[FieldTrip] problem with leadfield

[Schoffelen, J.M. (Jan Mathijs)]

Hi Philip,

Yes, indeed.

Best wishes,
Jan-Mathijs



On 12 Apr 2021, at 13:40, philip Joadavi <p.joadavi at gmail.com<mailto:p.joadavi at gmail.com>> wrote:

Dear Jan-Mathijs,

Thanks for your quick reply.
so if I sum over the channels on the selected dipole, I get these results :
so
for leadfield computed with 64 channels:

1.0e-14 *
   -0.1818   -0.6828    0.5815

for leadfield computed with 32 channels:

1.0e-13
    0.0046   -0.1538    0.0872

and for the subset of 32 extracted from 64 channels leadfield :

-0.3738   -0.4471   -0.2351

is that what you are referring to?

Best,
Philip

On Sun, Apr 11, 2021 at 6:59 PM Schoffelen, J.M. (Jan Mathijs) <jan.schoffelen at donders.ru.nl<mailto:jan.schoffelen at donders.ru.nl>> wrote:
Hi Philip,

If I understand you well, the 32 channel array is a subset of the 64 channel array?

Perhaps it may be useful to you know that FieldTrip computes EEG leadfields as if the channel array is average referenced. This means that - for a ‘leadfield’ topography for a given dipole - the average along the rows is 0, by construction. If you construct a 64-channel leadfield, which will have an average of 0 along its rows, a selection of 32 rows (i.e. your 32 channel subset) will not by default have a zero mean across the rows. However, had you pre-selected those 32 channels and computed the leadfield, the average would be 0.

Could this explain the difference you describe?

Best wishes,
Jan-Mathijs


On 11 Apr 2021, at 16:14, philip Joadavi <p.joadavi at gmail.com<mailto:p.joadavi at gmail.com>> wrote:

Dear all,

I'm trying to compute the leadfield for my EEG data.
I've done all the steps in the tutorial, here<https://www.fieldtriptoolbox.org/workshop/oslo2019/forward_modeling/>.
I run the bemcp, for computing the headmodel. Since in my data I have 2 electrode sets (64 channels and 32 channels), I tried to compute the leadfield for 64 and then extract the leadfield from that file for 32 channels. However, when I compute the leadfield for 32 channels separately, the results are different from when I extract the 32 channels from the 64 channel leadfield.

I have checked everything, the patterns that I get are exactly the same, but the values from the leadfield computed separately from 32 channels are higher when I plot the potentials (using the simulation in the tutorial).
Can someone explain why is this the case and where it happens in the code for computing the leadfield? why the number of channels is important for leadfield computation?

Unfortunately, I can not run the 'simbio' to compare.

Thank you very much!
Best,
Philip
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202

_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20210412/188dfff1/attachment.htm>