I am performing Seed to Voxel analysis with pre post conditions:
Subject effect selected as QA_Greymatter_vol ; subject-level-contrast 1
Only one seed is selected; between sources contrast 1
When I select Between condition (Post-Pre) contrast as eye(2) and open the result explorer:
I see multiple clusters
According to CONN FAQ:
[i]For seed-to-voxel or voxel-to-voxel analyses, typically the analysis results are considered appropriately corrected for multiple comparisons (across all brain/analysis voxels) as long as at least one of either the height (voxel-level) or the extent (cluster- or peak- level) thresholds uses an analysis-wise false positive control method (either FDR- or FWE- corrected p-values).[/i]
My questions are:
1. What does height and extent threshold means? When I change these thresholds the cluster sizes change why is that, how does the threshold affect clusters?
2. As said in the FAQ quoted above, results are appropriately corrected if one of the thresholds uses analysis wise false positive control: Does it mean that if I use FDR uncorrected for height threshold; the results are still appropriately corrected?
3. I am new to CONN. And I am a bit confused with what is the difference between F-statistics and T-statistics for ROI-ROI analysis? [F-test to determine whether group means are equal.t-tests are a type of hypothesis test that allows you to compare means] This is what I found about T and F test. I am not able to understand the results: what does higher/ lower F value means, what does +/- T value means in case of ROI1-ROI2?
Subject effect selected as QA_Greymatter_vol ; subject-level-contrast 1
Only one seed is selected; between sources contrast 1
When I select Between condition (Post-Pre) contrast as eye(2) and open the result explorer:
I see multiple clusters
According to CONN FAQ:
[i]For seed-to-voxel or voxel-to-voxel analyses, typically the analysis results are considered appropriately corrected for multiple comparisons (across all brain/analysis voxels) as long as at least one of either the height (voxel-level) or the extent (cluster- or peak- level) thresholds uses an analysis-wise false positive control method (either FDR- or FWE- corrected p-values).[/i]
My questions are:
1. What does height and extent threshold means? When I change these thresholds the cluster sizes change why is that, how does the threshold affect clusters?
2. As said in the FAQ quoted above, results are appropriately corrected if one of the thresholds uses analysis wise false positive control: Does it mean that if I use FDR uncorrected for height threshold; the results are still appropriately corrected?
3. I am new to CONN. And I am a bit confused with what is the difference between F-statistics and T-statistics for ROI-ROI analysis? [F-test to determine whether group means are equal.t-tests are a type of hypothesis test that allows you to compare means] This is what I found about T and F test. I am not able to understand the results: what does higher/ lower F value means, what does +/- T value means in case of ROI1-ROI2?