EEG reference methods in FieldTrip

Robert Oostenveld r.oostenveld at FCDONDERS.RU.NL
Tue Feb 15 15:36:05 CET 2005


Hi Laurens,

In the standard FieldTrip release there is indeed no function for the
computation of the scalp current density (or current source density, as
it is also called) from the EEG. I have implemented a
SCALPCURRENTDENSITY function in a fieldtrip style (based on spherical
splines as described by Perrin et al, Electroencephalogr Clin
Neurophysiol, 1989), but I found it to be very noise sensitive for high
channel densities (and also very slow), therefore I decided not to
include it.

However, what we also have been doing at the FCDC is to compute the
scalp current density in BrainVision Analyser (which uses a less noise
sensitive algorithm that smoothes over neighbouring channels) and to
export the resulting data from there to a BrainVision file.
Subsequently, the resulting file can be read into FieldTrip (of course
now without eeg referencing during PREPROCESSING) and you can do the
coherence computations in FieldTrip.

Alternatively, you can select two electrodes over sensory motor
regions, and compute a bipolar EEG channel from those electrodes (in
FieldTrip). That will result in an EEG signal that is very sensitive
for the underlying source of interest. However, you cannot use this in
combination with a topographic display of coherence.

I hope this helps,
Robert

PS have a look at http://www.egi.com/Technotes/SurfaceLaplacian.pdf,
which explains the motivation and technical background for doing this
scalp current density computation.


On 15 Feb 2005, at 10:21, Laurens Hogeweg wrote:

> Hello,
>
> I'm doing a study on EEG-EMG coherence during fatigue and when studying
> literature I came across the article of Mima and Hallett (T. Mima, M.
> Hallett / Clinical Neurophysiology 110 1896 (1999) 1892-1899) that
> describes
> the dependence of coherence strength on EEG reference method. They
> argue
> that it's best to use current source density or a Hjorth
> transformation. I
> wonder if it's possible to do such a transformation in FieldTrip,
> because I
> couldn't find it in the documentation.
>
> Thank you for suggestions,
>
> Laurens Hogeweg



More information about the fieldtrip mailing list