Hi,
Just a quick follow up.
We do see useful improvements from distortion correct but not in
all studies, although it doesn't seem to be bad. In fact, in the Functional
Image Analysis Contest (FIAC - OHBM2005) the main effect of interest
could only be recovered by standard GLM when fieldmap-based
distortion correction was used.
Therefore I would recommend it and I have seen useful improvements in
practice.
All the best,
Mark
Jesper Andersson wrote:
>Dear Gonzalo,
>
>
>
>> In which cases I must do a B0 Unwarping ?... Always ?... With all the
>>fMRI studies did in any MRI equipment ?... Or in some equipments only
>>(or some brands) ?...
>>
>>
>>
>
>I think it is probably the case that most people do not do B0 unwarping.
>The aim of the unwarping is to reduce the distortions in the EPI image,
>i.e. to render the images a little more like the "true" anatomy of the
>subject. The degree of distortion depends on where in the brain you are
>looking (bad distortions in frontal/orbito-frontal areas and in the
>temporal lobes) and on your scanner make and model (bigger distortions for
>old scanners, bigger distortions for higher fields).
>
>My personal take would be that if you have a "modern" scanner (no more
>than a couple of years old) and a reasonable field (no higher than 3T),
>are scanning with a 64x64 matrix and doesn't have a particular interest in
>the frontal or temporal areas you are probably OK not to do it.
>
>But, on the other hand one can always turn the question around and ask:
>given that it only takes a minute to acquire the fieldmap and that there
>are software out there that does the correction for you, why not just
>always do it?
>
>
>
>> Which is the difference between the fMRI results got with and without
>>doing the B0 unwarping in the same fMRI study ?...
>>
>>
>
>The main difference will be that any activations will be a little more
>accurately localised in relation to the true anatomy. There are
>indications (check Cusac et al) that your z-scores might increase in a
>multisubject study as a consequence of better alignment of activations
>across subjects, but I haven't personally ever seen that.
>
>Good luck Jesper
>
>
