Hi VM,

Please see below:

On 1 February 2017 at 02:38, neuroimage analyst <neuroimage.analyst@gmail.com> wrote:

Hi,

We extracted volume measures from 7 ROIs in 2 groups and compared
a) Mean of each ROI volume between group, and compared independently such as
i) whether mean of ROI 1 in group 1 is equal to mean of ROI1 in group 2
ii)whether mean of ROI 2 in group 1 is equal to mean of ROI2 in group 2 and so on

This means 7 comparisons, or 14 if you test both directions (tails) separately.

b) Plotted the mean ROI volume for 2 independent measures and compared whether the slope between and within group is significantly different. such as
i) Slope of ROI1 is positively associated with MoCA , age (grp1) ?
ii) Slope of ROI1 is negatively associated with MoCA, age (grp2) ?
iii) Slope of ROI1 vs MoCA (grp1) is significantly different than slope of ROI1 vs MoCA (grp2) ?
iv) Slope of ROI1 vs MoCA (grp2) is significantly different than slope of ROI2 vs MoCA (grp2) ?

These sound tests for main effects and interaction, with various combinations possible.

Q1) Do we have to perform multiple comparisons for (a) and (b)?

Yes.

Q2) If yes, do we correct across 7 measurements for (a) and 2 measurements for (b) or 7 measurements for both (a) and (b)?

You could put the values for the 7 ROIs in a table in .csv format, with one column per ROI and one row per subject. This would be the input data. Create a design matrix as:

EV1: 0 or 1 coding for group 1

EV2: 0 or 1 coding for group 2

EV3: MoCA group 1

EV4: MoCA group 2

EV5: Age group 1

EV6: Age group 2

etc

Then define a set of contrasts such as:

C1: [1 -1 0 0 0 0 ...] - This tests if mean for group 1 > mean for group 2

C2: [-1 1 0 0 0 0 ...] - This tests if mean for group 1 < mean for group 2

C3: [0 0 1 -1 0 0 ...] - This tests if the slope for MoCA for group 1 > slope for MoCA for group 2

C4: [0 0 -1 1 0 0 ...] - This tests if the slope for MoCA for group 1 < slope for MoCA for group 2

C5: [0 0 1 0 0 0 ...] - This tests if slope for MoCA for group 1 > 0

C6: [0 0 -1 0 0 0 ...] - This tests if slope for MoCA for group 1 < 0

etc

You'd run this analysis in PALM with something as:

palm -i input.csv -d design.mat -t design.con -corrcon -o myresults -logp

Hope this helps!

All the best,

Anderson

There is evidence to both performing and nonperforming multiple comparisons In the literature.

Any response (hopefully pointing to a literature) will be greatly appreciated.

Thanks

Regards

--VM

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

The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.