Dear experts,
I'm working with a dataset that has intermittent RF artifact. I was wondering if people have experience with preventing intermittent RF artifacts as well as with correcting the images during processing.
The artifact is caused by our acute pain machine, however, since it's intermittent, the minute we start testing systematically, we can't replicate it. We had an electrical engineer test the equipment and the set up, and everything seems fine. The only thing he didn't do yet is test whether the filter is functioning the way it's designed to function. Does anyone have any words of wisdom as to how to proceed?
Since we already have some data with the artifact, I was hoping to get rid of it during processing. ICA denoising doesn't seem like a solution, as it can't separate the artifact from the data perfectly, and I can see some residual flicker after I filter the data. At the same time, the artifact is very obvious on the timeseries. For each voxel, there are several huge intensity spikes. Is there a way to interpolate each voxel's timeseries across those spikes? A nuisance regressor wouldn't really work since the spikes come at different timepoints for each voxel and are limited to a subset of voxels.
Thank you in advance for your advice,
-Marina
|