History of AFNI updates
Suggested reading for beginners
Add yourself to the AFNI map!
|
Author: Gang Chen (---.nimh.nih.gov)
Date: 01-26-09 13:03
> (UpLeft +UpRight + DownLeft + DownRight) > Stop
> (UpLeft + UpRight) > (DownLeft + DownRight)
> (UpLeft + DownLeft) > Stop
For each group separately or with both groups combined? And regardless the scanning time (Immediate, 6WEEK, and 6MONTH)?
Use 3dcalc to combine those components on the left side (and the right side for the 2nd contrast), and then run a paired t-test with 3dttest. If these contrasts are done with scanning time collapsed, you may want to average instead of sum those component because the control group does not have data for 6WEEK. Also if some subjects didn't go through a stimulus, unfortunately they have to be rid of from the analysis for these contrasts.
> Then I would like to examine whether the differences in these contrasts vary
> over GROUP at the two applicable timepoints after treatment (Immediate
> and 6Month), for instance:
>
> Controls((UpLeft + DownLeft) > Stop) > Patients((UpLeft + DownLeft) > Stop)
Regardless the scanning time? If so, calculate the average (UpLeft + DownLeft - Stop) across the scanning times for each subject using 3dcalc, and run a two-sample t-test with 3dttest. Again if some subjects didn't go through a stimulus, unfortunately they have to be dropped from the analysis for these contrasts.
> Then I would like to do a GROUP x TIME interaction effect for the contrasts
> listed above (Including the 6Week scan that the Patients got but the Controls didn't?).
You mean an overall F test? If so, you may have to use 3dLME then.
> Then I would like to look within the Patient group to try and characterize a
> timecourse of recovery from treatment, asking specifically if the change is
> linear over time after treatment:
>
> Patient-Immediate(SomeContrast) > Patient-6Week(SomeContrast) > Patient-6Month(SomeContrast)
For each patient, get SomeContrast using 3dcalc for Immediate and 6Week separately, then run a paired t-test for Patient-Immediate(SomeContrast) > Patient-6Week(SomeContrast). Again if some subjects didn't go through a stimulus, unfortunately they have to be dropped from the analysis for these contrasts.
Do the same for Patient-6Week(SomeContrast) > Patient-6Month(SomeContrast)
HTH,
Gang
|
|
|
Daniel Schwartz |
01-23-09 17:54 |
|
Gang Chen |
01-25-09 12:21 |
|
Daniel Schwartz |
01-26-09 11:56 |
   |
Re: A rather painful study design... |
|
Gang Chen |
01-26-09 13:03 |
