Can neural evidence refute behavioral findings?

I've been thinking about this question lately: can neural evidence (e.g. from fMRI, EEG, TMS...) refute behavioral findings? It seems that the primary function of imaging studies is to explain the neural processes of certain well-established behavioral phenomena, but it's rarely the other way around. Is it even possible to debunk the behavioral data?

Brain-based lie detection

Sujeeta Bhatt and colleagues recently scanned 18 participants using fMRI and compared their brain activity across 3 conditions:

1). when the participants pointed out truthfully which face they'd seen earlier in a line of 3 faces
2). when they lied and pointed to a new face rather than the one they'd seen earlier
3). when all the faces were new but the participants lied and pretended to have seen one of them before

Bhatt et al. found that when people were lying their dorso- and ventro- lateral prefrontal cortices, the superior frontal gyri, the anterior cingulate gyrus, as well as the precuneu in the parietal lobe showed increased activity.

Awesome!

Deep brain stimulation and financial reward

This is an impressive paper by Kareem Zaghloul et al. at UPenn, in that they showed midbrain dopaminergic neurons (specifically in substantia nigra) encode unexpected financial rewards. The most remarkable part is that they used deep brain stimulation (not fMRI!) in Parkinson's patients during a probabilistic learning task. It's indeed an important contribution to reinforcement learning.