I concur with Donald McLaren.
I would add that the term deactivation rather refers to a contrast
task vs. fixation, not task vs. control (at least in the 'default
network system' literature). Besides, what you are testing here is the
interaction between questionnaire scores and task effects, not task
vs. whatever. Without looking at simple effects, you cannot easily
interpret the interaction.
If you want to know if they "activate more" or "deactivate less" in
the default network sense you should compute a statistic that
estimates the effect of task vs. fixation (if you have a fixation) and
then see if the areas detected in the interaction correspond spatially
to activations and deactivations in this contrast. I warn you however
that this is rarely done.
Best wishes,
Roberto Viviani
Quoting "MCLAREN, Donald" <[log in to unmask]>:
> I would use the terms less active and more active. From the data you
> presented, there is no way to know if both conditions activate relative to
> baseline or deactivate relative to baseline. If you knew what the tasks
> did, then you might be able to be more specific.
>
> On Tue, Feb 4, 2014 at 8:16 AM, L. Thomas <[log in to unmask]> wrote:
>
>> Dear all,
>>
>> I'm looking for help to interpret results from a fMRI study with block
>> design (experimental; control) correctly.
>>
>> I did the following:
>>
>> 1.) 1st level: contrast "experimental minus control"
>>
>> 2.) 2nd level: whole brain regression analysis with a psychological
>> parameter (continuous psychological parameter assessed by questionnaire) as
>> the predictor of interest. Results show positive correlation between
>> experimental-related activation (experimental minus control) in the insula
>> and the psychological parameter (the higher psychological parameter scores,
>> the more activation)
>>
>> 3.) Mean parameter estimates for the significant region (insula) from each
>> 2nd level subtraction result with marsbar (-> 1 value for each subject)
>>
>> 4.) SPSS: Correlation of parameter estimates with the psychological
>> parameter. -> Scatterplot: x-axis: psychological parameter; y-axis:
>> parameter estimates of the significant region)
>>
>> Result: The scatter plot shows a positive linear relationship between
>> parameter estimates and psychological parameter scores. The line goes from
>> parameter estimates -2.5 (people with very low psychological parameter
>> scores) to +2 (people with very high psychological parameter scores)
>>
>> Now I have troubles to interpret these results or I'm not really sure what
>> the "zero line" means?
>> I would say that subjects with higher scores in the psychological
>> parameter have more experimental-related (experimental minus control)
>> activation in the insula relative to subjects with lower scores. But can I
>> also talk about "activation" and "deactivation"? If the "zero line" means
>> baseline, I would say that a negative value indicates a relative
>> deactivation, while a positive value indicates a relative activation. Then
>> I would say that people with very low psychological parameter scores show a
>> deactivation in the experimental condition relative to the control
>> condition and people with very high scores show an activation in the
>> experimental condition relative to the control condition. But I'm really
>> unsure whether I interpret this correctly.
>
|