I understand that exclusionary criteria is not something that will be standardized, however, looking at my data, I’ve read a few papers that use “Fit Error %” as one parameter for inclusion/exclusion. I have Fit Error %'s on GABA+/Cr and Glx/Cr ranging anywhere from around 3 to 30%. I’ve read dozens of papers and none seem to mention any consensus (or near consensus) on acceptable Fit Error %. Could anyone point me toward any literature on the matter? I’ve read the MRSinMRS paper and other similar papers. Nothing on this though.
Hi @kickinghorse892,
It really depends on which brain region you collected your data, how big your voxel was, and for how long you scanned. You may use our Big GABA I paper and this paper to help guide you.
Another approach I recommend is to estimate the distribution of %FitErrors in your data to see what is the average error and what is an outlier. This can help you exclude datasets with high errors for your dataset.
Mark