Hi experts,
We are now analyzing resting-state data using melodic and dual-regression (and in turn randomize). Our design is a repeated measure (i.e., resting state after an experimental manipulation vs. resting state after a control manipulation on a different, counter-balanced day). We also have a physiological measure. We have already found that this physiological measure is changed as a function of our manipulation. Our question is whether changes in this physiological measure (slope?) are related to changes in resting-state ICs.
I’m not quite sure what design would allow us to answer this question. I realize that there have been many discussions in this forum about similar situations. Yet, it is unclear to me which one suits our question the most.
So far, I have three designs in mind. As a toy example, let says n = 4, and scores of the physiological measure for condition 1 = 1, 2, 3, 4, and for condition 2 = 2, 3, 4, 5. Thus the mean-centered values for this measure are -2, -1, 0, 1, -1, 0, 1, 2.
Design 1: Similar to below threads, each subject has his/her own group num, and his/her EV. Also, there is the EV1 (1, 1, 1, 1, -1, -1, -1, -1) indicating the manipulation condition. Crucially, there is also the EV2, which includes demeaned scores of subjects’ physiological measure multiplied with the EV1. The contrasts of our interest would be 1 and -1 at EV2, and 0 otherwise.
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1109&L=FSL&D=0&P=92340
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;4aad1d44.1011
group S1 S2 S3 S4 EV1 EV2
1 1 0 0 0 1 -2
2 0 1 0 0 1 -1
3 0 0 1 0 1 0
4 0 0 0 1 1 1
1 1 0 0 0 -1 1
2 0 1 0 0 -1 0
3 0 0 1 0 -1 -1
4 0 0 0 1 -1 -2
Contrasts = [[0 0 0 0 1 0], [0 0 0 0 -1 0], [0 0 0 0 0 1], [0 0 0 0 0 -1])
Design 2 is similar to Design 1, except that I also have EV3, which includes the demeaned scores of subjects’ physiological measures that are NOT multiplied with the EV1. Again, the contrasts of our interest would be 1 and -1 at EV2, and 0 otherwise. I think this model is similar to a “moderation” regression model used commonly in psychological research (i.e., main effect of a discrete variable (EV1), of a continuous variable (EV3) and of an interaction between the two(EV2)) .
group S1 S2 S3 S4 EV1 EV2 EV3
1 1 0 0 0 1 -2 -2
2 0 1 0 0 1 -1 -1
3 0 0 1 0 1 0 0
4 0 0 0 1 1 1 1
1 1 0 0 0 -1 1 -1
2 0 1 0 0 -1 0 0
3 0 0 1 0 -1 -1 1
4 0 0 0 1 -1 -2 2
Contrasts = [[0 0 0 0 1 0 0], [0 0 0 0 -1 0 0], [0 0 0 0 0 1 0], [0 0 0 0 0 -1 0])
Design 3 is similar to “Two Groups with continuous covariate interaction” in the glm manual http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/GLM#Two_Groups_with_continuous_covariate_interaction. That is, demeaned scores of subjects’ physiological measure without multiplication from condition 1 are in EV2, and those from condition 2 are in EV3. Then the contrasts of interest would be [1EV2 – 1EV3] and [-1EV2 1EV3].
group S1 S2 S3 S4 EV1 EV2 EV3
1 1 0 0 0 1 -2 0
2 0 1 0 0 1 -1 0
3 0 0 1 0 1 0 0
4 0 0 0 1 1 1 0
1 1 0 0 0 -1 0 1
2 0 1 0 0 -1 0 0
3 0 0 1 0 -1 0 -1
4 0 0 0 1 -1 0 -2
Contrasts = [[0 0 0 0 1 0 0], [0 0 0 0 -1 0 0], [0 0 0 0 0 1 -1], [0 0 0 0 0 -1 1])
Which one would be the most suitable? Are some of these equivalent to each other?
I am quite new to FSL and randomise. I would greatly appreciate your time and effort.
Thank you so much,
Narun
|