Hello I have just carried out my first fMRI study and am now in the process
of trying to analyse the data. I would really like to check that the
analyses I have planned are correct.
The study employed a within-subjects design: each participant completed 2
runs of the task in the drug condition and the placebo condition (total of 4
runs per participant). There were two different orders for these runs
(Order1 Run 1 and Run 2 & Order2 Run 1 and Run 2) and the stimuli were
different for run1 and run 2 for both orders. Half participants did Order 1
on both test days and the other half did order 2 on both test days.
First I plan to conduct a series of cluster-based whole brain analyses and
have decided to go through the following stages:
1. Run separate lower level feat analyses on each of the 4 runs for each
participant – set up 6 Evs and contrasts of interest – which will be the
same for each run. The only difference for each run is the timing of the
blocks/EVs.
2. Combine runs for drug and placebo day for each subject – using
multi-session option/example in Feat – this will give me a drug gfeat. and a
placebo gfeat. for each participant – does it matter at this stage that
block order and stimuli of the two runs are different – the contrasts are
exactly the same so I am guessing its fine to combine them in this way?
3.Combine gfeats. to get a group average for drug and a group average for
placebo. By doing this I will be able to look at the effects of the drug and
the effects of placebo alone for the contrasts of interest.
I am slightly confused here as to whether to feed the inputs in at this
stage as lower level feat directories as specified in the multi-subjects
example: 'Select Inputs are lower-level FEAT directories and select the 5
relevant directories created at second-level, named something like
subject_N.gfeat/cope1.feat/stats'.
Or as individual 3D Cope images from the second level analysis (i.e.,
cope.nii.gz). I am not sure what the difference is between the two inputs –
I understand it has something to do with the 'timepoints' at which the data
is feed in but wondered if anyone could give a further explanation?
4.Next compare the effects of drug and placebo using paired t-test set up
example in Feat – at this stage should I input the individual 3D cope images
that I am interested in rather than using the lower level feat directories
option mentioned earlier- not sure what I base this decision on?
I would really appreciate any suggestions or comments anyone. Does this
sound like a suitable method for analysing the data given the study design?
I would really like to know if I am on the right track.
Many thanks
Jess
|