[FieldTrip] left to right inversion after source interpolation

[Frederic Roux]

Dear fieldtrip users,

I have noticed that after interpolating a dummy source volume with ft_sourceinterpolate,
voxels that were located in the left side of the volume, apear to be located in the right
side of the volume.

The code that I use is as following:

dummy is a dummy structure of an lcmv beamformer map with the fields

dummy =
dim :        [15 18 17]
time:        [1x272662 double]
pos:         [4590x3 double]
inside:     [1x2144 double]
outside:   [1x2446 double]
method: 'average'
avg: [1x1 struct]

voxpos = dummy.pos(dummy.inside,1);% get the coordinates of voxels inside the volume
dummy.avg.pow(dummy.inside(sign(voxpos)==1)) = 0; % set activity in the left hemisphere to 0
dummy.avg.pow(dummy.inside(sign(voxpos)==-1)) = 1; % set activity in the left hemisphere to 1

cfg = [];
cfg.parameters = 'all';

int_dummy = ft_sourceinterpolate(cfg,dummy,template_MNI_mri);

cfg = [];
cfg.funparameter = 'avg.pow';
cfg.method = 'ortho';

%plot the non-interpolated volume
figure;
ft_sourceplot(cfg,dummy);

%plot the interpolated volume
figure;
ft_sourceplot(cfg,int_dummy);

As you can see from the figure in the attachement, the voxels that
corresponded to the left hemisphere prior to the interpolation,
no appear to be located in the rigth hemisphere.

This is also something that I have observed in real data.

Does anybody have an explanation why this is happening?
Am I doing something wrong?

Best,

Fred


-------------- next part --------------
A non-text attachment was scrubbed...
Name: non_interp_vs_interp.jpg
Type: image/jpeg
Size: 25092 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120905/ab01371d/attachment-0001.jpg>