This is a very complicated question because you first need to understand how much noise is in the signal and that depends on the quality of the recording. A lot of systems boast of 1000 Hz resolution of the ECG as is that would solve everything, but without a way to overread the actual beat detection and beat labeling and a way to limit the beat to beat files to the universe of actual intervals, you can get very misleading data from supposedly high res systems. It is vital to plot out the HR time series of putatively normal beats on a scale that makes it possible to identify outliers and verify them. Beyond that, once there is an accurate time series, you can go to Kubios online to get a lot of HRV measures. Really it comes down to what question you are asking of the data.
Phyllis , I apologize but I think that you misunderstood. Kubios gives classical linear and non linear measures. Go to the site www.saistmp.com and find the new method and software. The analysis is performed under a new and important perspective. Please follow also the updates that we are adding every day .I think just this evening we will have a new update but read the method and start to use the software to appreciate the method.
Had a quick look at your website (saistmp.com). It is interesting. I would like to contribute some my work through the site, but could not figure out how to use it.
I would be happy to work with you to check out the analysis in an existing dataset but you would have to do the analysis and be okay with old fashioned sampling rates. I do not have the time or resources to use what is on your site without a lot of incentive :-) If you want to do something with this offer, contact me directly at [email protected]. We have 18,000 annotated recordings in many, many datasets, mostly human:-)
I do not see the problem. Read the method that is in pdf. Download the software that just to day will be updated also to evaluate the asymmetry in heart rhythm not still in time asymmetry. Soon after install the software and use as file R-R usual time series that you certainly have . They must be in .txt. The software performs the analysis in the usual VLF,LF,HF bands as by using standard FFT method and of course compare the basic FFT method results with those obtained by the new method. One studies Power spectral Density ( PSD) and relative amplitude the new method instead takes in examination the Variability
Dear Phyllis , i just now went to read your comment. It must be clear that I do not ask help from anyone. If you are interested, go to the site , read carefully the method and if you intend it and appreciate it , you may download the software and apply it. This is a great courtesy and dissemination that I perform for the otrhers... if they want .. obviously.
how nice but how does it handle an annotate time series where there are normal, ventricular, supraventricular and unclassified beats? That is the form of our RR interval files and if we strip out the annotations, that loses data. There is an intermediate step to create the time series on which HRV is based, but I don't have easy access to that.
if you have not an accurate softwware for peak detectionunfortunately you cannot determine R-R intervals or using ECG or plethysmography. In a your previous E-mail you also mentioned 1000 Hz. Do I intend correctly? 1000 hz for an ECH : Is it not so much? It is not necessary. In any case my method has been elaborated just to take into account variability in R-R inetrvals and thus to examine also rhythm disfunctions. Of course standard methods do not enable aned NEVER should be used to examine other beats in addition to N.N intervals.