[FieldTrip] Does deleting bad channels in single subjects create problems for grand average later?

[Raquel Bibi]

Hi Kathrin,
I like having my data saved before (I can apply a montage later) and after
repairing channels.  As Hweeling suggests, you will have difficulty grand
averaging if the channel of interest is missing from one of your subjects.

I no longer have your original post and apologize if I am repeating
suggestions from others.  Did you review and display your neighbors before
repairing channels?  Are you sure your channels had neighbors?  I've used
this function for years - when I switched EEG systems, I was unaware of the
change in units, and ran into the same problem you are having.  I think you
are best off understanding why ft_channelrepair isn't working.  I also
found it helpful to have the ft_default settings provide me with feedback:
ft_default.trackconfig = 'report';
ft_default.checkconfig = 'pendantic';

Hopefully my suggestions help you determine what's going on.  If not, I
could review your neighbors, and data structure.

Best,

Raquel

On Mon, Aug 18, 2014 at 2:58 AM, Hwee Ling Lee <hweeling.lee at gmail.com>
wrote:

> Dear Kathrin,
>
> Re your previous post about repairing bad channels, I've also tried this
> on Fieldtrip and it works for me. I've not used the triangulation method,
> but maybe you could try again. I guess it depends on your template. For me,
> I was using Easycap 64 channels, but maybe it's different for you. Here's
> how I set up my repair channels options:
>
>     % repairing bad channels with averaged of its neighbours
>
> load('/home/leeh/Programs/fieldtrip-20140803/template/layout/easycapM11.mat');
>
>     cfg_neighb.method = 'template';
>     cfg_neighb.layout = lay;
>     neighbours = ft_prepare_neighbours(cfg_neighb, data2);
>
>     cfg = [];
>     if strcmp('S004', dir_analysis{sub}) == 1
>         cfg.badchannel = {'C4', 'TP9', 'FC4', 'C2', 'F4'};
>     end
>     cfg.trials = 'all';
>     cfg.method = 'average';
>     cfg.layout = lay;
>     cfg.neighbours = neighbours;
>     data3 = ft_channelrepair(cfg, data2);
>
> Re removing bad channels. Your question regarding if this will create
> problems for your grand average over subjects later. Well, it depends which
> channels you remove for individual subjects. As far as I'm aware, when you
> averaged your data across subjects, it will only pick the common channels
> across subjects for the statistical analyses. When you plot, you can always
> call for the template layout (e.g. 'EEG1020.lay'). This will plot the
> available channels across all subjects, and leave out the ones that are not
> in the data.
>
> I hope this helps.
>
> Cheers,
> Hweeling
>
>
>
> On 17 August 2014 11:23, KatrinH Heimann <katrinheimann at gmail.com> wrote:
>
>> Dear all
>> I still have problem with the channel interpolation (just not working at
>> all, without error message, see my previous post). Now I am thinking if I
>> just remove the bad channels for the single subjects and leave it like
>> that. My question is though: Will this create problems for my Grand average
>> over subjects later (as in each subject I remove different channels)? and
>> which layout do I have to use to plot my Grand average data later?
>> Every help is very welcome
>> Cheers
>> k
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>
>
>
> --
> =================================================
> Dr. rer. nat. Lee, Hwee Ling
> Postdoc
> German Center for Neurodegenerative Diseases (DZNE) Bonn
>
> Email 1: hwee-ling.lee<at>dzne.de
> Email 2: hweeling.lee<at>gmail.com
>
> https://sites.google.com/site/hweelinglee/home
>
> Correspondence Address:
> Ernst-Robert-Curtius Strasse 12, 53117, Bonn, Germany
> =================================================
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140818/8664fac5/attachment-0002.html>