Troubles importing EEG electrodes

Discussions
1

Hi Francois,

I'm having some troubles importing EEG electrodes from a .txt file using the EEG: ASCII: Name,XYX(.) option. The coordinates are flipped for some reason.

image.png695×592 58.2 KB

The coordinates were collected using a Zebris/BrainVoyager TMS neuronavigation system. I have the same problem if I try and upload using the polhemus .txt option instead of ASCII, or if I try using default anatomy instead of the individual anatomy. Note that the fiducial points are included in the file.

I have attached the .txt file for you to have a look at. Any help would be greatly appreciated.
014_electrode_positions.txt (2.8 KB)

Cheers,
Nigel

2

Hi Nigel,

The coordinates in your file are exported in a left-handed coordinate system (https://en.wikipedia.org/wiki/Cartesian_coordinate_system#In_three_dimensions), so a flip needs to be applied in one of the dimensions to be aligned on the scalp surface (everything in Brainstorm is stored in a right-handed coordinate system).

Is there an option somewhere in your software to export the positions in neurological (right-handed) instead of radiological (left-handed) coordinates?

Otherwise, you would have to edit the coordinates, with one of these options:

  1. Edit the text file and flip the sign of the Z coordinate for the electrodes
  2. Right-click on the channel file > Edit channel file, flip one of the dimensions, and re-align the electrodes on the scalp (automatic or manual).

Cheers,
Francois

3

Thanks Francois!

Flipping the sign of the Z coordinate in the text file worked a treat.

Another question. I'm importing some custom, individualised parcellations from freesurfer so that I can extract time series from them. I've followed the instructions in the link below, but would like some advice on the best order to do this:

Is it best to:

  1. import the new L + R pial surfaces and custom parcellations, downsample to 7500V to match the low resolution surfaces, merge them and set this new cortex as the 'default surface' before computing the head models; or
  2. generate the headmodels using the original surfaces and then project the sources to the imported surfaces containing the desired custom parcellations afterwards.

My intuition is that they should be identical, although (1) could save time . Thoughts?

Cheers,
Nigel

4

Hi Nigel,
I don’t know if there would be any advantage with option #2. I would go for solution #1, much faster and probably not very different.
Francois