A Novel Method for Baroreflex Sensitivity Estimation Using Modulated Gaussian Filter

Figures

• 

  Figure 1. BRS analysis with different values of κ (1.1, 2 and 3). It can be observed that the BRS value is identical for each value of κ. In this study, κ is constantly set to 2. The data points beyond the whiskers are marked by red + symbol.

• 

  Figure 2. Generation procedure for the modGauss filters in LF and HF bands (from time-domain perspective).

• 

  Figure 3. Generation procedure for the modGauss filters in LF and HF bands (from frequency-domain perspective).

• 

  Figure 4. Derived results for the interpolated SBP sequence in both the LF and HF band. Results from time-domain perspective with the B002LB data of EuroBavar dataset as example.

• 

  Figure 5. Derived results for the interpolated SBP sequence in both the LF and HF band. Results from frequency-domain perspective with the B002LB data of EuroBavar dataset as example.

• 

  Figure 6. Derived results for the interpolated IBI sequence in LF band and HF band (from time-domain perspective, with the data record B002LB of the EuroBavar dataset as an example).

• 

  Figure 7. Derived results for the interpolated IBI sequence in LF band and HF band (from frequency-domain perspective, with the data record B002LB of the EuroBavar dataset as an example).

• 

  Figure 8. Estimation of the BRS values by five different methods including modGauss, GAFD, AR spectrum, Welch’s periodogram and wavelet. The data points beyond the whiskers are marked by red + symbol.

• 

  Figure 9. Long-term BRS analysis by modGauss in comparison with sleep stage (with the sleep stages 1–4 being denoted by 01–04, the subject motion by 0M, rapid eye movement sleep by 0R, and the awake stage by 0W).

• 

  Figure 10. Long-term sleep BRS analysis by five methods including GAFD, AR spectrum, Welch’s periodogram and Wavelet.