Multi-taper Time Frequency decomposition

[Rasmita]

Hi Team,
I would like to perform time-frequency decomposition using Multi-taper analysis. I have few doubts wrt the parameters used.

• Is there any way, I can use logarithmic spaced frequencies instead of linear ?
• From the fieldtrip documentation : Time-frequency analysis using Hanning window, multitapers and wavelets - FieldTrip toolbox
  
  

1. As seen, t_ftimwin has frequency dependent window lengths. However, in the description
   
   
   , timwin seems to be static.
2. Also, can you help me understand what is the relation of time resolution parameter of brainstorm with the number of cycles as discussed in the fieldtrip documentation :
   
   
   
   ?
3. Is the parameter time_step in brainstorm equivalent to cfg.toi ?

• Can we perform hanning taper analysis as discussed in Time-frequency analysis using Hanning window, multitapers and wavelets - FieldTrip toolbox in brainstorm ?

[mcp0228]

Hi Rasmita, although I cannot answer your questions directly I can direct you to some previous forums that were discussing something similar which are as follows; not sure how useful these would be for you though.

Multitaper spectrogram - Prerau et al. (2017)

Hello there Brainstorm team, (sorry for the long post in advance) For the past few days, I've been trying to generate a multitaper spectrogram (in the context of sleep EEG) using the method provided by Prerau et al. (2017). The authors of the article have thankfully shared the MATLAB code online that can generate their multitaper spectrogram. I was able to successfully generate the below multitaper spectrogram (Fpz channel) using the codes provided by the authors. The spectrogram shows a parti…

Time-frequency multitaper parameters

Dear Francois, I think this could be a very easy question but I still have some doubts. I have performed Morlet wavelet ( Time definition = -2 sec to 2 sec; Frequency definition = 1:1:60 (Linear); Central frequency = 1Hz; Time resolution = 3s; Measure = Magnitude; Output = Save individual time-frequency maps -for each trial-) time frequency analysis by trials on my dataset using Brainstorm. We have already obtained statistically significant time-frequency results in gamma bands running FT t-te…

Multitaper method

hi Francois, I'm replicating some results from a previous analysis in Fieldtrip in Brainstorm, I did a multitaper time-frequency analysis and I have some doubts about the settings in Brainstorm, I previously computed power in Fieldtrip by setting: cfg.foi = 30:2.5:120; as frequencies of interest cfg.toi = -1.05:0.05:2; as times of interest cfg.t_ftimwin = ones(length(cfg.foi),1).* 0.4; as fixed time windows of 400 ms cfg.tapsmofrq = ones(length(cfg.foi),1).* 20; 20 Hz as fixed frequency s…

Multitaper time-frequency decomposition

Dear Brainstormers, I would like to perform a multitaper time-frequency decomposition as allowed by very recents updates in brainstorm. This involve the ft_mtmconvol function from fieldtrip but I encounteer some difficulties in setting the analysis parameters directly through the bstm interface. The ft_mtmconvol fieldtrip function need 4 inputs - i.e. the frequencies of interest, the time-interval of interest, the length of the sliding time-window in seconds and the width of frequency smoothin…

[Rasmita]

Hi @mcp0228,
Thank you for pointing to the previous discussions. I couldn't find answer to my earlier posed questions in any of them.
Any help on the above posed questions is highly appreciated!
Thanks,
Rasmita

[Rasmita]

Hello Experts,
I am experiencing errors while running multi-taper time-frequency decomposition method in brainstorm which is installed on a linux server.

Upon running the above config, i get the following error:

However, when I run the same config (used shorter time window as RAM is limited) on my windows based PC - the time frequency plot is calculated

but, I am getting errors when I click on it to view:

and below log after clicking OK

I am not sure what is causing this differential behaviour on windows PC and linux server when both of them are running the same brainstorm version.
Can you please help me out ?

[Sylvain]

As mentioned in the first error message, make sure you close all visualization windows before you attempt to visualize the Multitaper's contents. I also note the time window is different between the Windows and Linux screen captures share above.

[Rasmita]

Yes @Sylvain, I closed all the windows and attempted the Multi-taper process run.
My windows pc has 32 gb and was throwing me error of out of memory. Thus, I ran only the few first minutes.
However, I got the same error as shown above on linux server irrespective of the time-window.

I suspect, it could be bug in the code as I had tried multi-taper few weeks ago, it had worked but now, I couldn't recollect which version of brainstorm was it.

[Rasmita]

Hello @Sylvain @Raymundo.Cassani ,
I had passed the entire 20-280s recording for multi-taper. However, the output plot has time segments ranging from 25 to 274.5 seconds only.

Are the edge artefacts already removed before saving the file ?
Also, in the code - it seems edgeeffectmask is not being created for mtmconvol. Can you help me understand this better?

Thanks

[Sylvain]

The screenshot above (top right after Time: ...)) shows that your data run from 25 to 274.5s, hence the duration of the output. For the edge effect component, @Raymundo.Cassani will be able to look into this.

[Rasmita]

Hi Sylvain,
I ran the process from 20 to 280.

[Raymundo.Cassani]

It seems right, the times indicate the center of the analyzed windows:

With the Time resolution = 10s and Time step = 0.5s. The first window goes from 20s to 30s, center 25s, and the last window goes from 269.5s to 279.5s, center 274.5s

[Rasmita]

Got it!
Thank you!
@Raymundo.Cassani, what's with the edge artefact in mtmconvol ?

[Sylvain]

I meant to say that the first time point available in the data is at 25s (not 20) and the last one is at 274.5 (not 280). Simply check the All File check box instead of entering specific time points.