Short-time indices of heart rate variability (HRV) and cardiac regulatory dynamics (CRD) throughout a 24-hr recording may provide an alternative to the collection of a full 24-hr recording, however, methodological approaches need to be further evaluated. PURPOSE: The purpose of this study was to examine the robustness of various methodological approaches of short-time indices of HRV and CRD throughout a 24-hr period. METHODS: Eight healthy males completed two 24-hr visits. R-R intervals were recorded continuously using a heart rate monitor. Measures of HRV include the root mean square of successive R-R intervals (rMSSD) and the standard deviation of R-R intervals (SDNN), while CRD was assessed using sample entropy (SampEn). Each 24-hr recording was separated into 145 epochs to create a new time-series (HRVEP). Length and position of these epochs were varied around every 10th-min: the 3-min before every 10th-min (B3), the 3-min following every 10th-min (A3), the 3-min splitting every 10th-min (S3), and the 5-min splitting every 10th-min (S5). The dimensionality and complexity of each of these epoched profiles were subsequently analyzed. Tests of equivalence (TOST) were used to compare the raw values of rMSSD, SDNN, and SampEn between epoching methods at the individual level while paired TOST tests were used to examine the dynamics of these epoched profiles between epoching methods. RESULTS: TOST test between epoching methods of the raw values for rMSSDEPand SDNNEPat the individual level were equivocal (p≥0.05), whereas SampEnEPshowed equality (p≤0.05). Further analysis of paired TOST test comparing the embedding dimension and complexity of HRVEPshowed inequality in the optimal embedding dimension of these time-series and statistical equality (p≤0.01) between the complexity of these time-series. CONCLUSIONS: Epoch-by-epoch analysis of rMSSDEP and SDNNEP were not equal whereas SampEnEPwas equivalent across epoching methods. Although the optimal embedding dimension of these time-series varied between epoching methods, the complexity of these time-series were similar between methods for all indices of HRVEP.
Read full abstract