Aims: This study investigates dynamic heart rate variability (HRV) responses during exercise and recovery in athletes and non-athletes. The aim is to determine whether time-resolved and correlation-based HRV metrics, particularly dynamical detrended fluctuation analysis (DDFA), can better characterize differences in autonomic regulation and physiological adaptability between trained and untrained individuals.
Methods: ECG data from 12 participants (6 athletes, 6 non-athletes; age 17–33 years) were analyzed from the iAMwell dataset. Subjects completed a standardized treadmill protocol including rest, acceleration, steady running (85% HRmax), deceleration, and recovery phases. HRV metrics were computed using moving windows and included time-domain measures (SDRR, RMSSD), non-linear indices (SD1/SD2), and correlation-based measures (DFA alpha1 and DDFA alpha(t)). Temporal normalization of the different phases enabled group-level comparisons across exercise.
Results: During exercise, SDRR and RMSSD decreased markedly and similarly in both groups from rest, showing limited ability to distinguish training status under load. In contrast, athletes showed lower DFA alpha1 and DDFA alpha(t) during acceleration and steady phases, suggesting more adaptable and less correlated cardiac dynamics. Non-athletes maintained higher scaling exponent values, reflecting more rigid autonomic control. During recovery, all HRV measures increased, with DFA alpha1 and DDFA exceeding baseline values in both groups, indicating autonomic rebound. Athletes demonstrated a slightly stronger recovery response.
Conclusions: While conventional HRV measures show similar exercise responses in athletes and non-athletes, correlation-based metrics provide enhanced sensitivity to training-related differences in autonomic regulation. DDFA, in particular, captures dynamic changes in beat-to-beat correlation structure and may serve as a valuable tool for assessing physiological adaptability, training status, and recovery dynamics.