Re: [Freesurfer] contrast matrix

Yes, that -BPM should be +BPM, sorry about that. It will not change the p-values if you multiply by 0.5 or not. It will change the size of the contrast. Most people don't report this number, especially for an F-test, so I would not worry about it. doug

On 11/20/2012 09:17 AM, xiangbo_2010 wrote:

Dear doug Thank you for your reply! +AOM -BOM -APM -BPM 0 0 0 0 0 +AOM -BOM 0 0 -AQM +BQM 0 0 0 +AOM 0 -APM 0 0 0 -COM +CPM 0 +AOM 0 0 0 -AQM -COM 0 +CQM there should be use 1 -1 or 0.5 -0.5? whether the -BPM should be change BPM? Thanks Bo Xiang

At 2012-11-19 07:23:31,"Douglas Greve" <greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu> wrote:

Hi Bo, you can think of the Ftest as a logical 'OR' between the
t-test contrasts indicated in each row. Each row is a difference
of differences, so

1. (A-B)om - (A-B)pm --> Does the difference between A and B
differ between groups O and P for Males?
2. (A-B)om - (A-B)qm
3. (A-C)om - (A-C)pm
4. (A-C)om - (A-C)qm
5. (A-B)of - (A-B)pf --> Does the difference between A and B
differ between groups O and P for Females?
6. (A-B)of - (A-B)qf
7. (A-C)of - (A-C)pf
8. (A-C)of - (A-C)pf

I've put together the first 9 columns of the first 4 rows. The
last 9 columns are all 0s. For the last for rows, the 0s and below
matrix are swapped to give you the same for the females

doug

AOM BOM APM BPM AQM BQM COM CPM CQM
--------------------------------------------------------------------------
+AOM -BOM -APM -BPM 0 0 0 0 0
+AOM -BOM 0 0 -AQM +BQM 0 0 0
+AOM 0 -APM 0 0 0 -COM +CPM 0 &nb sp;
+AOM 0 0 0 -AQM -COM 0 +CQM



On 11/17/12 9:21 PM, xiangbo_2010 wrote:

...

Hi Freesurfer experts,

    I'm very sorry to bother you, but I am very confused with the
    following questions:

    My experimental design includes three discrete factors:
    factor 1 with three levels (A,B,C ); factor 2 with three
    levels (O,P,Q); gender (F, M), and one covariate.

    So I can get 18 classes: FAO,
    FAP,FAQ,FBO,FBP,FBQ,FCO,FCP,FCQ,MAO,
    MAP,MAQ,MBO,MBP,MBQ,MCO,MCP,MCQ.I want to perform the
    interaction between factor 1 and factor 2 regressing out the
    effect of gender and one covariate, but I don't know the
    rules for setting the contrasts for the F-test. The contrast
    matrix I used is:

    1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    1 -1 0 1 -1 0 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    1 1 1 0 0 0 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    1 0 -1 1 0 -1 1 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    0 0 0 0 0 0 0 0 0 1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    0 0 0 0 0 0 0 0 0 1 -1 0 1 -1 0 1 -1 0 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    0 0 0 0 0 0 0 0 0 1 0 -1 1 0 -1 1 0 -1 0 0 0 0 0 0 0 0 0 0 0
    0 0 0 0 0 0 0

    is it correct?

    Any help will be very appreciated.

    Best wishes,

    Bo Xiang











_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer