Thanks Dr. McLaren.
I think we will report the age-effect regardless of task first. Then we may also want to identify regions that show:
(1) task correlating with both age and performance (task & task corr age & tas corr performance triple conjunction)
(2) task correlating only with age (task & task corr age, excluding task corr performance)
(3) task correlating only with performance (task & task corr performance, excluding task corr performance)
I would like to know if these analyses are feasible if,
For (1) I put both age and performance as covariates in the same model and choose all 3 conditions and use conjunction
Since there is a .5 correlation between age and performance, I am also wondering if it is better to put them into different models.
For (2) and (3) since there is one inclusion and one exclusion, if we use a conjunction & mask approach, will the resultant t-values be accurate? If we instead save the t-map for these three and transform them to p maps, then do a multiplication and exclusion, will we get correct t-values?
Thank you very much for your help.
|