It depends on what you are trying to do with the statistic. But at first glance, I would guess that it is not since the statistic (mean and stddev) will be a mixture of across subjects and time points. It will be hard to interpret

On 5/8/2020 2:22 PM, Nasiriavanaki, Zahra wrote:

Hi Freesurfer experts

I have extracted the time course of a task fMRI data in a specific contrast and within a specific ROI, in 7 subjects. Consider my specific ROI has 1200 voxels and I extracted the time course for a duration of 5 TRs. After extracting the timecourse, I have a matrix of 1200x5 for each subject. To calculate the average timecourse, I concatenated the matrices for the 7 subjects and made one matrix of 8400x5. Then I averaged across columns which gives me one mean per column. I also calculated the standard error for each column. My question is if this voxel wise way of averaging and calculating the standard error is statistically correct or not. I deeply appreciate any comments on this.

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