[FieldTrip] Info about the power-line noise

[Stan van Pelt]

Hi Davide, Nice data (aside from what is indeed probably line noise). I see four approaches you could take, but not sure if it helps. You could just be unlucky with a very fluctuating line noise in this recording (see http://fieldtrip.fcdonders.nl/faq/why_is_there_a_residual_50hz_line-noise_component_after_applying_a_dft_filter). The options I see: - Check if there is actually a datapadding up to 10 s. With already preprocessed data (e.g., after 3rd order gradiometer correction), it is not directly possible to data pad for subsequent dft-filtering. It happened to me not so long ago, that's why I ask. I made a workaround by first creating trials of 10s length, then do the dft-filetering, and after that cut the actual trials out. - If not already done, apply 3rd-order gradiometer correction first (ft_denoise_synthetic, or in the CTF DataEditor) - Use a bandstop-filter instead of the dft - Source-project the data. With my data that usually also gets rid of most line noise (I use LCMV to reconstruct a virtual sensor time course) Hope that helps. Stan ----- Oorspronkelijk bericht -----
> Van: "Davide Rivolta" <drivolta81 at gmail.com>
> Aan: "Email discussion list for the FieldTrip project"
> <fieldtrip at donders.ru.nl>
> Verzonden: Maandag 2 juli 2012 18:20:13
> Onderwerp: [FieldTrip] Info about the power-line noise
> Dear fieldtrippers,
> I recently run into the attached figures (S01 files). It seems a
> failure of the power-line noise filter.
> I used the following settings quite succesfully for other subjects:
> cfg.dftfilter = 'yes';
> cfg.padding = 10;
> An example of good figures are the "KAT" files.
> These data come from a visual task where people see moving
> concentrical gratings. Each trial has a different length that range
> between around 0.5 until 3 seconds.
> Would the bandstop filter be the only possible solution here?
> Thanks,
> Davide
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
-- Stan van Pelt Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Kapittelweg 29, 6525 EN Nijmegen, Netherlands E-mail: stan.vanpelt at donders.ru.nl Website: www.ru.nl/donders/ Phone: (+31) (0)24 36 10981 Fax: (+31) (0)24 36 10989
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120703/d6ebf439/attachment-0001.html>