[Question] [Help] A example to run please

[vnl]

Thank you for actually working out the Frequency domain, but I am back again for a little help using the lib.
Also, please tell me off, if I need to read something before I even ask this question.

So I am using Kyto Reader to get the data in (I still need to get polar). It gives me hex data to read RR Intervals. Converting them, I get the following:
{967, 967, 983, 923, 923, 895, 895, 895, 937, 967, 967, 967, 983, 983, 952, 983, 983, 967, 1000, 1000, 1000, 967, 967, 983, 983, 983, 1016, 1000, 1000, 983, 1016, 1016, 1000, 1000, 1000, 967, 1000, 1000, 983, 1000, 1000, 967, 923, 923, 909, 952, 952, 983, 967, 967, 983, 923, 923, 895, 833, 833, 845, 869, 869, 869, 882, 882, 909, 937, 937, 983, 967, 967, 937, 952, 952, 967, 983, 983, 1016, 967, 967, 952, 869, 869, 895, 857, 857, 882, 937, 937, 895, 1052, 1052, 909, 869, 869, 857, 895, 895, 882, 882, 882, 909, 909, 909, 909, 895, 895, 937, 909, 909, 923, 952, 952, 952, 952, 952, 952, 923, 923, 923, 952, 937, 909, 895, 909}

Obviously, this is different from List data as in the data you mention.

class RrsData{
  RrsData(this.x, this.y);
  final int x; // time offset
  late final int y; // RR value
}

What is the time offset we looking for? As the input would be just RRI, correct?

Would you be able to share an example or if you can help me build it? I will be able to contribute to the example.

[Przemekeke]

Hello,
Sure, I've done it that way to easier visualize this data, but I see that it might be not so obvious.
RRs data is keeping the same data in 2 different formats:

• x (time offset) - lets say that first beat (onset) start at ms 0 than next will occur in 967 millisecond, and next in 967 + 967 = 1934 etc.
• y (rr values) - that's actually the list which you pasted above.

I've added that prototype in the docs, it works on events with separate rr:

// function to aggregate RrsData
  void _addToList(rr){
    rrs.add(RrsData((rrs.isEmpty ? 0 : rrs.last.x) + int.parse(rr.toString()), rr));
  }

It can be transformed to loop ofc:

rrs = [967, 967, 983, 923, 923, 895, 895, 895, 937, 967, 967, 967, 983, 983, 952, ...]
rrs_data = []
for (var rr in rrs) {
    rrs_data.add(RrsData((rrs_data.isEmpty ? 0 : rrs_data.last.x) + rr, rr));
  }

I will probably add simple list operations as well here.
Let me know if you have any more questions

[Przemekeke]

Just out of curiosity, what app are you building?

[vnl]

Just out of curiosity, what app are you building?

It is an application already out there (ios and android), called OptimalHRV. Using an API layer with hry_analysis python library that would do the job easily.

Wanted to take it a step further and try and see if I can produce real time charts for frequency domain, specifically for low frequency as the reading is taken. I do understand accuracy will take a hit at least the first 60 seconds...

Thank you again for sharing the above, it was really helpful.

One last question, how to determine rrsWindow? Or am I glaringly missing something and being silly.

[Przemekeke]

That's interesting. I've been similar thing POC at some point. Maybe we can collaborate somehow please write on e-mail: przemyslaw.marciniak2@gmail.com

[Przemekeke]

Regarding time window. In the research literature regarding HRV the 1 minute is absolute minimum. It is decreasing the power of the statistical artifacts within the signal. There is also a constrain of the frequency domain resolution which correlated to the points in time series.

Also I would recommend to use rrsWindow with a constant size since it would make the comparison between two HRV more accurate

ref: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8203359/