[FieldTrip] Magnetometers and gradiometers for source localisation

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

Hi Casper,

These are all very good and relevant questions. As a matter of fact, a group of colleagues and friends (i.e. methods people, epilepsy researchers and core developers of fieldtrip/mne_python/spm/brainstorm)) have been working on this topic on and off over the past years, meeting at Aston University, to try and understand the best way to do beamforming analysis on Elekta MEG data. We think that we now understand what’s the best way to approach this, and work is underway to disseminate this to the community.

In principle, gradiometers and magnetometers can be combined in a single analysis, but this is unlikely to be optimal. Reason: the different order of magnitude of the signals (which may be aggravated when changing the metric units of the data (e.g. going from T/m to T/cm for the gradiometers)) causes the covariance matrix (used for the beamformer spatial filter computation) to be differently weighing in the mags versus the grads. Indeed, if it’s deep sources you’re after, you may want to exploit the depth sensitivity pattern of the mags.

Some software packages address this by differently weighing the sensor types (by means of a baked-in default weighing scheme). A more principled way of addressing this would be to spatially whiten the data (and apply the same whitening to the leadfields). In FieldTrip this prewhitening can be done with ft_denoise_prewhiten, using the covariance estimated in a baseline window as the prewhitener. If the balanced gradiometer array is subsequently used for the forward computation, the same prewhitener is applied to the leadfield, so everything stays consistent.

One thing that is important to take into account, is the rank-deficiency of maxfiltered data. This can cause problems in the mathematical inversion of the covariance matrix, when done inappropriately. In order to have more flexibility in optimally inverting covariance matrices, we now have a dedicated ft_inv function (in fieldtrip/private), which allows for different ways of (regularized) matrix inversions. For maxfiltered filtered data, it would be recommended to use a ‘kappa’ parameter for inversion, i.e. to truncate the svd of the coviance matrix at (or just below) its rank.

Currently I am working on some documentation (that uses the Wakeman and Henson dataset) that demonstrates some of these issues. It will be presented at this years practicalMEEG workshop in Paris (http://practicalmeeg2019.org), so stay tuned! This documentation will become available soon.

Best wishes,
Jan-Mathijs




On 15 Nov 2019, at 10:18, Casper Kerren <C.Kerren at pgr.bham.ac.uk<mailto:C.Kerren at pgr.bham.ac.uk>> wrote:

Hi,

I am currently source localising an effect using LCMV on Elekta MEG data. I have both gradiometers and magnetometers and have until recently treated these as the same when preparing the leadfield. However, after having had discussions with my colleagues about the different units and strength for grads and mags, we are hesitant as to what might be the best way of selecting channels. I have searched the fiedltrip mailing list, and have found some answers, but nothing that I can say is very convincing. Could someone point me in the right direction regarding this? Perhaps someone who has done some checks on using both grads and mags and using only grads or mags. We are interested in a deeper source, and naively thought that magnetometers would be necessary to include. However, is this just as I say: naïve?

In the ft_prepareleadfield function I can see that gradiometers is the only channels that are thought of being included, so does fieldtrip actually expect us to exclude magnetometers?

Kind regards,

Casper Kerrén, PhD student
School of Psychology
College of Life and Environmental Sciences
University of Birmingham
Edgbaston
Birmingham B15 2TT
e-mail: cxk699 at student.bham.ac.uk<mailto:cxk699 at student.bham.ac.uk>



_______________________________________________
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/20191115/9ae2f6cc/attachment.html>