Psychological and psychophysiological substantiation and mathematical modeling for predicting pilot performance during simulator training

Introduction. Ensuring flight safety in modern conditions requires a quantitative assessment of the factors influencing a pilot’s professional reliability, including individual psychological characteristics, autonomic reactivity during professional activities, and subjective awareness of important aspects of flight operations. However, not all these issues have been studied in detail, especially within a comprehensive framework. In this regard, the development of objective tools for predicting the performance of flight personnel remains one of the fundamental issues requiring attention from specialists in aviation medicine. Objective. To validate a mathematical model for predicting pilot performance reliability during flight simulator training using psychological assessments of professionally important qualities and heart rate variability metrics. Materials and methods. The study involved 35 military aviators (11 second-year cadets and 24 instructor pilots) from Southeast Asia. We evaluated the subjective rank importance of 200 professional attributes grouped into nine domains, levels of neuroticism and extraversion (Eysenck Personality Questionnaire), and heart rate variability metrics (including the stress index) before and after a simulated emergency situation using the VNS-Ritm system (Neurosoft, Russia). Statistical analysis included Pearson’s chi-squared test, Spearman’s rank correlation, and the construction of a binary logistic regression model. Results. Unlike cadets, pilots prioritized the “sensory systems and perception” domain more frequently (53.2% vs. 40.6%), whereas cadets gave higher priority to “adaptive capabilities” (15.6% vs. 3.1%). The binary logistic regression model identified three predictors of high-risk inefficient performance (score ≤ 7.0). Specifically, a decrease in the perceived importance of the “sensory systems and perception” domain increases the odds of high-risk classification fourfold (odds ratio [OR] = 0.25; p = 0.008); a post-stress increase in the stress index also elevates these odds (OR = 1.004; p = 0.027). Additionally, the “adaptive capabilities” domain demonstrated a trend toward risk reduction (OR = 0.41; p = 0.064). The model demonstrated a classification accuracy of 85.3%, correctly identifying 80% of high-risk cases. Conclusions. A pilot’s professional development is accompanied by a structural transformation in professional attributes, with a shift from physical adaptation toward perceptual-cognitive efficiency. The perceived importance of sensory systems and autonomic stability constitute the key predictors of performance reliability under stress. The findings indicate that the developed predictive model can be applied for early identification of risk groups and simulator training optimization, though further refinement and validation remain necessary.