[FieldTrip] LCMV returns no time series?

Sarang S. Dalal sarang at cfin.au.dk
Fri Jun 17 18:54:04 CEST 2016


Hi Peter,

You likely have an Nvoxels x 1 cell array, which is as expected. Cells corresponding to voxels outside the brain will be empty, while the cells corresponding to voxels inside the brain should each contain an array of size 1xNsamples. (So effectively, you have Nvoxels x Nsamples, just repackaged into a cell array.)

Nutmegtrip is designed to visualize and navigate this entire output on your subject's MRI, or you can pull out some voxels of interest and plot the time series with ft_databrowser or with matlab's built-in plot function.

Cheers
Sarang

________________________________________
From: fieldtrip-bounces at science.ru.nl <fieldtrip-bounces at science.ru.nl> on behalf of Peter Goodin <pgoodin at swin.edu.au>
Sent: Friday, June 17, 2016 3:19 AM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] LCMV returns no time series?

Hi Sarang,

Regarding your point on that I should have the time series in source.avg.mom, the function only returns an n x 1 matrix. I've checked the averaged data and it does contain the channel x time series in av.avg, so I'm not really sure what's going on.

Peter



_______________________
Dr Peter Goodin,

Brain and Psychological Sciences Research Centre (BPsyC)
Swinburne University,
Hawthorn, Vic, 3122
________________________________________
From: fieldtrip-bounces at science.ru.nl [fieldtrip-bounces at science.ru.nl] on behalf of Sarang S. Dalal [sarang at cfin.au.dk]
Sent: Thursday, 16 June 2016 12:10 AM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] LCMV returns no time series?

Hi Peter,


If you've used the 'lcmv' method for ft_sourceanalysis, you ought to have the time series already contained in source.avg.mom.


I've been working a lot lately on improving the LCMV functionality, but unfortunately haven't had a chance to document much yet. Especially if you are not contrasting different conditions, you might want to try these parameters:

cfg.lcmv.weightnorm   = 'nai';  % this is a new option :-)
cfg.lcmv.fixedori     = 'yes';

I also prefer to turn off reducerank, but YMMV:
cfg.lcmv.reducerank   = 'no';



Finally I've also contributed a viewer that will greatly help visualizing your virtual time series together with the MRI. You can find it in the latest distributions of FieldTrip under fieldtrip/contrib/nutmegtrip. You'll also need a full version of SPM8 in your path (the external in FieldTrip doesn't have all the necessary functions).


You​​ can use it like this:

cfg = [];

cfg.mripath = '/path/to/mri/nii'  % must be in nifti (nii) format

cfg.funparameter = 'avg.mom'

% cfg.atlas = ft_read_atlas('ROI_MNI_V4.nii'); % [optional] you can get atlas labels, if your MRI and grid points are based on the MNI grid
nmt_sourceplot_spm8(cfg,ft_convert_units(source,'mm'));

Cheers,
Sarang


________________________________
From: fieldtrip-bounces at science.ru.nl <fieldtrip-bounces at science.ru.nl> on behalf of Peter Goodin <pgoodin at swin.edu.au>
Sent: Wednesday, June 15, 2016 3:32 PM
To: FieldTrip discussion list
Subject: [FieldTrip] LCMV returns no time series?


Hi FieldTrip list,

I'm using LCMV on averaged data (covariancewindow='all') but find the returned source estimates only have a single point. Is there a way to specify the analysis returns the time series for each grid point?

Thanks,

Peter


_______________________________________________
fieldtrip mailing list
fieldtrip at donders.ru.nl
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip




More information about the fieldtrip mailing list