Over at The Frontal Cortex, Jonah Lehrer muses on what he calls the “value of neuroscience”:
[L]earning about the brain can help constrain our theories. We haven’t decoded the cortex or solved human nature – we’re not even close – but we can begin to narrow the space of possible theories. We know, for instance, that the rational agent model of Homo Economicus isn’t particularly accurate, at least from the perspective of the brain, and that the deliberative prefrontal cortex is often out-shouted by emotional brain areas like the nucleus accumbens, insula, etc. This supports, of course, lots of observational studies that demonstrate that people rarely rely on explicit calculations of utility (or explicit calculations of anything, really) when making decisions. The anatomical details, in other words, can help settle the argument.
His answer is interesting and well thought out. The Popperian scientific dogma is that nothing can be proven but merely disproven, so we are in a sense with each study reducing the search space for possible theories.
However, this answer leaves out one critical component of neuroscience research: applications. Medically, fMRI and EEG studies are building empirical data from which scientists are extrapolating exciting possibilities like allowing immobilized but conscious patients to communicate. Due to the necessity of confirming that a patient is properly anaesthesized, even consciousness studies are becoming practical. So yes, constraining theory is one useful component of neuroscience research, but it is happily far from the only one.