help
help > RE: 2nd-level analysis
Jun 12, 2015 02:06 PM | Alfonso Nieto-Castanon - Boston University
RE: 2nd-level analysis
Hi Noah,
The "import values" option will always extract the original connectivity values entered into your second-level analyses for each subject (without any adjustment). When including multiple conditions or multiple sources (within-subjects terms) in your second-level analyses then:
a) for ROI-to-ROI analyses, the "import values" option will return a second-level covariate separately for each of the included conditions and source ROIs.
b) for seed-to-voxel analyses, the "import values" option will return a second-level covariate combining all of the included conditions and source ROIs using the corresponding between-conditions and between-sources contrast used in your second-level analysis (e.g. if you select conditions 'pre' and 'post' and enter a contrast [-1 1], then the returned connectivity values for each subject will represent the difference in connectiity between the post and pre conditions). You can change that behavior by clicking on the option that reads 'extract effect-sizes from this analyses' and selecting a different/alternative second-level analysis there if you instead would like to extract a different set of connectivity values within each cluster (in that case it will return the connectivity values representing the combination of conditions/sources specified in your alternative analyses instead of your current analysis). See for example this post https://www.nitrc.org/forum/message.php?msg_id=13352 for an example of how to use this to extract the connectivity values separately for each of your conditions.
Hope this helps
Alfonso
Originally posted by Noah Philip:
The "import values" option will always extract the original connectivity values entered into your second-level analyses for each subject (without any adjustment). When including multiple conditions or multiple sources (within-subjects terms) in your second-level analyses then:
a) for ROI-to-ROI analyses, the "import values" option will return a second-level covariate separately for each of the included conditions and source ROIs.
b) for seed-to-voxel analyses, the "import values" option will return a second-level covariate combining all of the included conditions and source ROIs using the corresponding between-conditions and between-sources contrast used in your second-level analysis (e.g. if you select conditions 'pre' and 'post' and enter a contrast [-1 1], then the returned connectivity values for each subject will represent the difference in connectiity between the post and pre conditions). You can change that behavior by clicking on the option that reads 'extract effect-sizes from this analyses' and selecting a different/alternative second-level analysis there if you instead would like to extract a different set of connectivity values within each cluster (in that case it will return the connectivity values representing the combination of conditions/sources specified in your alternative analyses instead of your current analysis). See for example this post https://www.nitrc.org/forum/message.php?msg_id=13352 for an example of how to use this to extract the connectivity values separately for each of your conditions.
Hope this helps
Alfonso
Originally posted by Noah Philip:
Hi Alfonso - this brings up a question regarding
second level analysis and extracted betas/Zs. When we are
performing an ANCOVA in conn, and then extract the betas, are we
extracting the "corrected" values, i.e., that have been subjected
to an ANCOVA, or are we extracting subject-level betas without
adjustment?
Thanks!
Noah
Thanks!
Noah
Threaded View
| Title | Author | Date |
|---|---|---|
| Ben Xu | Jun 10, 2015 | |
| Alfonso Nieto-Castanon | Jun 11, 2015 | |
| Ben Xu | Jun 12, 2015 | |
| Alfonso Nieto-Castanon | Jun 12, 2015 | |
| Noah Philip | Jun 12, 2015 | |
| Alfonso Nieto-Castanon | Jun 12, 2015 | |
| Noah Philip | Jun 12, 2015 | |
| Ben Xu | Jun 15, 2015 | |