding the output of ANN connected with the control. Also disclosed with the corresponding device.