Below is a comment I left, in a comment thread to a characteristically wonderful blog post by Eric Macdonald. It's pretty long so I'm reproducing it here.
Eric, I feel that Daniel Dennet won the argument on sui generis characteristics of subjectivity (or lack thereof) back when he wrote Consciousness Explained. For me, the main takeaway was that philosophers mistake failures of imagination for insights into necessity, and pre-emptively draw up limits to explanatory power of science.
It may be hard to imagine that brain scans can tell me anything important about my own subjectivity. But when I learn that my brain hears, in the violin, the same expressiveness it hears in a human voice, I've learned something about myself and my relationship to the music. Of course Bach did not do brain scans on anyone before writing Air on A G String. But did he know a listener might hear the sorrowful voice of a friend, or a mother? I do not doubt this. His genius, as well as that of Michelangelo, consisted in seeing connections now reflected by our best understanding of biology.
Similarly, when I'm banging my head to the endlessly repetitive rhythms of my favorite metal band Isis, it widens my relationship with the music to know my cerebellum (which regulates motor function) is at that moment furiously active in rhythm processing, and that it has massive connections to the amygdala which is tasked with remembering emotional events. And that the cerebellum is the most primal reptilian part of our brain.
It may be hard to imagine how we might separate out legitimate experiences of religious euphoria (if any) from erotic fantasies. But, presumably we'd do science on this the same way we do science on other things: with attention to detail, with control subjects, with creativity. Attending to the specifics of subjective reports, and the details of brain activity, we can compare what the nuns feel to actual sexual fantasy, explicitly tested for in other subjects, or actual experiences of euphoria from atheists. Do you think we would not tease out differences? Or do you think that, having failed to spot differences, this in itself wouldn't be positively interesting as opposed to neutral? We could find that experiences explode into a dozen or a thousand different categories, or that they are the same.
On correlations: it appears to me the history of science with respect to any phenomenon, is a history of encroaching correlations that eventually infiltrate the phenomenon itself. When this happens, they are no longer correlations but simply a description of what the phenomenon is. Why should it be the case that we can map out the physiological basis for chills in response to music, as Zatorre and Blood have, but not the higher-order circuitry that preferentially responds to qualitative aspects of music to induce those chills? Where exactly is it, in the interpretation of art, that science isn't supposed to be able to get to?
We suffer from an impoverished language that rarely ever does justice to our euphorias, or religious experiences. This I think, together with the fact of our being drenched in 2000 years of human culture where the only attempts at language-making with respect to aesthetics and euphorias have come from people wholly ignorant of brains, has constrained our intuition and our ability to form hypotheses that might relate these experiences to the objective world (i.e. to brains). Even since the scientific revolution brain science has only just come of age, or maybe hasn't even.
The history of religious (and philosophical!) retreat in the face of scientific explanation has not finished, and atheists are wrong to concede the point with respect to aesthetics. It's not that we would wipe out art (a misguided fear that I think is not unlike religious fear of losing their humanity if they become atheists) but that we would come to greater understanding of ourselves. Feynman said it like this: "It does not do harm to the mystery to know a little more about it. For far more marvelous is the truth than any artists of the past imagined it."