Forum: open-discussion

Monitor Forum | Start New Thread

Nested Flat Threaded Ultimate

Show 25 Show 50 Show 75 Show 100

RE: Comparison of HRV values with HRV Toolkit [ Reply ] By: Abraham Otero on 2014-09-15 14:26

[forum:41458]

Hi Erik, A dirty secret in heart rate variability analysis is that it depends on several parameters/implementation solutions that are not standardized. Some of these are: -The algorithm to filter beats (filtering false positives or abnormal heartbeat) -The algorithm used to interpolate the RR to obtain a time series of constant sampling frequency. Common solutions are linear and cubic interpolation. But the latter can be problematic if there is a big gap in the RR series and at some point it is necessary to use some other solution. Linear interpolation and cubic are not the only proposed in the literature. -How to remove the DC component (from all the RR series, from each window...) -The window type (Hamming, Welch,…), window size and window overlap used in the Fast Fourier Transform. The end result of all this is that it is not possible to compare the results of HRV analysis made ​​with different implementations. That said, the results obtained with different libraries should be consistent qualitatively. And the results of an experiment should be reproducible if the same library used in the original experimen is used in the reproduction. One purpose of RHRV is precisely that everyone can access a free and open source implementation of these algorithmsn so that the results of different studies on hrv analysis are comparable. An interesting reading in this regard is: http://link.springer.com/article/10.1007/s10916-009-9383-5#page-1

Comparison of HRV values with HRV Toolkit [ Reply ] By: Erik Nlsson on 2014-09-15 10:43

[forum:41457]

Greetings! I have been comparing the data output from the RHRV package with Physionets "HRV Toolkit". The Frequency-domain variables give very dfferent results in RHRV: ULF (RHRV) = 25 ULF (HRVToolkit) = 2705 (!) Other parameters also differ 30-100%. I fed RHRV the data from the example presented at: http://physionet.org/tutorials/hrv-toolkit/#iv-representative-measurements-of-hrv Using the following settings: hrv.data<-CalculatePowerBand(hrv.data, indexFreqAnalysis=1, size=300, shift=30, ULFmin=0, ULFmax=0.0033, VLFmin=0.0033, VLFmax=0.04, LFmin=0.04, LFmax=0.15, HFmin=0.15, HFmax=0.4, type="fourier") The ULF max is changed from default 0.03 to 0.0033 (as used in the HRV toolkit and I believe is recommended in the litterature). Using the default limit of 0.03 gives ULF = 277, still far from the Toolkit example. Do you have any thoughts on why the difference is so large?