Hello everyone!

I've got data from 30 subjects concerning two states (200 volumes each): resting-state and a focused attention state. I intend to study the correlation maps generated in First-Level Analysis. However, I now noticed that two different functional sessions (which would be the 2 states) cannot be defined as two different conditions in Setup.
Is there a way to study these two different states without defining as if it was acquired from different subjects?

Hi Gustavo

That is perfectly fine, you can define two different functional sessions as two different conditions. Simply define two conditions (e.g. 'rest' and 'attention') and then associate the 'rest' condition with the first session (in Setup.Conditions select the "rest" condition, then select "session 1" and enter onset "0" and duration "inf", and then select "session 2" and leave the onset and duration fields empty there) and the "attention' condition with the second session (in Setup.Conditions select the "attention" condition, then select "session 2" and enter onset "0" and duration "inf", and then select "session 1" and leave the onset and duration fields empty there).

In general, leaving the onset/duration fields empty (or entering [] -empty brackets- in these fields) for some subject/condition/session simply indicates CONN that this condition is not present in this session/run for this subject. 

Hope this helps
Alfonso


Originally posted by Gustavo Pamplona:

Thanks, Alfonso! It worked fine!

Hi Alfonso,

We have one scan with multiple blocks. We inputted it into conn similarly to what you advised above. When looking at the second level analysis, I see under "Targets" multiple "Effects of...[each condition]". What exactly does this mean? 

Thanks!

Hi Kaylah,

The "Effect of [condition]" regressors represent the shape of the expected BOLD response associated with each condition (block timeseries convolved with the hemodynamic response function). You should typically include those in the 'Confounds' list in Denoising, so that those effects are removed from the BOLD timeseries of interest before proceeding. If you do so they will then not appear listed as potential sources in the first-level analysis tab (any ROI, condition effect, or first-level covariate that is not included in the 'Confounds' list is passed as a potential seed for seed-to-voxel or ROI-to-ROI first-level analyses). 

Hope this helps
Alfonso 
Originally posted by Kaylah Curtis: