[FieldTrip] append CTF grad structures

Schoffelen, J.M. (Jan Mathijs) janmathijs.schoffelen at donders.ru.nl
Thu Nov 18 09:25:26 CET 2021


Hoi Matthias,

If you’re confident enough that the participant’s head position relative to the sensors is sufficiently similar I agree that the best way would be to obtain an average of the two gradiometer structures. You might want to consider using a weighted average if the length of the data-of-interest in the two chunks is out of balance.

Your best friend in this case is the ft_average_sens function.

I’d think that if you’re after the grad structure that is closest to the one, had the acquisition not been stopped, then you need to use the grad that pertains to the first chunk of data, because the position information in (any) CTF dataset is taken from the initial headpositioning step.

In other words, although it is not often done as far as I know, it may in general be better to register the head to the average position throughout the data segments of interest, rather than to time 0. If you’re worried about subject movement, or if you feel adventurous you could have a look at the ft_headmovement function (provided that your data contains continuous head position information).

ft_megrealign tampers with the data, and I would in general be reluctant to use it.

Best wishes and I hope all is well!

Jan-Mathijs




On 16 Nov 2021, at 18:25, Matthias Franken via fieldtrip <fieldtrip at science.ru.nl<mailto:fieldtrip at science.ru.nl>> wrote:

Hi all,

I have a question with respect to the ‘grad’ structure in my CTF MEG data. For a given participant, data acquisition was stopped and restarted, leading two (or more) raw data files, with only a short break in between. The participant did not leave the MEG in between.
I would like to append the two datasets to preprocess them together, as if they were acquired in a single run. I am wondering though how to deal with the grad structure, given that the ft_appenddata seems to remove it. I am guessing this is due to slight differences in the coil and chan pos/ori matrices, presumably a result of slight head movements. Up til now, I was simply using the first file’s grad structure, but I guess there should be a better way. Would the most straightforward way be to average the positions and orientations in the grad structures of each run to get at a grad structure for the appended dataset? I.e., would this be closest to the grad structure that I would have gotten if data acquisition wouldn’t have been interrupted?
Or should I realign each using ft_megrealign? Though I am guessing this is not what just doing acquisition in one run would lead to either.

Thanks!



---
Matthias Franken, PhD
Postdoctoral Researcher
Motor Neuroscience Lab
Department of Psychology
McGill University

matthias.franken at mail.mcgill.ca<mailto:matthias.franken at mail.mcgill.ca>
https://frankenmatthias.github.io/homepage<https://urldefense.com/v3/__https://frankenmatthias.github.io/homepage__;!!HJOPV4FYYWzcc1jazlU!tsSYmQgffgOCZMMBN2nsV1tijivDU_neQi5yknO1kBQlHOIvWyZtNkSU5miyOhpalkUsaTeqzLvw01g$>




_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!tsSYmQgffgOCZMMBN2nsV1tijivDU_neQi5yknO1kBQlHOIvWyZtNkSU5miyOhpalkUsaTeqzJtxNXk$

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20211118/53a5c7b6/attachment.htm>


More information about the fieldtrip mailing list