In 3dttest++ afni while performing a correlation analysis of one factor in the whole brain, suppose cue reward activity and across participants interaction effect in behaviour, do we need to normalize the whole brain reward activity for each participant?
To clarify, is the estimated BOLD effect of cue reward activity your response variable, while a behavioral measure serves as the predictor? There is no need to normalize either variable. Instead, you can obtain the correlation by converting the t-statistic for the regression coefficient using the following formula
r = \sqrt{t^2 / (t^2 + DF)}
Gang Chen
Thank you for the heads up Gang!
Just curious, is there any situation in which a across participation Z-scoring
normalization would be required for the whole brain activity?
Discussing normalization in general can be challenging. Providing a specific, concrete example of why normalization is necessary would help improve the discussion clarity.
Gang Chen
1 Like