Research on AI-based methods for modeling, analyzing and forecasting biomedical time series by applying a digital twin concept

Expected Results of the Project Implementation

In accordance with the project objectives and the implementation of the work program, the expected results are as follows:

1. Project management and reporting.
   Timely submission of reports for all Work Packages and Project Stages. Development of a comprehensive methodology for project sustainability and risk management. Achievement of the planned scientific results with high quality, delivered on time and in compliance with ethical regulations and the protection of volunteers’ data.
2. Methodological framework for HRV analysis.
   Development, testing and validation of a new methodological framework for HRV analysis, integrating classical, fractal, entropic and nonlinear methods into a unified concept. Proven effectiveness and universality of the framework for real-world data. The framework will include a minimal yet informative set of indicators capable of reliably distinguishing physiological, stress-induced and pre-risk states.
3. Synthetic cardiological data.
   Objective assessment of the advantages and limitations of different approaches for generating cardiological data and selection of the most effective and reliable method for practical application. Formulation of principles and criteria for building synthetic databases with high physiological fidelity and sufficient variability and statistical coverage for training and validation of AI algorithms.
4. New indicators, dependencies and biomarkers.
   Discovery of new indicators and correlations enabling more accurate assessment and differentiation of physiological states. Identification of new dependencies between HRV parameters and physiological states (rest, cognitive load, stress, pre-risk states) that cannot be clearly revealed by classical methods. Definition of potential biomarkers and combined indices for improved state stratification. Identification of new regularities and latent factors that enrich existing knowledge of autonomic regulation and serve as a basis for personalized AI models and digital twins.
5. VR/MR as a research tool.
   Establishment of a scientific rationale for considering VR/MR environments as an independent research problem. Definition of criteria for reliability, applicability and reproducibility of virtual stimuli in the study of psychophysiological responses.
6. Personalized digital twins and clinical validation.
   Study of patients with different diagnoses using personalized digital twins, including:
   • Identification of new dependencies between HRV indicators and clinical conditions;
   • Development of biomarkers for risk assessment and recovery dynamics;
   • Demonstration of the applicability of digital twins as scientific tools for studying pre-risk and pathological conditions.
7. Dissemination of results and publication activity.
   Open access scientific publications (total for both project stages), including:
   • For the entire project: 8 publications with Impact Factor (IF) and 4 publications with SCImago Journal Rank (SJR);
   • Additionally, 6 more publications in refereed and indexed journals;
   • Participation in international and national seminars and conferences;
   • Implementation of an open access repository for the purpose of open source and open access, prepared recommendations, public publications, etc.
8. Project website.
   Creation and periodic updating of the project website.