[FieldTrip] No backprojection following ICA

Schoffelen, J.M. (Jan Mathijs) jan.schoffelen at donders.ru.nl
Fri Jun 12 20:35:04 CEST 2015

Dear Peter,

Depending on what you want to do, you need or need not worry about the warning you get. Since you mention that in first instance you want to look at connectivity at the channel level, I think you are safe. The warning pertains to the application of the spatial transformation to the low-level description of the channel array (the elec-structure), and not to the actual data. In short, the back projection is applied to the data itself, but there is insufficient information available in the sensor description to be able to apply the same projection to this as well. The latter would be needed if at a later stage you want to do source reconstruction, for which you need to compute leadfield matrices, where the forward model should include information about the spatial topographies that have been removed.


On Jun 10, 2015, at 2:00 AM, Peter Lyons <plyons at udel.edu<mailto:plyons at udel.edu>> wrote:

Hello All,

My name is Peter Lyons, and I am a researcher at the University of Delaware.  Currently, I am conducting resting state EEG research with the intention of using Fieldtrip to calculate the power of different frequency bands as well as phase lag index between various sensors.

I have come across a problem when I attempt to call ft_rejectcomponent following an ICA of my data.  The ICA was performed using ft_componentanalysis, and the data was viewed using ft_databrowser to determine unwanted components.  Following these steps I ran this code:

cfg = []
cfg.component = [3 5]
data_clean = ft_rejectcomponent(cfg,data_ica)

which returned the following message:

"not applying the backprojection matrix to the sensor description"

My question is how do I proceed to allow the application of the backprojection matrix?
I assume I am leaving some vital piece of code out?

Prior to calling ft_componentanalysis my continuous data was bandpass filtered using ft_preprocessing, and the reference was removed using the code:

cfg = []
cfg.channel = {'all','-E65'}
data_64 = ft_preprocessing(cfg,data_filt)

I am using an EGI Hydrocel 64 electrode Sensor Net and sampling at 1000 hz.

Any advice would be greatly appreciated!


fieldtrip mailing list
fieldtrip at donders.ru.nl<mailto:fieldtrip at donders.ru.nl>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20150612/86952811/attachment-0002.html>

More information about the fieldtrip mailing list