I have preprocessed this particular dataset, but each functional run has resting state data for a certain duration both prior and post task based data ( Extra 33 and 6-s rest periods were added to the beginning and end of each run). How to delete this particular segment of data and then preprocess the data? Also, I am assuming there shall not be any issue with this if I preprocess the data as it is, given that the steps performed is only subjective to the location and temporal dynamics and not the actual value inside each voxel. Will the outputs be different if the data is cleaned beforehand?
Hi @Satwick,
I don’t know what your analysis entails, but in many cases you don’t need to remove those time segments before preprocessing. You can just not model them in your GLM when statistically analyzing the data. For example, you can use --dummy-scans to create these regressors if you know how many TRs have data you do not want to analyze later.
Best,
Steven
So, in case of the dataset I am currently using TR is of 3s and 33s is the first block of resting state scan, I just need to ignore the first 11 acquisition slices. I am estimating GLM weights in python, so I while model estimation I just need to ignore the following slices, right?
Hi @Satwick,
Yes, you would just create 11 one-hot encoded vectors that regress out the first 11 volumes, for example:
1 0 0 …
0 1 0 ….
I was thinking to just chunk out the entire restin state duration using one regressor in the on-off design matrix ? would that work as well?