Hi,
You can only attached very, very small files to any FSL list emails
(in order to prevent lots of wasted bandwidth to the hundreds of
recipients). However, I don't think it is crucial to see your design.
Steve's comments relate to the fact that when you talk about
"significant difference in volumes between group1 > group2
when compared to, group1 > group3" it sounds like you are
just looking at the total number of voxels declared significant
(even if these are all in the hippocampus). Hence you might be
seeing a different set of voxels in each case, and this is then
comparing what happens in one set of voxels with what happens
in a different set of voxels. That is, at different voxels (points)
in the brain. It is easy to see that different things may show up
in different voxels. However, even at the same voxels, the results
you get are sensible. Your main concern seems to be that
groups 2, 3 and 4 are not showing significant differences, yet
you see that the comparisons to group1 (i.e. group1>group2,
group1>group3, etc) differ considerably. This is fine as the
fact that group2>group3 is not _significant_ does not mean that
there is no difference there. It simply means that the difference
is not enough to be over a certain threshold. However, there
probably is a difference and that will make more or less voxels
in the group1 comparison get over the threshold. For example,
if the threshold was 3 (this is an example only) and your values
in groups 1, 2 and 3 were 3.5, 0.4, 0.6, then group1-group2
would be over the threshold, but group1-group3 would be under
the threshold, while group2-group3 is no where near the threshold.
Obviously changing these numbers a little would then give you
any combination of results where group2-group3 was not
significant but the comparisons with group1 were or were not.
I hope this helps explain things. What you are describing is
certainly possible and is in no way contradictory. It is likely that
a lot of it is because you are just looking at the total number of
voxels and hence comparing what is happening in different
voxels, as Steve said. However, the above situation with
respect to thresholding changes might also be happening.
All the best,
Mark
On 3 Nov 2010, at 14:25, Rajagopalan, Venkateswaran wrote:
> -----Original Message-----
> From: FSL - FMRIB's Software Library on behalf of Stephen Smith
> Sent: Wed 11/3/2010 12:10 AM
> To: [log in to unmask]
> Subject: Re: [FSL] VBM results
>
> Hi Steve,
>
> Thanks, my model is very simple (I haven't included any covariates or anything, not even gender or age) just comparing different groups. I am attaching my design and contrast matrix so that you can have a look at it. Our email system prevents attaching certain file extensions so design.mat file was renamed to design.m (sorry for any incovenience)
>
> I am not able to understand this "this is explained by having these different results at different points in the brain, so you're not really comparing the same comparisons in the different cases?". Can you please rephrase or give me little bit more explanation.
>
> Thanks
>
> Venkat
>
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library on behalf of Stephen Smith
> Sent: Wed 11/3/2010 12:10 AM
> To: [log in to unmask]
> Subject: Re: [FSL] VBM results
>
> Hi - I suppose either you have an error in your models and contrasts, or this is explained by having these different results at different points in the brain, so you're not really comparing the same comparisons in the different cases?
>
> Cheers.
>
>
> On 2 Nov 2010, at 14:58, Rajagopalan, Venkateswaran wrote:
>
>>
>> Dear All,
>>
>> I would like to clarify the following with regard to my VBM results. VBM was done to compare GM volumes in 5 groups say Control, group1, group2, group3 and group4. Group 1 to 4 are patient subgroups. The results are as follows (I am giving the contrast that i used and the results)
>>
>>
>> Contrast Results
>>
>> Control > group1 Significant
>>
>> Control > group2 NOT Significant
>>
>> Control > group3 NOT Significant
>>
>> Control > group4 NOT Significant
>>
>> group1 > group2 Significant
>>
>> group1 > group3 Significant
>>
>> group1 > group4 Significant
>>
>> group2 > group3 NOT Significant
>>
>> group2 > group4 NOT Significant
>>
>> group3 > group4 NOT Significant
>>
>> I would like to know whether my results are correct in the sense more areas (in the sense for instance if we take a same region let us say hippocampus more hippocampus volume atrophied in group1>group2 when compared to group1>group3) have shown significant difference in volumes between group1 > group2 when compared to, group1 > group3 similarly when compared to group1 > group2 and group1 > group3 results less regions are found to be significant in group1 > group4 comparison, since there is no significant difference between control and groups2,3 and 4 also no significant difference was observed when comparing group2 > group3, group2 > group4 and group3 > group4, I am wondering whether the results above of more regions showing significant difference in group1 > group3 comparison when compared to group1 > group2 and group1 > group4 is possible when no significant difference was observed between group2 >group3. If so can you please tell me what makes this possible with the statistical approach we are using like the point spread function used to estimate blobs to do FWE can be different or any other reason for this difference in (I don't have much background on VBM and statistical methods used). However the clinical symptoms in these groups 1, 2, 3 and 4 are different.
>>
>> Thanks
>>
>> Venkat
>>
>>
>> ===================================
>> P Please consider the environment before printing this e-mail
>>
>>
>>
>> Cleveland Clinic is ranked one of the top hospitals
>> in America by U.S.News & World Report (2009).
>> Visit us online at http://www.clevelandclinic.org for
>> a complete listing of our services, staff and
>> locations.
>>
>>
>> Confidentiality Note: This message is intended for use
>> only by the individual or entity to which it is addressed
>> and may contain information that is privileged,
>> confidential, and exempt from disclosure under applicable
>> law. If the reader of this message is not the intended
>> recipient or the employee or agent responsible for
>> delivering the message to the intended recipient, you are
>> hereby notified that any dissemination, distribution or
>> copying of this communication is strictly prohibited. If
>> you have received this communication in error, please
>> contact the sender immediately and destroy the material in
>> its entirety, whether electronic or hard copy. Thank you.
>>
>
>
> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
> Associate Director, Oxford University FMRIB Centre
>
> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> ---------------------------------------------------------------------------
>
>
>
>
>
>
> ===================================
>
> P Please consider the environment before printing this e-mail
>
> Cleveland Clinic is ranked one of the top hospitals
> in America by U.S.News & World Report (2009).
> Visit us online at http://www.clevelandclinic.org for
> a complete listing of our services, staff and
> locations.
>
>
> Confidentiality Note: This message is intended for use
> only by the individual or entity to which it is addressed
> and may contain information that is privileged,
> confidential, and exempt from disclosure under applicable
> law. If the reader of this message is not the intended
> recipient or the employee or agent responsible for
> delivering the message to the intended recipient, you are
> hereby notified that any dissemination, distribution or
> copying of this communication is strictly prohibited. If
> you have received this communication in error, please
> contact the sender immediately and destroy the material in
> its entirety, whether electronic or hard copy. Thank you.
>
|
