[FieldTrip] ft_headmodelplot coordinate system origin problem
jan-mathijs schoffelen
jan.schoffelen at donders.ru.nl
Thu Jan 5 12:36:05 CET 2012
Hi Ion,
Did you use the mri with the correct coordinate system (i.e. MEG-based) for the interpolation?
JM
On Jan 3, 2012, at 2:22 PM, Ion Lavado wrote:
> Thank you very much Jan! Now everything fits using ft_headmodelplot(cfg). You solved our problem of the week.
>
> Just one final problem we have:
>
> - we are doing the beamformig source analisys.
>
> - our vol, grid and grad fits ok.
>
> -we have the sourcePre = ft_sourceanalysis(cfg, freqPre);
> sourcePost = ft_sourceanalysis(cfg, freqPost);
>
> -we then interpolate the data with
>
> cfg = [];
> cfg.downsample = 2;
> sourcePostInt = ft_sourceinterpolate(cfg, sourcePost , mri_real);
>
> - finally we use method SLICE to plot the results: (WITH MRI_REAL)
>
> %METHOD SLICE:
> sourceDiff = sourcePost;
> sourceDiff.avg.pow = (sourcePost.avg.pow - sourcePre.avg.pow) ./ sourcePre.avg.pow;
> cfg = []
> cfg.downsample = 2;
> sourceDiffInt = ft_sourceinterpolate(cfg, sourceDiff , mri_real);
> cfg = [];
> cfg.method = 'slice';
> cfg.interactive = 'yes';
> cfg.funparameter = 'avg.pow';
> cfg.maskparameter = cfg.funparameter;
> cfg.funcolorlim = [0 2];
> cfg.opacitylim = [0 2];
> cfg.opacitymap = 'rampup';
> figure('Name', 'SLICE Method')
> ft_sourceplot(cfg, sourceDiffInt);
>
>
> The problem comes now (not before), that when plotting the result don't fit the mri in the figure. I send you the .jpg so you can see it and try to help us whith this FINAL problem! In the script i send you, the problem i think starts after line 156.
>
> You are really kind, and thank you very much for your usefull help.
>
> Ion
>
>
>
>
>
>
>
>
> 2012/1/3 jan-mathijs schoffelen <jan.schoffelen at donders.ru.nl>
> Hi Ion,
>
> If the positive z-axis (denoted by +Z in the figure that is created by ft_determine_coordsys) is not pointing to the top of the head you made an error in ft_volumerealign and probably exchanged left with right in setting the lpa and rpa locations. This results in the coordinate system being a left-handed one and causes all kind of trouble later on. If this is the case, you need to repeat ft_volumerealign (with the original MRI) and type 'l' where you typed 'r', and the other way around. Ft_volumesegment relies on spm for the segmentation, which relies on the mri-data being defined in a meaningful right-handed coordinate system (which can be either mni-based, or MEG system based).
>
> If however the +Z is just pointing downward but the anatomy is 'upside-down' (neck pointing to the ceiling) then the coordinate system is a valid, i.e. a right-handed RAS (right-anterior-superior) coordinate system. If this is the case then I wouldn't know what's going on, because I would suspect that ft_volumesegment Perhaps some other neuromag aficionados could jump to the rescue here?
>
> As such ft_determine_coordsys does not change the coordinate system of your mri, it just changes the coordsys field, which helps FieldTrip in the interpretation of the real-world coordinates. The mapping of voxel locations to real-world locations is defined by mri.transform which can be changed by ft_volumerealign.
>
> Best wishes,
>
> JM
>
>
> On Jan 3, 2012, at 1:20 PM, Ion Lavado wrote:
>
>> The Z axis is pointing to the oposite direction (inferior), so i have the mri upside down. How can i change this? When Matlab plots the figure it asks me this:
>>
>> The positive x-axis is pointing towards the right
>> The positive y-axis is pointing towards anterior
>> The positive z-axis is pointing towards superior
>> Do you want to change the anatomical labels for the axes [y, N]?
>>
>> I then mark to change the axis and mark the positive z-axis as (i)inferior, but when plotting again is as the begginig... What do i have to do?
>>
>> Thank you very much Jan.
>>
>>
>>
>> 2012/1/3 jan-mathijs schoffelen <jan.schoffelen at donders.ru.nl>
>> Hi Ion,
>>
>> I'd suggest first to have a look at mri_real. How does it look in ft_determine_coordsys?
>> In particular, is the z-axis pointing in the correct direction?
>>
>> JM
>>
>>
>> On Jan 3, 2012, at 12:45 PM, Ion Lavado wrote:
>>
>>> Hi Jan,
>>> I am sorry, I have been trying with your method but now what is incorrect is the segmentation.
>>> Here is the script. Maybe you understand better the problem...I am getting crazy. The problem comes from line 53
>>> Thank you very much,
>>> Best,
>>> Ion
>>>
>>>
>>>
>>> 2012/1/3 jan-mathijs schoffelen <jan.schoffelen at donders.ru.nl>
>>> Hi Ion,
>>>
>>> What is the coordinate system in which the vol and mri are defined? To me it does not look like the neuromag coordinate system, but rather like MNI-space. I would expect the origin to be a bit (~4 cm) lower and a bit more anterior (because the origin is defined on the line going through the pre-auricular points. In the 'vol.jpg' it seems that the origin is in the anterior commissure. Although in principle it shouldn't matter, it is convention to express the coordinates in the MEG-system's coordinate system, i.e. in neuromag space in this case. Therefore I would say that the vol and grid variables need to be adjusted, rather than the sensor positions. As mentioned in my previoius message, this can be achieved by coregistering the anatomy to the neuromag coordinate system using ft_volumerealign. You need to manually identify the locations of the fiducials that define the coordinate system, i.e. the nasion, lpa and rpa. The required cfg is: cfg.interactive = 'yes', cfg.coordsys = 'neuromag'. Once the anatomy is coregistered you can proceed with creating the vol and grid variables.
>>>
>>> Best,
>>>
>>> Jan-Mathijs
>>>
>>>
>>> On Jan 3, 2012, at 11:38 AM, Ion Lavado wrote:
>>>
>>>> Thank you for your help Jan. I tried all you told but i continue with my problem. The coordinates for the head (volume) and mri are the same, with the function ft_determine_coordsys i see this.
>>>> I think that the problem is about the sensor (helmet) origin of coordinates. Please see the images i adjunt here, the sensors origin is very low(nearly out of the helmet) and when ploting everything i suppose that every origin align to each other, so that the brain comes a little bit low.
>>>>
>>>> Best whishes,
>>>>
>>>> Ion
>>>>
>>>> 2012/1/2 jan-mathijs schoffelen <jan.schoffelen at donders.ru.nl>
>>>> Dear Ion,
>>>>
>>>> Are you sure that the sensor positions are defined in the same coordinate system as are the volume conductor model and the source positions? If not, then this is the likely cause of the shift. You can have a look here for more information about the different coordinate systems: http://fieldtrip.fcdonders.nl/faq/how_are_the_different_head_and_mri_coordinate_systems_defined
>>>>
>>>> In order to coregister the anatomical mri to the sensor array, you can use ft_volumerealign.
>>>>
>>>> Best wishes,
>>>>
>>>> Jan-Mathijs
>>>>
>>>>
>>>> On Jan 2, 2012, at 5:11 PM, Ion Lavado wrote:
>>>>
>>>>> Hello, i'm working with MEG data from Elekta Neuromag. My problem comes when i try to plot the headmodel with the function ft_headmodelplot(cfg). This should plot the head model(volume) + brain grid + sensors. As you can see in the adjunted image, my head volume and grid is not centered and part of the head could not be read by the sensors. Is it possible to modify this so i can center the head in the sensors(helmet)?
>>>>> Thank you very much.
>>>>>
>>>>> Ion Lavado
>>>>> <headmodelplot.jpg>_______________________________________________
>>>>> fieldtrip mailing list
>>>>> fieldtrip at donders.ru.nl
>>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>>
>>>> Jan-Mathijs Schoffelen, MD PhD
>>>>
>>>> Donders Institute for Brain, Cognition and Behaviour,
>>>> Centre for Cognitive Neuroimaging,
>>>> Radboud University Nijmegen, The Netherlands
>>>>
>>>> Max Planck Institute for Psycholinguistics,
>>>> Nijmegen, The Netherlands
>>>>
>>>> J.Schoffelen at donders.ru.nl
>>>> Telephone: +31-24-3614793
>>>>
>>>>
>>>> _______________________________________________
>>>> fieldtrip mailing list
>>>> fieldtrip at donders.ru.nl
>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>>
>>>> <helmet.jpg><vol.jpg>_______________________________________________
>>>>
>>>> fieldtrip mailing list
>>>> fieldtrip at donders.ru.nl
>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>
>>>
>>> Jan-Mathijs Schoffelen, MD PhD
>>>
>>> Donders Institute for Brain, Cognition and Behaviour,
>>> Centre for Cognitive Neuroimaging,
>>> Radboud University Nijmegen, The Netherlands
>>>
>>> Max Planck Institute for Psycholinguistics,
>>> Nijmegen, The Netherlands
>>>
>>> J.Schoffelen at donders.ru.nl
>>> Telephone: +31-24-3614793
>>>
>>>
>>> _______________________________________________
>>> fieldtrip mailing list
>>> fieldtrip at donders.ru.nl
>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>
>>> <script.m>_______________________________________________
>>>
>>> fieldtrip mailing list
>>> fieldtrip at donders.ru.nl
>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>>
>> Jan-Mathijs Schoffelen, MD PhD
>>
>> Donders Institute for Brain, Cognition and Behaviour,
>> Centre for Cognitive Neuroimaging,
>> Radboud University Nijmegen, The Netherlands
>>
>> Max Planck Institute for Psycholinguistics,
>> Nijmegen, The Netherlands
>>
>> J.Schoffelen at donders.ru.nl
>> Telephone: +31-24-3614793
>>
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
> Jan-Mathijs Schoffelen, MD PhD
>
> Donders Institute for Brain, Cognition and Behaviour,
> Centre for Cognitive Neuroimaging,
> Radboud University Nijmegen, The Netherlands
>
> Max Planck Institute for Psycholinguistics,
> Nijmegen, The Netherlands
>
> J.Schoffelen at donders.ru.nl
> Telephone: +31-24-3614793
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
> <plot.jpg><script.m>_______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
Jan-Mathijs Schoffelen, MD PhD
Donders Institute for Brain, Cognition and Behaviour,
Centre for Cognitive Neuroimaging,
Radboud University Nijmegen, The Netherlands
Max Planck Institute for Psycholinguistics,
Nijmegen, The Netherlands
J.Schoffelen at donders.ru.nl
Telephone: +31-24-3614793
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120105/b877a79b/attachment-0002.html>
More information about the fieldtrip
mailing list