Forum: open-discussion

RE: Question about frequency and time domaine parameters [ Reply ] By: Frederic Briend on 2021-07-02 11:10	[forum:49021]
Thanks for all Abraham! Best.

RE: Question about frequency and time domaine parameters [ Reply ] By: Abraham Otero on 2021-07-02 05:47	[forum:49019]
Forget about non-linear statistics. You need even larger time intervals. It's okay to try with frequency, but I would put my best hope in the time statistics.

RE: Question about frequency and time domaine parameters [ Reply ] By: Frederic Briend on 2021-07-01 14:47	[forum:49018]
Thanks, really Abraham. I am lost, and do not know where I can find all these information. Thus, to sum up, for my protocol: Frequency Domain (only HF, by Episode +/=/-): window: 10s, intervals: 5s Time Domain: window: 10s, intervals: 5s *Is a non-linear analysis is possible? Best.

RE: Question about frequency and time domaine parameters [ Reply ]
By: Abraham Otero on 2021-07-01 10:42

[forum:49017]

My opinion is that they are too small intervals and it is going to be very difficult to find something in the frequency indices. The only one that might make any sense to calculate at those timescales is HF. I'll try something like 10 sec window and 5 scroll and do a visual inspection of some registers to see if it really looks like there is any change. But overall I'd have more hope of finding something with the temporal indices.

Another observation: I hope you have a good mechanism to synchronize the moment in which the images are displayed and the moment of occurrence of the heartbeat. A few seconds of error in the alignment of beats and pictures when the pictures are only displayed for 6 seconds would invalidate the experiment.

RE: Question about frequency and time domaine parameters [ Reply ]
By: Frederic Briend on 2021-07-01 09:23

[forum:49016]

Dear Abraham,
Thanks for your quick reply.

In fact, I have: 5s a scrambled pictures / 6s of Positive, negative or neutral pictures / 15s of scrambled pictures. Is it always not possible to extract information from that entire interval (6s image + 15s of scrambled image)?

They are not in serie, my software presented each picture in a random manner. If I concatenate all the signal according to the valence, is-it possible to extract information about the 3 valences (+, - and =)?

Thanks for your valuable help.

RE: Question about frequency and time domaine parameters [ Reply ]
By: Abraham Otero on 2021-06-30 16:50

[forum:49015]

Showing each picture for 6 seconds, and with a 5-second interval between photographs, I don't see any feasibility of extracting information from each individual photograph using HRV. It just doesn't have enough resolution. If all the negative, neutral and positive pictures are shown in series, and you want to extract information from that entire interval, it may be feasible.

In general, the smaller the time interval you want to analyze, the smaller the window size should be. But this means that, especially for the lower frequency bands, reliable estimates will not be extracted. If you wanted to use all the frequency bands, I would say that at least you would have to work with 5 minute intervals. Below 20 seconds, I'd say only HF *may* make sense. T shift you can set it to a half or a third of the size of the window.

Question about frequency and time domaine parameters [ Reply ]
By: Frederic Briend on 2021-06-30 13:42

[forum:49014]

Dear all,

Firstall, thanks for this great package!
Second, as a newbie, I am lost about the parameters of the time or the frequency domain analysis (CreateTimeAnalysis or CalculatePowerBand).

I have an elicitation of emotions protocol. I have 3 sets of visual stimuli including negative, neutral and positive pictures. Each picture was displayed for 6 s. A gray-scale slide with the mean luminosity of the presented pictures was displayed during the 5 s intertrial interval.

How should I proceed to calculate the size, the interval (or for the FD, the shift) of my protocol (for example for the positive images)?

Thanks!

Frederic