Hi Steve,
I have been reading all past messages on percentage signal calculation and
have some questions I wanted to get clarified:
1. Is calculating percent signal change for a [masked] ROI merely averaging
the cope (or pe) values for all voxels in the mask (multiplied by the
formula on the featquery page)? Or does it recompute the parameter
estimates from the mean raw time series of in the masked region?
2. Is the max_pe1_ts.txt the time series for the voxel with highest percent
signal change for the first PE in the mask? And the reason why this time
series could be different for pe2 (max_pe2_ts.txt) is because a different
voxel within the same mask has a higher value for the second PE?
3. What is in the mean_mask_ts.txt? Is this the average of the voxels in
the mean functional image that is used in percent signal change calculation?
4. Based on earlier posts on this topic, what does it mean to choose zstat
or tstat images as inputs for featquery? Does the output of featquery on
these images give a sense of whether the corresponding contrast in the
selected ROI is significant for each of the subjects queried on?
5. Finally, the voxel coordinates in standard space have decimal places.
Why is this so? Is this because my outputs from first level are not in the
standard space and my mask is in standard space. Should I convert my
outputs from first level to standard space before running feat query?
Thanks
Vinod
|