To expand what Gang said a little:
When you have a single group of subjects, then using 3dANOVA to compare results among a set of tasks is a generalization of a two-sample t-test (generalizing from two tasks to multiple tasks). This is the appropriate thing to use when the subjects in different tasks are different people.
Using 3dANOVA2 mixed effects is a generalization of a paired t-test. This is the appropriate thing to do when the subjects in different task are all the same people.
When you have different groups of subjects that you also wish to compare (e.g., men and women, patients and controls), then you get up to 3dANOVA3, which would allow for a layout like (tasks)X(subject type)X(subject) -- this is the 'nested' design.