Combining participants with different slice counts

Dear all,

Would it be fine to combine subjects with different slice counts while doing analysis on SPM? Would it cause any issues particularly for the 2nd level analysis?

Thanks.

That depends…

For group analysis, different number of slices, but otherwise same voxel resolution may be treated as missing data in programs like 3dLME.

For my case, voxel sizes will be the same. I am planning of running either a multiband sequence with additional slices or decrease the TE from 30ms to 25ms on my 32 slice protocol to get a few more slices.

If say, I have 30 participants acquired on 32 slices and now start collecting data with more slices would it ruin anything?

I want to get more slices for full brain coverage (mainly anterior temporal lobes on huge brains)

I’m sure this is another “it depends”. The echo time is an important factor not just for coverage, but you can get differences in BOLD response using one versus another. If you use one echo time for one group and a different for another, you won’t be able to tell if the difference came from echo time differences. The results can also vary throughout the brain too. There may be more or less dropout or distortion. Multi-band studies may introduce similar kinds of issues. In any case, you will have to check for these kinds of differences and show no differences and justify your methods to reviewers.

https://doi.org/10.1016/j.neuroimage.2021.118617

1 Like

For my case this is a within subjects design, but there is a possibility of doing a between-subject comparison as well. So as you said, different TE’s across participants would cause issues. But what I understand is that there shouldn’t be any issues on a within-subject basis as long as I don’t compare groups with different slice settings?