Hi tino, A t-test doesn't feel like the right thing, really. If you did a paired t-test and found significant results, that certainly doesn't say much for reliability.
But absence of a significant result doesn't assure reliability. A correlation does seem meaningful, but I'm not sure how to test significance with it. (Others may know.) Here's what little I do know -- the usage for this relevant caret_command option: caret_command -metric-correlation-coefficient-map <input-metric-file-name-A> <input-metric-file-name-B> <output-metric-file-name> For each node, compute a correlation coefficient from the node's values in the two input metric files. The two input files must have the same number of columns and column 'j' in the two files should contain data for the same subject. Donna On 11/04/2010 02:07 PM, Quintino Mano, Ph.D. wrote: > hello caret users, > > i'm interested in performing an inter-session comparison between two > functional maps. > > specifically: > > "picture > control" in session 1 (i.e., scanning session on day 1) > > COMPARED to > > "picture > control" in session 2 (i.e., scanning session 40 days later). > > i'm wondering how reliable these maps are across different scanning > sessions... so should i simply run a correlation or t-test? if so, can > this be performed using metric > calculations? > > can anyone think of a better way to answer the question of inter-session > reliability? > > many thanks in advance! > > tino > _______________________________________________ > caret-users mailing list > [email protected] > http://brainvis.wustl.edu/mailman/listinfo/caret-users > _______________________________________________ caret-users mailing list [email protected] http://brainvis.wustl.edu/mailman/listinfo/caret-users
