Hi Michael,
Great! After looking at it, I managed to create my own 2x3 ANOVA!
Thank you so much.
If I wanted to add a covariate that is stable (educational level), how would I do that without causing problems to the model?
Thanks!
Heidi
================================================
Dr. Heidi Jacobs
Postdoc researcher
Faculty of Health, Medicine and Life Sciences
School for Mental Health and Neurosciences
Division Cognitive Neuropsychiatry and Clinical Neurosciences
Alzheimer Center Limburg
[log in to unmask]
www.maastrichtuniversity.nl
www.heidijacobs.nl
Dr. Tanslaan 12, 6229 ET Maastricht
P.O. Box 616, 6200 MD Maastricht, The Netherlands
T +31 43 38 84 090 F +31 43 38 84 092
================================================
________________________________________
Van: FSL - FMRIB's Software Library [[log in to unmask]] namens Harms, Michael [[log in to unmask]]
Verzonden: maandag 24 juni 2013 16:10
Aan: [log in to unmask]
Onderwerp: Re: [FSL] Overall test for two groups - 3 conditions
Hi Heidi,
Turns out I posted on this recently (May 14, 2013) under the subject "2 x
4 repeated measures ANOVA".
For convenience, I'm reattaching the example design matrix.
For simplicity of illustration, the design matrix uses just 4 total
subjects, 2 in each group.
Columns 1-3 model the condition effects for G1 (group 1).
Columns 4-6 model the condition effects for G2 (group 2).
Columns 7-10 model the subject specific means.
"F1" represents the main effect of condition, and "F2" the condition x
group interaction.
In your case, the "condition" effect is "time", and you need to reduce
appropriately to just 3 (rather than 4) repeated measures.
To reiterate, the main effect of group would not be valid in this design,
and thus there is no such contrast included.
cheers,
-MH
--
Michael Harms, Ph.D.
-----------------------------------------------------------
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
660 South Euclid Ave. Tel: 314-747-6173
St. Louis, MO 63110 Email: [log in to unmask]
On 6/21/13 1:43 AM, "Jacobs H (NP)" <[log in to unmask]>
wrote:
>Hi Michael,
>
>No I could not find a 2x3 design, so I tried to combine the information
>from a 3x2 ANOVA design and a 2x2 design that I saw on the forum.
>I did read that the group effect is not valid and should be calculated
>afterwards with fslmerge and fslmaths (and than a simple two group
>comparisons).
>But I am specifically interested in the interaction.
>It would be great if you could send me your example, so that I know how
>to proceed with this.
>
>Thank you very much
>Heidi
>
>________________________________________
>Van: FSL - FMRIB's Software Library [[log in to unmask]] namens Harms,
>Michael [[log in to unmask]]
>Verzonden: vrijdag 21 juni 2013 3:32
>Aan: [log in to unmask]
>Onderwerp: Re: [FSL] Overall test for two groups - 3 conditions
>
>Which example on the GLM page are you using as a template? You need to
>model the repeated measures -- specifically, you need a 2 group X 3
>repeated levels design, of which there isn't an exact example on the GLM
>Wiki page. I had to do something similar recently, so I can send you an
>example design matrix after OHBM. Note though that only the effect of
>time, and the group X time interaction will be valid. The main effect of
>group will not be valid within that model. Given that, which specific
>effects are you trying to evaluate?
>
>cheers,
>-MH
>
>--
>Michael Harms, Ph.D.
>
>-----------------------------------------------------------
>Conte Center for the Neuroscience of Mental Disorders
>Washington University School of Medicine
>Department of Psychiatry, Box 8134
>660 South Euclid Ave. Tel: 314-747-6173
>St. Louis, MO 63110 Email: [log in to unmask]
>
>
>
>
>On 6/20/13 4:40 PM, "Jacobs H (NP)" <[log in to unmask]>
>wrote:
>
>>Hi Michael,
>>
>>Thanks for the feedback. I have been studying the website, but still find
>>it very complex.
>>Would this be a suitable design (I have more subjects, but just to make
>>it easy: 12 subject: 2 groups - 3 session each
>>
>> group Age session Interaction Mean
>>Input1 1 1 1 1 0
>>Input2 1 1 1 1 0
>>Input3 1 1 -1 -1 0
>>Input4 1 1 -1 -1 0
>>Input5 1 1 0 0 0
>>Input6 1 1 0 0 0
>>Input7 1 -1 1 -1 0
>>Input8 1 -1 1 -1 0
>>Input9 1 -1 -1 1 0
>>Input10 1 -1 -1 1 0
>>Input11 1 -1 0 0 0
>>Input12 1 -1 0 0 0
>>
>>
>>
>>Contrast F1 F2 F3
>>C1 0 1 0 0 0 x
>>C2 0 0 1 0 0 x
>>C3 0 0 0 1 0 x
>>C4 0 -1 0 0 0
>>C5 0 0 -1 0 0
>>C6 0 0 0 -1 0
>>
>>
>>I appreciate your feedback on this!
>>Thanks
>>Heidi
>>
>>================================================
>>Dr. Heidi Jacobs
>>Postdoc researcher
>>Faculty of Health, Medicine and Life Sciences
>>School for Mental Health and Neurosciences
>>Division Cognitive Neuropsychiatry and Clinical Neurosciences
>>Alzheimer Center Limburg
>>[log in to unmask]
>>www.maastrichtuniversity.nl
>>www.heidijacobs.nl
>>
>>Dr. Tanslaan 12, 6229 ET Maastricht
>>P.O. Box 616, 6200 MD Maastricht, The Netherlands
>>T +31 43 38 84 090 F +31 43 38 84 092
>>================================================
>>________________________________________
>>Van: FSL - FMRIB's Software Library [[log in to unmask]] namens Harms,
>>Michael [[log in to unmask]]
>>Verzonden: donderdag 20 juni 2013 17:32
>>Aan: [log in to unmask]
>>Onderwerp: Re: [FSL] Overall test for two groups - 3 conditions
>>
>>No, that design is not correct. See the GLM page on the Wiki.
>>
>>cheers,
>>-MH
>>--
>>Michael Harms, Ph.D.
>>
>>-----------------------------------------------------------
>>Conte Center for the Neuroscience of Mental Disorders
>>Washington University School of Medicine
>>Department of Psychiatry, Box 8134
>>660 South Euclid Ave. Tel: 314-747-6173
>>St. Louis, MO 63110 Email: [log in to unmask]
>>
>>
>>
>>
>>On 6/20/13 10:00 AM, "Jacobs H (NP)" <[log in to unmask]>
>>wrote:
>>
>>>Dear FSL-developers,
>>>
>>>I have fMRI data of 2 groups (each group 20 people) over 3 time points.
>>>If I want to perform a repeated measures test with a within variable and
>>>a between-group variable, would the mat and con file be ok?
>>>(G1_T1 = group 1, time 1). So, I guess I have to do an F-test statistic.
>>>
>>>I also want to correct for GM. However, I only have one T1-weighted
>>>image
>>>for person. How can I include grey matter as a covariate (I don't have
>>>120 files) without harming the model?
>>>
>>>Thanks
>>>Heidi
>>>
>>>
>>>
>>>
>>>================================================
>>>Dr. Heidi Jacobs
>>>Postdoc researcher
>>>Faculty of Health, Medicine and Life Sciences
>>>School for Mental Health and Neurosciences
>>>Division Cognitive Neuropsychiatry and Clinical Neurosciences
>>>Alzheimer Center Limburg
>>>[log in to unmask]
>>>www.maastrichtuniversity.nl
>>>www.heidijacobs.nl
>>>
>>>Dr. Tanslaan 12, 6229 ET Maastricht
>>>P.O. Box 616, 6200 MD Maastricht, The Netherlands
>>>T +31 43 38 84 090 F +31 43 38 84 092
>>>================================================
>>
>>
>>________________________________
>>The materials in this message are private and may contain Protected
>>Healthcare Information or other information of a sensitive nature. If you
>>are not the intended recipient, be advised that any unauthorized use,
>>disclosure, copying or the taking of any action in reliance on the
>>contents of this information is strictly prohibited. If you have received
>>this email in error, please immediately notify the sender via telephone
>>or return mail.
>
>
>________________________________
>The materials in this message are private and may contain Protected
>Healthcare Information or other information of a sensitive nature. If you
>are not the intended recipient, be advised that any unauthorized use,
>disclosure, copying or the taking of any action in reliance on the
>contents of this information is strictly prohibited. If you have received
>this email in error, please immediately notify the sender via telephone
>or return mail.
________________________________
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.
|
