Hi Felix,

> I am doing comparison between this HCP MEG preprocessed data and another 
> dataset.
> But that dataset (another one) is in the MNI space, and we computed the 
> leadfield with it's own surface (including faces and vertices) and mri data 
> which is in the MNI space.
>       2. But here the HCP MEG data is in the bti space, and I computed the 
> leadfield with the anatomy data which is in the bti space last time.
> It seems not fair to do the comparison between two datasets in different 
> space.  
> so I need to do the conversion.
> 
> And, I need the surface to do the comparison on the same vertices (or 
> regions) between the two datasets.


I don’t understand why the ‘difference in space’ is the key factor that holds 
you back from doing a structured comparison. You mention that the other dataset 
has (I guess) source-reconstructed MEG data defined on a surface topology that 
is different from the one used in HCP, so expressing the surface coordinates of 
the HCP MEG data in another coordinate system will not bring you much closer to 
making a easy comparison between the two datasets, without the need for an 
interpolation of any kind (although given the relatively low spatial resolution 
of MEG I don’t think that this is too big a problem).  

> Other problems:  "If you still think you want to do the conversion, you 
> probably would need to surface-resample (using wb_command -surface-resample) 
> the high-resolution surfaces (from the MNInonlinear ones in the structural 
> preprocessing package) to the 4k/8k resolution representation."
> 
> where is the structural preprocessing package? is it in the MEG preprocessed 
> dataset?

No, it’s in the output of the structural processing pipeline, as documented 
starting from page 21 in the 1200 subjects reference manual.

> the command "wb_command -surface-resample" is in the megconnectome or 
> freesurfer?

Neither, it’s in the workbench software. 
https://www.humanconnectome.org/software/connectome-workbench 
<https://www.humanconnectome.org/software/connectome-workbench>

Still, I think that this would be an overkill, and you might want to try to 
just use the subject specific *.surf.gii files that are in the MEG/anatomy/ 
package. As mentioned in the reference manual (p.48) (and as far as I remember) 
these are expressed in ACPC aligned RAS, and I’d say that this is close enough 
to what you want. To be very specific, you can do all of your MEG-data based 
processing using the anatomical data expressed in BTI-space (as in the *.mat 
files), and post-hoc replace the relevant information (i.e. the vertex 
positions) with the ones from the gifti-files.

Best wishes,

Jan-Mathijs


> 
> 
> 发件人: [email protected] 
> <[email protected]> 代表 Jan Mathijs Schoffelen 
> <[email protected]>
> 发送时间: 2018年6月21日 13:05:13
> 收件人: [email protected]
> 主题: Re: [HCP-Users] how to convert BTI space to MNI space
>  
> Dear Felix,
> 
> Could you try and explain why you think you need this conversion? The 
> transformation to MNI space is a non-linear one, which invalidates any 
> subsequent analyses on the MEG data. If you think you need it for 
> visualization purposes, the transformation should also not be needed, since 
> the individual vertices are surface realigned across subjects, which mean 
> that you can directly average/combine  across subjects.
> If you still think you want to do the conversion, you probably would need to 
> surface-resample (using wb_command -surface-resample) the high-resolution 
> surfaces (from the MNInonlinear ones in the structural preprocessing package) 
> to the 4k/8k resolution representation. For this you’d need the spherical 
> meshes that are in the megconnectome/template directory.
> 
> Best wishes,
> Jan-Mathijs
> 
> 
> 
> 
>> On 18 Jun 2018, at 07:26, Felix Yuan 
>> <[email protected] 
>> <mailto:[email protected]>> wrote:
>> 
>> Hi, HCP experts,
>> Sorry to bother you!
>> I know the anatomy data and MEG sensors are in the bti space of the MEG 
>> preprocessed data.
>> But now I want to convert the anatomy T1w data space and MEG sensor space 
>> toMNI space, how can I do this?
>>  
>> I also want to get the surface (vertices and faces) of MEG data on MNI 
>> space, could you help me with this?
>>  
>> Best wishes,
>> Felix
>> _______________________________________________
>> HCP-Users mailing list
>> [email protected] <mailto:[email protected]>
>> http://lists.humanconnectome.org/mailman/listinfo/hcp-users 
>> <http://lists.humanconnectome.org/mailman/listinfo/hcp-users>
> _______________________________________________
> HCP-Users mailing list
> [email protected] <mailto:[email protected]>
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users 
> <http://lists.humanconnectome.org/mailman/listinfo/hcp-users>

_______________________________________________
HCP-Users mailing list
[email protected]
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Reply via email to