Hi Elizabeth - 

This doesn't really work - subtracting both ev1 and ev2 from ev3 will require effects in ev3 that are larger than ev1 and ev2 combined, which doesn't usually make sense.

Did you try contrast masking?




--

Eugene Duff, PhD
Analysis Group, Centre for Functional MRI of the Brain (FMRIB)
Nuffield Department of Clinical Neurosciences
John Radcliffe Hospital 
University of Oxford, OX3 9DU

Ph: +44 (0) 1865 222 523

--



On 9 November 2012 17:54, Woytowicz, Elizabeth <[log in to unmask]> wrote:

Hi,

 

Thank you for your guidance earlier with this question!  We applied the contrasts EV3-EV1.  However, from there we used fslmaths to do: contrast(EV3-EV1) – EV2, since we are interested in which regions are different or show greater activations.  Would this be a correct method?  Also, is there a way to use fslstats or another function to determine which of these activations are significantly greater or different?

 

I tried to sum up the main question below.

 

Contrast (EV3-EV1)      –     EV2                                     = X activations

(Showed significant        (subtracted significant        (how to determine

Differences)                       activations from EV2)          if these are significant activations?)

 

Thanks!

Liz

 

From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Eugene Duff
Sent: Sunday, October 28, 2012 7:26 PM


To: [log in to unmask]
Subject: Re: [FSL] 3 event contrast question

 

Hi Liz

 

On 28 October 2012 23:11, Woytowicz, Elizabeth <[log in to unmask]> wrote:

Hi Eugene,

 

Thanks for your quick response.  Is there a way to set up a contrast to get EV1-(EV2+EV3)?

 

 Well that of course defines a (not particularly interesting) contrast directly, but I'm guessing what you really want:  EV1 is stronger than both EV2 and EV3.  A straightforward approach is to define E1-E2 and E1-E3 contrasts, and use "contrast masking" to find the conjunction (where they are both true).

 

See:

 

Cheers,  

 

Eugene

 

Thanks,

Liz

From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Eugene Duff [[log in to unmask]]
Sent: Sunday, October 28, 2012 7:02 PM
To: [log in to unmask]
Subject: Re: [FSL] 3 event contrast question

Hi Liz,

This contrast will show you regions where there is a difference between EV1 and the average of EV2 and EV3.  Note that you can't interpret this contrast as one identifying regions where EV1 is stronger than both EV2 and EV3.

Eugene

On 28 October 2012 21:57, Elizabeth Woytowicz <[log in to unmask]> wrote:

Hi,

I am a graduate student new to FSL and wanted to make sure I was setting up my contrasts correctly for the question I was asking.

I want to create a contrast that would show significant activation in EV1 with EV2 and EV3 both subtracted from it, rather than individually.  Basically, EV1-(EV2+EV3).  I set the model as [2 -1 -1].  Is this correct?

Thanks!
Liz