artifact rejection

[Robert Oostenveld]

Hi Esther,

On 26 Jun 2006, at 15:04, Esther Meeuwissen wrote:
> and get the following error message:
>
>  Error using ==> horzcat
> CAT arguments dimensions are not consistent.
>
> Error in ==> rejectartifact at 182
>     cfg.artfctdef.type = [{'file'} cfg.artfctdef.type];

I suspect here that cfg.artfctdef.type is a cell-array with strings
represented as a vertical cell-vector, and that line 182 tries to
combine it with something that Matlab treats as horizontal cell-
vector. It is unclear to me why it does that, since the
cfg.artfctdef.type is built up internally (you did not explicitely
specify it).

> Error in ==> preprocessing at 317
> [cfg] = rejectartifact(cfg);
>
> Error in ==> preproesther1 at 45
> data = preprocessing(cfg);
>
> By trying to make it work I figured out that it is possible to do
> both the
> artifact rejection parts separately using (parts of) the script
> above but
> (Fieldtrip wants to combine two matrices to 1 matrix and that is
> what goes
> wrong). Does anyone know how (/whether) it is possible to do a
> artifact
> detection/rejection in Fieldtrip partly using functions of Fieldtrip
> itself and partly read markers of EEG parts containing a artifact (the
> file made by brain vision analyzer)?

I have tried replicating your problem (with the data that I found on /
home/electromag/estmee/GOs/analyses/datagoed), but did not encounter
any problems up to the definetrial, artifatc_muscle and
rejectartifact. However, the problem comes only in when preprocessing
is called. For backward compatibility with other peoples old scripts,
preprocessing is also calling definetrial and rejhectartifact, and
rejectartifact seems confused by being called twice.

I made a small change to rejectartifact that hopefully fixes the
problem. The updated version is on home/common now and will be on the
ftp this evening.

best regards,
Robert