When I made my initial comments, I had only read the press release that Ivan linked to, and not the original article. I had misunderstood a sentence they wrote in it (even after reading the original article and going back, it took me another read through to realize what they meant by the sentence I misunderstood).
Broca's area is in the left frontal lobe, and it does look like there is decreased activity there during glossolalia (but I'm not an expert at reading these types of images...I generally know enough to get by with the brain areas that interest me, so I may be wrong). You're never going to see a complete deactivation, just a statistically significant decrease. That's the nature of this type of test.
I think the confusion is that they are talking about loss of control and language in the same sentence, but they haven't directly tested that it is loss of control of language centers. The "loss of control" is that blood flow in the prefrontal cortex is decreased, and decreases in activity in the PFC are associated with loss of inhibition.
Wernicke's area is more relevant to language processing, i.e., in understanding what someone else is saying (and is located close to an auditory center in the temporal lobe as well). So, in this case, it's Broca's area that's more relevant.
That's pretty much what they were using it as a measure of...just generalized brain activity.
In the original article, they are more specific about which of the basal ganglia differed in activity, and that was the caudate nucleus. This nucleus receives input from the frontal cortex, so the difference in activity seen there might just be related to the difference in activity in the frontal cortex. This region can also be involved in learning/memory, though I'm not sure how that's relevant here, unless it's as simple as being more active when singing because of the need to recall the words to the song. Lastly, there's a report in Science that the caudate is involved in distinguishing semantics in bilingual people, so maybe this difference in activity is associated with the glossolalia being like a second language to these people?
Here's the reference for the Science article:
J. Crinion, et al., Language Control in the Bilingual Brain, Science 312:1537 - 1540, 2006.
And the link to the abstract (you need a subscription or access through your library for the full article) http://www.sciencemag.org/cgi/conten...urcetype=HWCIT