help > RE: Multiple conditions for a contrast
Apr 6, 2021  04:04 PM | Y. Chang
RE: Multiple conditions for a contrast
Hi Alfonso,

Thank you so much. This is very helpful for me to understand how the contrast works!

best regards,
YaNing

Originally posted by Alfonso Nieto-Castanon:
Dear YaNing,

The last one is the correct one ([1 0 0 -1; 0 1 0 -1; 0 0 1 -1]), that will evaluate whether (A>baseline) OR (B>baseline) OR (C>baseline). An alternative option would be a contrast like [1/3 1/3 1/3 -1], which evaluates instead whether (A+B+C)/3>baseline. In general a contrast like [a b c d] evaluates an hypothesis of the form a*A + b*B + c*C +d*D >0 (so [1/3 1/3 1/3 -1] for example is evaluating A/3 + B/3 +C/3 -D > 0, which is just the same as (A+B+C)/3 > D), and a contrast with multiple rows simply combines the hypotheses of each row with OR conjunctions (so [1 0 0 -1; 0 1 0 -1; 0 0 1 -1] is just evaluating A-D>0 OR B-D>0 OR C-D>0, which is just the same as A>D OR B>D OR C>D)

Hope this helps
Alfonso
Originally posted by Y. Chang:
Dear all,

I am new to the CONN toolbox and have a question about setting a between-condition contrast at the second level analysis. I have four conditions (A, B, C, and baseline). A, B and C are the reading task. After the ROI-ROI connectivity analysis, I would like to check if there is an overall task effect in reading versus baseline (i.e., (A, B, C) > baseline). But I am not sure how to set the contrast. Shall I give a contrast, [ 1 1 1 -3] or [1 1 1 -1] or [1 0 0 0;0 1 0 0; 0 0 1 0; 0 0 0 -1] or [1 0 0 -1; 0 1 0 -1; 0 0 1 -1]? I have tried all of these options, and the results are different so I would like to make sure which one is correct (or none of them!)? I think that I have not had clear understanding about how the conditions can be combined so I would appreciate your help and suggestions!

best,
YaNing