When we design fMRI tasks, typically we want both groups to be able to perform the tasks so that we're actually looking at activity in the brain related to the type of processing we're interested in. We design the tasks so that they're not necessarily simple, but doable. We want the performance to be similar. That's the strength of fMRI, that they can do the task, they're successful at doing the task, with the same reaction time and the same errors, but their brains are doing something different. They're having to compensate. They're having to use more brain resources to perform the tasks.
This is the challenge, that when you get out into the real world and the tasks are more difficult, that compensation might not be good enough, essentially.