Wavelet Parameters ISPC

Hi. I am trying to decide on optimal parameters for my wavelet deconvolution. I understand from ANTs that temporal precision is important for estimating PAC and therefore you recommend using relatively few cycles for wavelet length. It would seem to me this logic might also extend to estimating ISPC, but I wanted to make sure I’m understanding this correctly. Is it also a good idea to emphasize temporal precision over frequency precision for measures looking at phase difference across sites? I can’t quite wrap my head around how these parameters affect the estimation of phase.

Thanks for all you have done in methods teaching!


Hi Katy. Of course, there’s no single correct answer for this. However, you wouldn’t want the filter to be too narrow (thus, too much spectral precision), otherwise there will be no variance and hence perfect “synchronization.” That is, if the time series has exactly one frequency component (imagine a perfect sine wave), then any two time series at the same frequency will have phase angle differences all clustered around the same phase angle.

Anyway, my advice is to use the same parameters that you would use for a regular time-frequency power plot. That should be fine for ISPC as well.