functional brain imaging and native language

[gusl]

Around here, fMRI studies usually require subjects to be native English speakers, meaning they were immersed in it when no older than ~5 years old, regardless of their current level of fluency.

Clearly, the purpose is to eliminate an irrelevant source of variation (assuming the study is not about the effects of native-language). But why filter by native-language, rather than other cognitively relevant dimensions?

How can native-language affect the results of functional brain imaging studies? Is it because:

(1) non-native speakers, regardless of how well they perform on language tests, are never quite as fluent as native speakers, and wouldn't understand the instructions as well (or as quickly)

(2a) non-native speakers use a different part of the brain when interpreting spoken English (i.e. they use a "foreign-language region", rather than the "native-language region")

(2b) each language uses different physical regions in the brain (e.g. "Portuguese is a rather dorsolateral language")

(3) people raised in different languages will tend to perform the same tasks differently, i.e., even if you translated all instructions into the subject's native language, they would have a different way of performing them.

?