M.Yu. Kostenkova1

1 State budgetary institution of health care “Regional Psychiatric Hospital n.a. K.R. Yevgrafov” (Penza, Russia)
1 marikost83@mail.ru

The development of personalized psychiatry is faced with the choice of therapy tactics in the presence of concomitant diseases in a patient identified by instrumental, laboratory and other diagnostic methods; the search for diagnostically significant combinations of blood parameters in mental pathologies; the definition of clear biomarkers in blood tests that would indicate a specific mental disorder. The main problem is the complexity of clinical interpretation by a doctor of extensive heterogeneous diagnostic data and the development of integrated analytical systems that combine this information. Development of a system for visual presentation of diagnostic data for its prompt interpretation based on patient blood parameter data. Creation of a mathematical model of informative blood parameters using cognitive graphics for data visualization and analysis. Definition of the feature space of "Norms" and "Pathologies" of blood parameters to predict the patient's condition and identify hidden relationships between the parameters. A mathematical model of informative blood parameters has been developed, on the basis of which a system for visualizing multidimensional data in the form of a cognitive image "Heat Map" has been created. Diagnostically significant feature spaces have been defined, allowing to differentiate normal and pathological conditions for subsequent application of machine learning algorithms. The developed mathematical model for blood parameter analysis served as the basis for creating the Medical Data Management System "MedData" and provided comprehensive processing of clinical information. The use of the cognitive graphics method made it possible to clearly visualize heterogeneous data, facilitating the diagnostic process for clinicians. The formed "norm-pathology" feature space formed the basis for finalizing machine learning algorithms aimed at identifying diagnostically significant biomarkers. The study makes a significant contribution to the digitalization of psychiatry, ensuring the transition from subjective assessments to evidence-based diagnostic methods.

Kostenkova M.Yu. Mathematical modeling and color coding for visual analytics of patient clinical parameters. Biomedicine Radioengineering. 2025. V. 28. № 5. P. 64–68. DOI: https:// doi.org/10.18127/j15604136-202505-13 (In Russian)

1. Sv-vo o gosudarstvennoj registracii programmy dlya EVM №2024662143 (RF). Sistema upravleniya medicinskimi dannymi «MedData» (SUMD «MedData»)/ M.Yu. Kostenkova, A.M. Molodezhev. 2024 (In Russian).
2. Kostenkova M.Yu., Sidorova M. A. Primenenie IT-tekhnologij kognitivnoj vizualizacii medicinskih dannyh. Izmerenie. Monitoring. Upravlenie. Kontrol'. 2023. № 4(46). S. 72–80. DOI 10.21685/2307-5538-2023-4-9. EDN IXFNUR (In Russian).
3. Bishop C.M. Mixture density networks. 1994 (In Russian).
4. Gusev A.V., Novickij R.E., Ivshin A.A., Alekseev A.A. Mashinnoe obuchenie na laboratornyh dannyh dlya prognozirovaniya zabolevanij. Farmakoekonomika. Sovremennaya farmakoekonomika i farmakoepidemiologiya. 2021. T. 14. № 4. S. 581–592. DOI 10.17749/2070-4909/farmakoekonomika.2021.115. EDN ANFRKW (In Russian).
5. Ivanov A.I. Nejrosetevoj mnogokriterial'nyj statisticheskij analiz malyh vyborok. Proverka gipotezy nezavisimosti: Spravochnik. Penza: Izd-vo PGU. 2022. 218 s. (In Russian).