What, if any, is the connection between, on the one hand, the tCr normalized metabolite values (in the tCR column), and the metabolite levels in non-gray matter? Or are the metabolite values corrected for tissue composition to be found in any of the other data columns generated by Osprey? The data set in question includes voxel fraction values from co-registration and segmentation of the MRS voxel in the T1 weighted image
I’m not sure I understand the question, specifically, what you mean by “metabolite levels in non-gray matter”. Could you clarify?
The term “non-gray matter” comes from a referee who complained that we reported metabolite levels normalized against total creatine and that in doing so we were ignoring variations in tissue composition. I am not quite sure myself what he/she means by “non-gray levels”, maybe “tissue corrected water scaled” and/or “CSF water scaled” (?).
Thanks for clarifying. The .tsv files containing the string TissCorrWaterScaled are what you are looking for - they have been fully tissue- and relaxation-corrected according to the popular Gasparovic method.
The term “non-gray levels” is completely wrong and should not be used in any context. All metabolites appear in grey and in white matter alike. Ratios to total creatine self-correct at least for the amount of CSF in the measurement volume (edit: except for compounds like Lac, which can also occur in CSF). They are not perfect, but neither are water-scaled estimates, since they necessarily rely on literature values for water and metabolite relaxation and water content, and will therefore also be weighted by the inaccuracies in these values, and their potential for change in pathology and between individuals.
I think it is useful to interpret both tCr ratios and tissue-corrected estimates together and in context with the actual tissue fraction estimates.