Hi - a 3-group (e.g. 3-EV) simple GLM modelling will allow simple
voxelwise ("univariate") testing between the groups.
If you want to look for group effects interacting over multiple voxels
then yes you'll need a multivariate test. For example, you can feed
TBSS-preprocessed data into MELODIC.
Cheers.
On 4 Jun 2008, at 00:11, Longchuan Li wrote:
> Hi Steve
>
> Sorry for the confusion I brought in the previous mail.
>
> My situation is like this: I have three groups: control, grp1, which
> has the subjects with severe disease A, and grp2, which include
> subjects with less severe degree of disease A. Also, for each group,
> there are six region of interests. The aim of this analysis is to
> find out which group is significantly different from the other
> groups. So, my questions are like this:
>
> 1) if it is necessary to use multivariate tests to find out the
> relations among different ROIs?
>
> 2) After ANOVA (or MANOVA), if it is necessary to use multiple
> comparison correction?
> One opinion is that it is not, as it usually has less sensitivity
> and pair wise comparison between each pair should be enough;
> However, I tried both and found that the results with and without
> multiple comparison corrections are quite different, indicating the
> differences among groups might not be very significant.
>
> I really want to hear some suggestions on this kind of analysis
>
> Your help is greatly appreciated
>
> Longchuan
>
> ----- Original Message ----- From: "Steve Smith"
> <[log in to unmask]>
> To: <[log in to unmask]>
> Sent: Tuesday, June 03, 2008 3:44 PM
> Subject: Re: [FSL] about statistics for three groups
>
>
>> Hi - I'm not quite sure what you're asking - but yes you can put 3
>> groups into TBSS without problem.
>> Cheers.
>>
>>
>> On 3 Jun 2008, at 17:39, Longchuan Li wrote:
>>
>>> Hi Steve,
>>> Thank you for the reply. Also, if I use these three groups to do
>>> TBSS analysis, instead of ROI, do I still need to separate them
>>> instead of putting them together?
>>>
>>> Thank you
>>>
>>> Longchuan Li
>>> ----- Original Message ----- From: "Steve Smith" <[log in to unmask]
>>> >
>>> To: <[log in to unmask]>
>>> Sent: Tuesday, June 03, 2008 3:34 AM
>>> Subject: Re: [FSL] about statistics for three groups
>>>
>>>
>>>> Hi,
>>>>
>>>> I would just run separate (t-test) comparisons for each pair of
>>>> groups - this is sensitive to any group differences and very
>>>> easy to interpret.
>>>>
>>>> Cheers.
>>>>
>>>>
>>>>
>>>> On 2 Jun 2008, at 18:46, Longchuan Li wrote:
>>>>
>>>>> Dear FSL experts
>>>>>
>>>>> I have a question regarding the statistics of multiple group
>>>>> analysis. Right now, I have three groups, group 1 has the
>>>>> patients with the most severe state of disease A (grp1), group
>>>>> 2 has the patients with less significant degree of disease A
>>>>> (grp2) and group 3 are unexposed controls (grp3). What is the
>>>>> best statistics if I want to see the differences among the
>>>>> three groups? i.e., if grp1 and grp2 are significant different
>>>>> from grp3 and if there is difference between grp1 and grp2.
>>>>>
>>>>> I am using one-way ANOVA to find out if there is any
>>>>> difference among three groups and then use multiple comparison
>>>>> to pick out which group is different. But I was also suggested
>>>>> that two sample t tests between each group might be a better
>>>>> choice.
>>>>>
>>>>> Your help is greatly appreciated.
>>>>>
>>>>> Longchuan
>>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------------
>>>> 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
>>>> ---------------------------------------------------------------------------
>>>>
>>>>
>>>
>>
>>
>> ---------------------------------------------------------------------------
>> 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
>> ---------------------------------------------------------------------------
>>
>>
>
---------------------------------------------------------------------------
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
---------------------------------------------------------------------------
|