plots and stats with biosemi128.lay

[Robert Oostenveld]

Dear Katya

If I do
 >> cfg.layout = 'biosemi128.lay'
 >> layoutplot(cfg)

then I get attached figure (which you should be able to reproduce).
Looking carefully it seems that D39 is lying on top of another channel.

If I then do
d = dist(lay.pos')
imagesc(d==0)

then I see that there are multiple off-diagonal elements in the
distance matrix (i.e. a 128x128 matrix containing the distance bwteen
all pairs) that are zero.  So there are multiple channels overlapping.
I don't know why that is and at the moment I don't know where the
biosemi layouts were coming from. The only information that I have
from the log is
   revision 1.1
   date: 2005/06/15 08:21:08;  author: roboos;  state: Exp;
   new layout files, based on excel sheet from biosemi

I don't know the details of this excel sheet any more, but probably
the excel sheet was already wrong.

But to come to your prime question: does it affect clustering? Yes it
will. You are using the layout to define neighbours. If the layout is
flawed, the neighbour definition might also be flawed. Please use the
neighbourselection function to make a neighbour definition explicitely
and check the neighbours by looking at the output of the
neighbourselection function.


best regards,
Robert

PS it would be nice if you could provide a better layout for the
biosemi128 cap



----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/neuroimaging/fieldtrip.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Picture 1.png
Type: image/png
Size: 37404 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20090408/9750cdb3/attachment-0002.png>
-------------- next part --------------



On 7 Apr 2009, at 13:56, Katya Vinnik wrote:

> Dear FT users,
>
> has anyone of you encountered this warning?
> Duplicate x-y data points detected: using average of the z values.
>
> I'm simply plotting my data with       cfg.layout      =
> 'biosemi128.lay';
>
> and it gives me strange plots  like the one here:
> picasaweb.google.com/lh/photo/qmd3BgpeEDegPo4kHdRW0w?feat=directlink
>
> basically data from channels B19 and 21 and D22 and 24 are overlayed
> (circled in red on the pictures).
> This message appears also with topoplot functions. In the layout file,
> howether ,positions of this electrodes are different (see below).
>
> Do you think it can affect cluster-based multiple comparison
> correction?
>
> For example, if I use
>
> cfg.layout      = 'biosemi128.lay';
> cfg.correctm = 'cluster';
>  [stat] = freqstatistics(cfg,x,y)
>
> does it calculate it properly?
>
> Thank you,
> Best regards,
> Katya
>
> ----------------------------------
> The aim of this list is to facilitate the discussion between users
> of the FieldTrip  toolbox, to share experiences and to discuss new
> ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html
>  and http://www.ru.nl/neuroimaging/fieldtrip.
>


----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/neuroimaging/fieldtrip.