N.A. Al-Khulaidi, R.V. Isakov, L.T. Sushkova

The use of neural networks in medicine usually associated with integral or differential diagnosis systems. One of the main applications of neural networks is the diagno sis of heart disease. In the present, the analysis of heart rate variability (HRV) is considered as the most informative and noninvasive method for quantitative evaluation of organism functional state. For this reason, this method is often used in scientific research. The purpose of this research is consideration of questions of neuron networks application expedience and features in the automated determination of changes in variability of heart. Comparison of multi-layered perceptron and modular construction of neuron network organization was in-process produced, as variants of the automated analysis of heart rate variability. Two databases were formed: base of histogram and scattergram. Researches rotined that enhanceable sensitiveness to pathologies, a low error and possibility of analyzable pathologies number unlimited increase does a modular construction, probably, by an optimum choice for the decision of hart rate variability.