Hi Jaret,

In the first level you'd model the waveform associated with the task. Unless the treatment happened while the FMRI sequence was running, in principle the treatment (pre/post) would be investigated at the second level. So no variables should be used for pre or post at the first level.

All the best,


On 4 August 2017 at 18:30, Jaret Hodges <[log in to unmask]> wrote:
I tried to do a second level analysis comparing two sets of scans. Unfortunately, it said that my .feat did not have any associated statistics with them.

My design is as follows-
1. pre scan
2. treatment received
3. post scan

The variable of interest is the contrast between pre and post scan.

When I set up my first level, in my statistics set up, I included 1 explanatory variable (pre or post). I selected the shape of empty.

I believe what I should have done is select square and set the skip portion to zero (so that instead of having zero across all time points, it would be one across all time points)