I am really new to fMRI analysis and I wanted to ask a methodological question regarding preprocessing in fMRIPrep.
In my experiment I have to use a 64 channel coil and I plan to look at the whole brain. I am aware that this coil is not as good as a 32 channel in terms of homogeneity of the signal. However I have been informed that this can be accounted for by the scanner having intensity normalization during reconstruction.
My question is the following (on the basis that I will be using fMRIprep for the preprocessing):
Would the different reconstruction in a 64 channel coil pose an issue , when using fMRI prep for the preprocessing? Is there anything additional that I would need to take into account then?
Hi,
We are using the 64CH coil in our Center as well on a 3T Siemens Prisma System. I never used the 32ch myself but from different reports I recall that the signal homogeneity and signal is relatively similar between the two coils (The 64CH have extra coverage of the neck compare to the 32ch). We ran fmriprep on several studies with the 64CH coil and we don’t encounter any issue, the bias is well corrected by the bias correction algorithm (SPM segment, Ants N4BiasCorrection, HCP T1wxT2w procedure…). From what I recall the acceleration capabilities of the 32ch and 64ch are also quite similar.
In my opinion, the Siemens 20 channel head(16)/neck(4), 32-channel head, and 64-channel head(48)/neck(16) coils are all competent coils, and the choice between them is based on human factors:
The 20-channel has the most space inside. Unlike the Siemens 12-channel it was designed with multi-band in mind. It shows competitive SNR as long as the multi-band is restrained (e.g. 2-3), as shown for fMRI and DTI. This coil is ideal for participants where the other coils are too confining. For example, this reduces discrimination of people based on their weight. In my experience, it is better for clinical populations like stroke where obesity is a risk factor and kyphosis is a common consequence.
The 32-channel coil has a smaller cavity than the 20-channel. It is widely used and many sequences from experienced groups are tuned for this coil (e.g. ABCD and HCP).
The 64-channel coil has a clever aperture at the top that is ideal for simultaneous EEG. The internal cavity is the same size as the 32-channel, however the integrated posterior neck coil protrudes quite a bit. The consequence of this is that the head orientation is pitched upward. This rotation pushes the frontal lobes further away from the (smaller) anterior elements. Further, many individuals who can fit in the 32-channel can not fit in the 64 as their cheek bones protrude anteriorly. As I recall, historically this coil required bandwidth compression, but that may have been resolved.