Modelling long-term heart rate variability: an ARFIMA approach.

Area of Science:

• Physiology
• Biomedical Engineering
• Time Series Analysis

Background:

• Long-term heart rate variability (HRV) analysis is crucial for understanding cardiovascular health.
• Traditional time-variant autoregressive models may not fully capture long-range dependencies in HRV data.
• HRV recordings often exhibit non-negligible dependence between distant observations, indicating long-range correlations.

Purpose of the Study:

• To develop and apply a novel method for capturing long memory in long-term HRV recordings.
• To improve the description of low- and high-frequency components in HRV spectral analysis.
• To investigate the temporal dynamics of long-memory parameters in HRV.

Main Methods:

• Selective adaptive segmentation combined with fractionally integrated autoregressive moving-average (ARFIMA) models.
• Application of the proposed method to long-term HRV recordings.
• Analysis of circadian variations in the long-memory parameter.

Main Results:

• The proposed method effectively captures long memory in HRV series.
• An improved description of low- and high-frequency components in HRV spectral analysis was achieved.
• The long-memory parameter demonstrated circadian variation, with distinct patterns during day and night periods in a 24-h recording.

Conclusions:

• The combination of selective adaptive segmentation and ARFIMA models provides a powerful tool for analyzing long-term HRV.
• This approach enhances the understanding of HRV dynamics, particularly long-range correlations and circadian patterns.
• The findings suggest that HRV analysis can benefit from models that account for long memory and time-varying characteristics.