I wouldn't exactly say "more significant activations", because it's hard
to say without knowing your experiment. But if you're doing e.g. a
passive viewing task and the contrasts you take to the 2nd level are
'visual stimulation > crosshairs', then yes you can say that.
Regarding the second part, let's say in your ROI group 1's beta is 5,
and group 2's is 3. Let's also say that your group 1 analysis shows a
cluster of activation in the ROI, while group 2 does not (group2 might
have much higher variance). When you do a 2-sample t-test, and look at
'group1 > group2', you cannot assume that your ROI should show a
significant result just from the single-group results.
Alvaro Muro wrote:
> Hi Chris,
>
> When you say:
>
>
>> Second, the 2 sample t-test [1 -1] is showing where the contrast value
>> for group 1 is significantly greater than for group 2.
>>
>
> does that mean that there is more significant activations in those areas in group 1 than in group 2?
>
>
>
>> You shouldn't
>> expect to see any differences just based on looking at the single-group
>> results.
>>
>
> That is because the 2 sample t-test is a random effects analysis which takes into acount group variablility, isn't it?
>
>
> Thank you very much for your kind help.
>
> Best wishes,
>
> Alvaro
>
>
>
>
>
>
|