I would like to calculate the tSNR. The analysis I am working on is trying to extract CerebroVascular reactivity from resting state fMRI so the preprocessing of the data is a little different.
I motion correct and blur the data. Then I linear detrend and bandpass filter. And finally I run the regression which includes 6 motion regressors and .

My question is if I want to calculate the tSNR do I use the signal that is motion corrected and blurred (but not linear detrended and bandpass filtered) and divide this by the residuals from the regression?

Thank you very much for any help in advance!
Carolin

For the record, it looks like you could run that pipeline using afni_proc.py, and it would not separate the bandpass step from the other linear regression. In that context, “-volreg_compute_tsnr yes” could be used to compute the TSNR after the volreg block, in addition to the one after regression.

As for which is better, I am not really sure. They are slightly different numbers. TSNR values reported by scanner folks are often coming from the unprocessed data. To me it makes more sense to at least run volume registration first. These steps affect the results. Volume registration should increase the values, as should the linear regression. One can make arguments for any of these. The most important thing is probably just to report which numbers the TSNR results are based on.

rick

The
National Institute of Mental Health (NIMH) is part of the National Institutes of
Health (NIH), a component of the U.S. Department of Health and Human
Services.