Hybrid Modeling of Cardiopulmonary System, HRV and Consciousness

Processing of Heart Rate Variability (HRV) signal becomes more and more important for functional assessment of patient’s health. When developing Hybrid Models of cardiopulmonary system we need to make possible taking into account heart rate variability. Our aim is to generate signals that have properties similar to HRV in norm and in different pathologies. Hybrid models of cardiopulmonary system with features including generation of HRV and new signal processing algorithms may be very helpful in medicine for noninvasive functional assessment. We decided to use Elman’s neural networks to generate HRV signals off-line and we are working on application of this method on-line. To evaluate if the generated signal is sufficiently good to model HRV we compare its Higuchi’s fractal dimension with that of the signal used to teach the network. While generating HRV of a person suffering with epilepsy we observed significant decrease of HRV-signal’s fractal dimension that is accompanying epileptic seizure.