[FieldTrip] Edge effect/artifact in time-frequency representation

[Stephen Whitmarsh]

Dear Carlota,

I would take a look at the time-course data, e.g. using ft_databrowser. If
there are eye movements/blink you should see them there.
If you don't have too many trials, you can also inspect individual trials
of the TFR to get a better sense of the consistency of the artefact, using
cfg.keeptrials = 'yes' and plotting them in a loop of some kind.
Lastly, since your artefact is at the end of the trial, and you have some
ideas of what might happen then, you might consider extending your trials
to get a better sense of the signal after your trial ends.
Because it is frontal, some muscle artefact might be present, again a good
reason to start looking at the timecourses, timelocked or raw, and maybe
remove artefacted trials or trial periods.

I hope this helps,
Stephen




Op ma 14 dec. 2020 om 21:33 schreef MUSTAFA YAVUZ <m.yavuz.18 at ogr.iu.edu.tr
>:

> Dear Carlota Pages,
>
> I'm not 100% sure about my expertise but I'll try to do a few
> comments. You stated that whether this would be due to a preparation for a
> blink which occurs after the end of the trial, however, as you also stated
> that this effect disappears once you calculate the difference between
> conditions; hene, it would be unlikely to be due to a random blink effect.
> I was wondering whether this would be due to line noise since the script
> here does not contain any cfg input for line noise removal (for
> ft_preprocessing). Hope my limited knowledge does not cause any problem.
>
> Best wishes,
> Mustafa
>
> PAGES PORTABELLA, CARLOTA MARIA <carlota.pages at upf.edu>, 14 Ara 2020 Pzt,
> 23:07 tarihinde şunu yazdı:
>
>> Dear Fieldtrip community,
>> I am Carlota Pagès, a Ph.D. candidate in the CBC of UPF (Barcelona).
>> Currently, I'm exploring the neural oscillations of music unexpectedness.
>>
>> I have run time-frequency analysis on my EEG data via wavelet convolution
>> and applied a dB baseline correction for my different conditions. However,
>> when I plot the output, it appears some activity that (in my opinion) does
>> not look as brain-related because of its high power values and might be
>> masking the effects of interest (best seen in electrode Fz). When I
>> calculate the difference between conditions, it disappears. Therefore, it
>> is equally present in all conditions.
>> I wondered whether it might be an artifact due to a preparation of blink
>> (which happens right after the end of the trial) or an effect of the
>> baseline correction (baselines are statistically equal across
>> conditions). Please find the image and my script (for a single subject)
>> below.
>> Have any of you observed something similar and have a clue on what could
>> be causing it? And if so, what would be a good way to correct it?
>> Any help would be appreciated.
>>
>> cfg = [];
>>
>> cfg.demean = 'yes';
>>
>> data_demean = ft_preprocessing(cfg, data);
>>
>> % wavelet convolution
>>
>> cfg = [];
>>
>> cfg.method     = 'wavelet';
>>
>> cfg.output     = 'pow';
>>
>> cfg.foi        = 30:1:60;
>>
>> cfg.width      = 10;
>>
>> cfg.toi        = -0.7:0.03:1.2;
>>
>> cfg.pad =  2;
>>
>> cfg.trials     = data.trialinfo == 1;
>>
>> wavelet_cond1 = ft_freqanalysis(cfg, data_demean);
>>
>> % db normalization
>>
>> cfg = [];
>>
>> cfg.baseline = [-0.2 -0.05];
>>
>> cfg.baselinetype = 'db';
>>
>> wavelet_cond1_db = ft_freqbaseline(cfg, wavelet_cond1);
>>
>> % multiplot
>>
>> cfg = [];
>>
>> cfg.colorbar     = 'yes';
>>
>> cfg.channel = {'all', '-Fp1', '-Fp2'};
>>
>> cfg.showlabels = 'yes';
>>
>> cfg.layout = acticap32;
>>
>> figure; ft_multiplotTFR(cfg, wavelet_cond1_db);
>>
>>
>> [image: multiplot.png]
>>
>>
>>
>>
>>
>> --
>> Carlota Pagès
>>
>> PhD Student
>> <https://www.upf.edu/web/lcc/entry/-/-/139611/adscripcion/carlota-pag%C3%A8s>
>> Language & Comparative Cognition Group <https://www.upf.edu/web/lcc> l Center
>> for Brain and Cognition <http://cbc.upf.edu/> l UPF
>>
>> *Office 24.326*
>> *Mercè Rodoreda Building*
>>
>> *c\ Ramon Trias Fargas, 25-27
>> <https://maps.google.com/?q=c%5C+Ramon+Trias+Fargas,+25-27+08005+Barcelona&entry=gmail&source=g>08005
>> Barcelona
>> <https://maps.google.com/?q=c%5C+Ramon+Trias+Fargas,+25-27+08005+Barcelona&entry=gmail&source=g>*
>>
>> _______________________________________________
>> fieldtrip mailing list
>> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>> https://doi.org/10.1371/journal.pcbi.1002202
>>
>
>
> --
> *Mustafa Yavuz *
>
> M.Sc. Candidate in Neuroscience
>
> Department of Neuroscience
> İstanbul University
>
> _______________________________________________
> fieldtrip mailing list
> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> https://doi.org/10.1371/journal.pcbi.1002202
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20201215/56fe4cdc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: multiplot.png
Type: image/png
Size: 97725 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20201215/56fe4cdc/attachment.png>