Back to the main page.

Bug 3200 - grad structure not correctly updated after ICA for Elekta neuromag data

Status NEW
Reported 2016-11-03 09:39:00 +0100
Modified 2016-11-03 11:33:27 +0100
Product: FieldTrip
Component: preproc
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P5 normal
Assigned to: Robert Oostenveld
Depends on:
See also:

Weiyong Xu - 2016-11-03 09:39:22 +0100

Hi, I am using ICA(first PCA) to remove eye blinks for neuromag data. However, when I checked the data structure after ICA, the channel type and channel unit information were not kept. then I did ft_rejectvisual, and the channel type and channel unit were added again, but with all the 306 channels having values of 'megmag' and all the channel unit 'T'. Before ICA, the grad structure has 102 'megmag' and 204 'megplanar' for channel type, and 102 'T' and 204 'T/m' for channel unit. I am wondering if this would affect the results for ft_prepare_leadfield and later the source analysis?

Weiyong Xu - 2016-11-03 09:43:43 +0100

You can find the test data and script at:

Robert Oostenveld - 2016-11-03 11:33:27 +0100

yes, I think it would affect source reconstruction results. I recommend as work-around for now that you copy the grad structure from the data prior to ft_componentanalysis into the data structure that you obtain after ft_rejectcomponent. I'll download your data.