[FieldTrip] Ft_VolumeRealign Shifting Manually Set Fiducials

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

Hi,

Given the low number of red dots in your figures, I think that the ICP does not make sense. The algorithm only works if you have a decent point-cloud with hundreds of points that define the headshape in sufficient unambiguous detail. In your case I would stick to the manually defined fiducials, as long as you are sure that these indicated positions coincide with the ones used during the measurements.

Good luck,
Jan-Mathijs


On 9 Jul 2024, at 03:22, Parker via fieldtrip <fieldtrip at science.ru.nl> wrote:

Hi,

I have a question about MEG anatomical coregistration, particularly when using ft_volumerealign to do fiducial selection and coregistration refinement with headshape and ICP.

My current steps with ft_volumerealign are: 1) adding ‘coordsys = neuromag’ field to my mri by manually selecting N,LPA,RPA fiducials 2) refining fit with cfg.method = 'headshape' and cfg.headshape.icp = 'yes' with the headshape provided from my MEG file 3) plotting realigned mri to do a visual check on coordinate system with ft_determine_coordsys. (My code is also below)

After doing the refinement step, I notice the fiducial positions I initially set are almost always shifted. I have linked a picture for example showing 1) coordinates after my fiducial selection, 2) coordinates after headshape refinement - https://drive.google.com/drive/folders/13SAOL7FYM2yCiUPTV9StJh1HfNr_hDGP?usp=sharing<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdrive.google.com%2Fdrive%2Ffolders%2F13SAOL7FYM2yCiUPTV9StJh1HfNr_hDGP%3Fusp%3Dsharing&data=05%7C02%7Cfieldtrip%40science.ru.nl%7C296f77d8c00f4683f9e608dcaa3d7602%7C084578d9400d4a5aa7c7e76ca47af400%7C1%7C0%7C638572428148357773%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=FTbITxBCNhbzdE2qSG%2FgVCuYBLxzQ8lz5ggKVav%2FOIg%3D&reserved=0>

In the second image, the nasion is shifted upwards to what I would consider an incorrect position. (I also noticed that when plotting my source model, the source points appear to be slightly shifted up compared to one of the subjects from FT tutorials).

I am guessing the reason why, for ex, my nasion fiducial shift upwards after headshape refinement is because my headshape variable from my MEG file has a fiducial field, which are instead the fiducials that are being used instead of my manually selected ones.

Nevertheless, I was wondering how I could still perform headshape refinement with ICP and ensure MRI/MEG shape alignment but instead re select the fiducial positions.

I appreciate the help,
Parker

Step 1 - Fiducial selection
cfg                           = [];
cfg.method                    = 'interactive';
cfg.coordsys                  = 'neuromag';
cfg.flip                     = 'no';
mri_realigned               = ft_volumerealign(cfg, mri);

Step 2 - headshape refinement
cfg = [];
cfg.method              = 'headshape';
cfg.headshape.headshape = My_MEG_fif_file;
cfg.headshape.icp       = 'yes';
mri_realigned           = ft_volumerealign(cfg,mri_realigned);

Step 3 - Plotting Realigned mri + coordinate system:
cfg = [];
cfg.method              = 'headshape';
shape                   = ft_read_headshape(filename);
shape = ft_convert_units(shape,'mm');



ft_determine_coordsys(mri_realigned, 'interactive', 'no');
hold on;
drawnow;

ft_plot_headshape(shape);
_______________________________________________
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/20240722/df9fee1c/attachment.htm>