[Mne_analysis] Guide For The Selection of cwt_n_cycles

[balandongiv at gmail.com]

        External Email - Use Caution        

Hi Alex,

Thanks for the insight, it helps refine my search on the net.

I would like to confirm, say one is interested in finding the connectivity
between several channels within the frequency band between 1 to 5 Hz with
1Hz increment. Assume, the number of complex Morlet wavelet cycle is 3
(based on your rule of thumb). Does the value 3 represent the number of
cycles globally? Such that, in the spectral_connectivity module, the
parameter for cwt_n_cycles is set  equal to 3. 

Or something that I found unlikely (i.e., cycle for each frequency) but
presented in the aforemention basic mne example. Such that, the  cwt_freqs =
np.arange(1, 5, 1), then the
cwt_n_cycles = cwt_freqs /3?


Regards
Rodney

----------------------------------------------------------------------

Message: 1
Date: Tue, 17 Nov 2020 08:53:07 +0100
From: Alexandre Gramfort <alexandre.gramfort at inria.fr>
Subject: Re: [Mne_analysis] Guide For The Selection of cwt_n_cycles
	{Disarmed}
To: Discussion and support forum for the users of MNE Software
	<mne_analysis at nmr.mgh.harvard.edu>
Message-ID:
	<CADeotZo-vqkE5d_Xef_rAK=VdhDoKC-d=_U-FD=2eEMNneQXzg at mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

        External Email - Use Caution        

hi,

well I can just share with you my rule of thumb. If you truly look for
oscillations and not just evoked responses your time-frequency atom
(wavelet) should have at least 3 cycles. The more cycles the better the
frequency resolution and the worse the time resolution. So it's always a
trade-off and it depends on your data and your question.

HTH
Alex


On Tue, Nov 17, 2020 at 7:42 AM <balandongiv at gmail.com> wrote:

>         External Email - Use Caution
>
> Dear Group,
>
>
>
> I am trying to understand the tutorial with the title: Compute 
> seed-based time-frequency connectivity in sensor space, which is 
> accessible via the link
>
> *MailScanner has detected a possible fraud attempt from 
> "secure-web.cisco.com" claiming to be* 
> https://secure-web.cisco.com/1MR6NJbbVbHaHThcDpc-5P1Ke1Z02OLPDuH0pakIY
> 0P_kwRec-izqBnYljKZCvNBHHaOu0zz68P7Ng4-m4M7C4RpqFMoAT-qabSggYOy1kPJpK2
> SXZ0-Qsh-FHPg53t-42gH_Rn0ltMg43LCUPAHWma30bPEmFfKNhkEjlx_bJgybHWo34vMu
> YzDVWtUN29AkpRKswJEWEIqlGyXw1hhk2NFgaPhUqEox7WvfUO5l7zwasDzeI0xC_wPPh6
> 5c8_osFLGu77OBgi-fs_hHsdvgIQ/https%3A%2F%2Fmne.tools%2Fstable%2Fauto_e
> xamples%2Fconnectivity%2Fplot_cwt_sensor_connectivity.html
>
<https://secure-web.cisco.com/1C2D8Ha4l_YR0obIvl4YWwj-OVYq74zsi_-kmSQBMHGhWr
kYsCvbnT6C3TBjML1P0jYz0iTMCmzj8VuWgAwrBl-1VZ0fSDeiknpWg4TokGeow9brn6Pvv30Ql2
vShGCxTFoolf0eMuKQMUHw5cN9hmWrrALUPfbBoakcJbmYaUV0ByOW9nTwXlPP6pXTnHfXcdo-0s
HZcx0zH8TQFqw2EmN4_VBeR0CZS_5yMMfaDzyjHt0iEDAKf05fW0YOyDNmiMF_3fMJQ6rnt83RTK
t_19w/https%3A%2F%2Fmne.tools%2Fstable%2Fauto_examples%2Fconnectivity%2Fplot
_cwt_sensor_connectivity.html>.
> The tutorial objective is to compute the connectivity between a 
> seed-gradiometer close to the visual cortex and all other 
> gradiometers. The connectivity is computed in the time-frequency 
> domain using Morlet wavelets, and the debiased squared weighted phase 
> lag index is used as a connectivity metric.
>
>
>
> In the example, the band of interest are cwt_freqs = np.arange(7, 30, 2).
> Then,  the number of cycles is defined as cwt_n_cycles = cwt_freqs / 
> 7. May I know what is the logic/engineering explanation for the 
> selection of value SEVEN, here? I have searched the internet, but nothing
relevant comes out.
> This may be due to incorrect keyword use.
>
>
>
> Appreciate if someone kind enough to shed some light and direct me to 
> good reading material about this topic.
>
>

>
>
> Thanks
>
>
>
> Regards
>
> Rodney
> _______________________________________________