Cluster correction for the group-level stat after multivariate analysis

AFNI Message Board
1

Dear AFNI experts,

I have a question how to go about FWE cluster-wise correction after a mild smoothing applied only for the group analysis.

I ran a multivoxel pattern analysis on a native space (without any spatial smoothing), but I wanted to apply a very mild smoothing just prior to the group analysis (i.e., smoothing right after spatial normalization step).
So in this case, I would like to know how I should run 3dClustSim. I read that -acf option takes residual time series (in the standard space), and in my case I only want the smoothing for the group-level… Should I regenerate residual time series in non-smoothed native space data, and run 3dFWHMx to get parameters for acf?

Also, I make use of 3dLME because of some covariates with multiple factors (but also sometimes 3dttest++).

Thanks for your help,

Best, Michelle.

2

Michelle,

If you have the most recent version of AFNI, there is an option -resid in 3dLME that allows you to output the residuals at the group level. Then you can use 3dFWHMx to estimate the ACF values based on the group residuals.

3

Dear Gang,

Thank you very much for your input. I have been using afni version from this year January, 2016. I am trying to run with the new version. I hope the update does not affect too much of the prior analyses that I did.

Best,
Michelle