help
help > RE: Seed-to-voxel - determining effect sizes?
Jan 26, 2016 04:01 PM | Jon Dudley
RE: Seed-to-voxel - determining effect sizes?
Hi Jeff,
Did you only report that connectivity between a seed and some cluster(s) was different in one condition compared to another at some threshold of significance? If so I would assume that the reviewer wants to see the beta values from the seed-to-voxel analyses. In your second-level seed-to-voxel results window in CONN, there are two buttons on the lower right hand side:
"Display Values" will show you the effect sizes for your given contrast for each suprathreshold cluster. See attached, in this image I have one seed that has four clusters with significant between-group differences. The bar graph shows the between-group effect size (the fisher-transformed groupwise difference in connectivity) for each cluster and the table below gives the effect sizes/beta values as well as the T-statistic and p-values. The line graph at the bottom shows the subject-specific effect for the selected cluster (here I have twelve subjects and you can see subjects 1, 2, 4, 11, and 12 who make up group A have lower values than the other subjects in group B. The effect size of 0.39 in the table above is the difference in means of groups A and B).
"Import Values" will save the subject-specific effects as second-level covariates. From there it's easy to copy and paste the values elsewhere if you want to plot them out in another another fashion like a box-whisker.
Hope this helps,
Jon
Did you only report that connectivity between a seed and some cluster(s) was different in one condition compared to another at some threshold of significance? If so I would assume that the reviewer wants to see the beta values from the seed-to-voxel analyses. In your second-level seed-to-voxel results window in CONN, there are two buttons on the lower right hand side:
"Display Values" will show you the effect sizes for your given contrast for each suprathreshold cluster. See attached, in this image I have one seed that has four clusters with significant between-group differences. The bar graph shows the between-group effect size (the fisher-transformed groupwise difference in connectivity) for each cluster and the table below gives the effect sizes/beta values as well as the T-statistic and p-values. The line graph at the bottom shows the subject-specific effect for the selected cluster (here I have twelve subjects and you can see subjects 1, 2, 4, 11, and 12 who make up group A have lower values than the other subjects in group B. The effect size of 0.39 in the table above is the difference in means of groups A and B).
"Import Values" will save the subject-specific effects as second-level covariates. From there it's easy to copy and paste the values elsewhere if you want to plot them out in another another fashion like a box-whisker.
Hope this helps,
Jon
Threaded View
Title | Author | Date |
---|---|---|
Jeff Browndyke | Jan 26, 2016 | |
Jon Dudley | Jan 26, 2016 | |
Jeff Browndyke | Mar 6, 2016 | |
Jon Dudley | Mar 7, 2016 | |
Jeff Browndyke | Mar 7, 2016 | |
Jon Dudley | Mar 8, 2016 | |
Jeff Browndyke | Mar 8, 2016 | |
Jeff Browndyke | Jan 26, 2016 | |