I suggest looking at the raw images (dicoms) before preprocessing to narrow down the source of the problem. Some issues (e.g., https://mvpa.blogspot.com/2020/02/when-making-qc-or-most-any-images-dont.html, https://mvpa.blogspot.com/2017/09/voxelwise-standard-deviation-at.html) are much more obvious in unmasked raw images. Mango (and other software) lets you view a set of dicom images from a run as a movie, which can make it clear if there's only a handful of high-motion frames at fault or a bigger issue (e.g., https://mvpa.blogspot.com/2021/09/yes-im-still-glad-were-censoring-our.html). Checking if other participants have the same artifact also helps narrow down the issue.
good luck!
Jo
|