[Eeglablist] Removing Components makes "index exceeds matrix dimension" error in SIFT

[boglyas]

Hy there,

I tried to remove components from EEG data set for better results, but I
cannot use the 3D brain movie in SIFT after that,because it gives an error
"Index exceeds matrix dimensions".

Basically it was a 61 channel recording with 60 (or 61) components, but I
had a data set that was already pruned, it had only 16 components/dipoles.
Some of the dipoles from the 16 were not in the cortex, they were deep
inside the brain (eg. thalamus), so I decided to remove those in order to
have better results later. After removing these components, if i check the
plotting in the EEGlab, or the  EEG <1x1 struct> in MATLAB, it all says
that it has 9 components left.

At the same time, if I try to run SIFT, every tool thinks that i have 16
components (and not 9). Even in pre-processing, the default "components to
keep" goes like: 1; 2; 3; 4; 5; 6; 7; 8; 9; 10; 11; 12; 13; 14; 15; 16;

Meanwhile it doesn't give error during the pre-pocessing or the model
fitting, and the validation also seems ok. I can visualize the connectivity
result in a Time-Frequency grid (and it only plots 9 x 9 figure), but when
I try to visualize in BrainMovie3D, it crashes, error messages, and so on.

What should I do, or what did I do wrong?

Thanks,
Balazs


2013/8/6 boglyas <kamanbalazs at gmail.com>

> Hy there,
>
> I tried to remove components from EEG data set for better results, but I
> cannot use the 3D brain movie in SIFT after that,because it gives an error
> "Index exceeds matrix dimensions".
>
> Basically it was a 61 channel recording with 60 (or 61) components, but I
> had a data set that was already pruned, it had only 16 components/dipoles.
> Some of the dipoles from the 16 were not in the cortex, they were deep
> inside the brain (eg. thalamus), so I decided to remove those in order to
> have better results later. After removing these components, if i check the
> plotting in the EEGlab, or the  EEG <1x1 struct> in MATLAB, it all says
> that it has 9 components left.
>
> At the same time, if I try to run SIFT, every tool thinks that i have 16
> components (and not 9). Even in pre-processing, the default "components to
> keep" goes like: 1; 2; 3; 4; 5; 6; 7; 8; 9; 10; 11; 12; 13; 14; 15; 16;
>
> Meanwhile it doesn't give error during the pre-pocessing or the model
> fitting, and the validation also seems ok. I can visualize the connectivity
> result in a Time-Frequency grid (and it only plots 9 x 9 figure), but when
> I try to visualize in BrainMovie3D, it crashes, error messages, and so on.
>
> What should I do, or what did I do wrong?
>
> Thanks,
> Balazs
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20130806/540af86f/attachment.html>