Hello,

First of all I would like to thank you very much for developing such a great toolbox.

I wanted to see 1) the change in connectivity over time, and 2) the difference between groups in the connectivity change.
With the help of many questions here and Prof.Zalesky, I was able to build a design matrix for 2 group, 2 session (attached).
I was referring to the following link (https://www.nitrc.org/forum/forum.php?th...).

I have some questions here.

1. Did I build the interaction column correctly?

If so:
2. The interaction is determined by group x time, I was wondering how the interaction was determined here? (with control=0 and patient=1?)
3. if this interaction is significant, I'm wondering how to interpret it. I'm confused about time1 for control and time2 for patient being tied to the same 1.

Thank you so much for the help.

Chris

Hi Chris, 

the interaction column can be generated by multiplying (elementwise multiplication) the group column by the time column. 

Based on the xls spreadsheet, I don't think that the interaction column is correct. Multiplying the group and time columns is a simple way to compute the interaction column. 

The interpretation of the interaction is no different from any other statistical model. It is the extent to which the difference between groups is different at the two time points that you are considering.

andrew

Originally posted by Chris Pae:

Hi Andrew,

Thank you for the comment. I guess I was overthinking it.

According to your comment, only the "patient group" and "time=1" sections in the interaction column should be assigned a value of 1, while the rest should be 0. Please let me know if I have understood correctly.

Thank you again for your kindness in answering my very basic question. 

Chris

Yes - that sounds reasonable. 

Also, the interaction term can be modelled in several ways. 
Originally posted by Chris Pae: