Researchers gave each of eighteen subjects a playing card, then offered them money to lie to a computer about the identity of the card while undergoing a scan by functional magnetic resonance imaging (fMRI).
Conclusion (this supports the idea)
fMRI can eventually form the basis of an effective lie detector.
Evidence
When 18 subjects lied about a playing card, fMRI scans showed increased activity in regions of the brain known to be stimulated by stress.
Evaluate
The evidence tells us that the fMRI picked up on activity in stress-related brain regions. The conclusion i s claiming the fMRI is detecting a lie.
So we need to assume that activity in these regions clearly signals a lie. The author seems to be thinking,
But does lying cause stress? If not, this test could provide false negatives, because people would be lying but it wouldn't detect them.
And what if other things cause stress, even though the person is telling the truth? Or what if other things besides stress cause these brain regions to light up? If so, this test could provide false positives, because it might look like they're lying even though they aren't.
Goal
Since we want to weaken, let's look for a way to say that this test wouldn't be reliable. It would provide false negatives or false positives.