A.A. Mikheev, V.N. Loktukhin, O.V. Melnik

The possibility of personalization of physical therapy procedures by individual selection of the therapeutic effect parameters based on the rapid assessment to current functional status of the patient is analyzed. To improve the reliability of assessment of the current functional status the process of diagnostic conclusions forming presented in the form of linked stages of selection of informative parameters of the original signal, identification of informative features that serve as markers of changes in the functional status of the body, and making decisions based on neural network technology. The ways to improve the reliability of obtaining the source for the formation of the diagnostic conclusion of the data, in particular, the amplitude-time parameters of elektrocardiosignal is demonstrated. The methods for selection of the additive low-frequency noise ensuring minimal loss of elektrokardiosignal informative parts are proposed. The requirements to be met by methods of construction of informative signs and an approach to developing such methods, based on the use as a diagnostic sign of integral parameters in the form of spectral coefficients in a special orthogonal basis. Developed and investigated methods to identify informative shape parameters elektrocardiosignal elements in their spectral composition of the bases of the Legendre polynomials and the Walsh functions. We propose a three-tier neuro-bionic structure of the system forming diagnostic conclusions, using the results of the exact "tool" of control and interpretation of diagnostically significant parameters of biomedical signals and received from them informative features as input to the device neuro-fuzzy decision support. The concept of building subsystems for decision-making support based on fuzzy logic.