Dear Matthias,
You can but you have to take into account that the resels per voxel are
estimates themselves, subject to noise. If you assume stationarity and
average over voxels, you will have a very precise estimate, otherwise
have a look at the following article to see how it can be taken into
account:
http://www.fil.ion.ucl.ac.uk/spm/doc/papers/HayasakaNichols_NonstCls.pdf
Another option might be to locally smooth the RPV image and proceed the
way you described, see spm_est_smoothness.m lines 206-230 and spm_list.m
lines 383-410.
Best regards,
Guillaume.
On 20/01/17 07:54, Stangl, Matthias /DZNE wrote:
> Dear Guillaume,
>
> thank you for your reply. This is already helpful for me, but I would have
> a follow-up question on that.
> After a GLM has been estimated, is it also feasible to just read the
> values within an ROI of the RPV.nii image that is automatically saved? For
> my understanding, this tells me the resels-per-voxel for all my voxels
> within the ROI. Could I not simply add all the resels-per-voxel estimates
> within the ROI to get the ROI resel count? However, I tried this and the
> values I get seem to be roughly but not exactly the values that SPM prints
> out when doing a small volume correction.
> Would it nevertheless be feasible to take within-ROI values from RPV.nii
> as a measure of smoothness of individual voxels and further derive the
> number of resels within the ROI from RPV.nii?
>
> Best,
> Matthias
>
>
>
>
>
>
>
> On 19/01/17 11:50, "Guillaume Flandin" <[log in to unmask]> wrote:
>
>> Dear Matthias,
>>
>> Have a look at spm_VOI.m, the SPM function where "small volume
>> correction" is implemented. It calls spm_resels.m that will return the
>> Resel counts given a search space and the estimated smoothness (assuming
>> stationarity).
>>
>> Best regards,
>> Guillaume.
>>
>>
>> On 19/01/17 08:47, Stangl, Matthias /DZNE wrote:
>>> Hi all,
>>>
>>> I am currently looking for a way to correct the results of a statistical
>>> test for spatial smoothness of the (fMRI) data. This should be done
>>> outside SPM using custom Matlab code. However, I would need to calculate
>>> the number of resels within an ROI for that purpose.
>>> Is there any way (ideally an SPM function that I could call) which
>>> calculates the number of resels within an ROI/volume in SPM?
>>>
>>> I would be very grateful if someone knows an easy way how to calculate
>>> the number of resels or how to get to this value in SPM.
>>>
>>> Thank you,
>>> Matthias
>>>
>>> --
>>> Matthias Stangl
>>> Aging & Cognition Research Group
>>> German Center for Neurodegenerative Diseases (DZNE)
>>> Leipziger Str. 44, 39120 Magdeburg, Germany
>>> phone: +49-391-67-24520
>>> email: [log in to unmask] <mailto:[log in to unmask]>
>>> website: www.wolberslab.net
>>>
>>> --
>>> Matthias Stangl
>>> Aging & Cognition Research Group
>>> German Center for Neurodegenerative Diseases (DZNE)
>>> Leipziger Str. 44, 39120 Magdeburg, Germany
>>> phone: +49-391-67-24520
>>> email: [log in to unmask] <mailto:[log in to unmask]>
>>> website: www.wolberslab.net
>>
>> --
>> Guillaume Flandin, PhD
>> Wellcome Trust Centre for Neuroimaging
>> University College London
>> 12 Queen Square
>> London WC1N 3BG
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|