Strange error with fmriprep 1.4.0

I have been running fmriprep 1.4.0 on my data set for several months now, and recently have been encountering a weird error in which the bbregister looks like it is having serious issues (see images). When I re-ran fmriprep on some folks, it self-corrected on some of them, but others still continue to turn out like this. It generally only happens to 1 run of 1 task out of 2 sessions, and I can’t find any obvious errors in the raw data. Any suggestions on how to correct this or ideas of what is going wrong?

Thanks so much!
Lauren Ericksen
University of Utah