[FieldTrip] ERP image from EEGlab in fieldtrip and power normalisation without baseline

[Arnaud Delorme]

Dear Vitoria,

there is no need to "export" data from Fieldtrip to EEGLAB to plot ERPimage figures. Simply call the "erpimage" function of EEGLAB and fill in the parameter using your fieldtrip data. The only thing you might have to do it is transform your data from a cell array to a regular 2-D array containing sample points in one dimension and trials in the other. Anything that is being shown on the EEGLAB ERPimage figure can be reproduced using this function on the command line.

To sort single trials by alpha power or alpha phase (for example), you do not need a baseline period. You can compare ERPimage accross subjects and perform statistics on them using the STUDY structure of EEGLAB.

Best,

Arno

--
Arnaud Delorme, PhD
Centre de Recherche Cerveau et Cognition - UMR 5549
Pavillon Baudot, Hopital Purpan, BP 25202
31052 Toulouse Cedex, France

On Nov 18, 2013, at 6:33 AM, Vitoria Piai <v.piai.research at gmail.com> wrote:

> Hi everyone,
> 
> In EEGlab, there is an interface to make the so-called ERP image: single-trial data is sorted, for example on the basis of RT, and plotted with a vertical line showing stimulus onset and another line showing the RTs. An example is shown here:
> http://sccn.ucsd.edu/wiki/File:1ERPimageinf.gif
> Has anyone ever implemented this in fieldtrip in a straightforward way? Or managed to interface with EEGlab and make these plots with FT structured data?
> I scripted some code outside of FT to make this kind of figure but they didn't look as nice - neither the code nor the figure :)
> I've never tried exporting from FT to EEGlab, maybe that's the only way to go about... but if anyone has thought about this as well, I'd be happy to hear.
> 
> The other thing I was wondering about is whether it'd work to make the same kind of figure but instead of using single-trial phase-locked activity, I'd use single-trial alpha power over time (averaged over a frequency range) instead.
> The problem I have in this case is that I don't have a valid baseline period to normalise against. I was reading some earlier threads about power normalisation (e.g., http://mailman.science.ru.nl/pipermail/fieldtrip/2012-March/004862.html) but many of the suggestions made still rely on pre- vs. post-stimulus normalisation. The usual (cond1 - cond2) / (cond1 + cond2) and its variations won't work here either since it's single-trial data. Should I give up or is there something I actually could do?
> 
> I'm looking forward to hearing any thoughts you may want to share!
> Cheers, Vitória
> 
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip