I am designing an fMRI study for my dissertion and I got consfused as I got to this part. Since, I have to signify the study variables one by one in a chart and provide theoritical and operational definitions for them, I recognized that I need to clarify it to myself.
I have got 2 groups of subjects alzheimers disease subjects and a healthy group matched. I aim to COMPARE the brain activation dedicated to a specific cognitive function between these 2 groups.
What would be my fMRI variables (dependent, independent, etc) ? Is it like the BOLD signal or something?
Thanks for your kind guidance
If you need to define the scope of your variables, from a linguistic point of view (and language is a good access door to the mind), I would suggest that predication (attention-driven phrases + peripheral information) might reveal to be an interesting approach, since predication variables (well defined) are reach in predictive problems (such as (a) both levels of meta-information, (b) ortho-information and even (c) para-information).
Your dependent variable, or the one in which you expect to see "changes", would be, effectively, the BOLD signal.
Your cognitive task is "manipulating" the activity of the brain, therefore, this one is the independent.
André Wlodarczyk Smart yet complicated! Thanks for sharing your answer.
Jonatan Ottino-González Thanks a lot for your answer, Jonatan! this is exactly what I thought that I have to mention in my Thesis proposal. But actually I've been told that I'd better mention z-score, number of voxels and location (x,y,z) as my variables. any ideas?
ok, I think now I understand your question. They're asking you to determine which are going to be the outcomes from the statistical analysis. This is, how are you going to report your results.
In that case, yes, either z or t-scores are the ones normally reported. Sometimes the beta-weights if you've conducted a regression. Coordinates are frequently reported in MNI space. And typically along with the Z, p-value and coordinates one would like to report the cluster size (the extension of the spatial map of "activation") in number of voxels.
So t/Z-score, p-value (adjusted or not for multiple comparisons), cluster size and coordinates of the peak in MNI coordinates.
Parnian Rafei If you tell me what is the title of your study, perhaps I will be able to explain my suggestion in more detail.
Jonatan Ottino-González Thanks a million Jonatan. your second answer was the exact answer to my question ! I was wondering how should I report the statistical analysis of my study.
- Badges
- Science topic