sawfoot _:
JC, Jaynes is a fun read but its pretty nuts! Certainly our concept of consiousness has changed a lot in the last few thousand years, and this must have changed our experience of self considerably. But the current best guesses on the evolution of human culture is that "modern" humans have been around for at least 50,000 years - though some put the date back as far as 200,000-400,000 years (in Africa). And you seem to be equating "self" with "consciousness". But I agree in principle - social relations and increased cultural complexity probably had a significant influence on the development of consiousness and "selfing processes".
Something "happens" in the brain when I go to the toilet. Something also happens when I have a stroke. In one case there are profound changes with lasting consquences and in the other less so. The claim in the DhO is that some brain events are more significant than others. With neuroimaging (e.g. fmri or EEG) you can pick up a lot about brain activity, so my best guess is that you would be able to pick up an event like stream entry, or at least "blips" involving a transient loss of consciousness, given their likely EEG signature.
So just a random example of what we can detect to contrast with your poem example:
http://www.sciencemag.org/content/340/6132/639
While Jaynes uses the word "consciousness," he's talking about the idea of the self. Using his terminology, schizophrenics, enlightened people, and people before a few thousand years ago were all not "conscious." But clearly they were aware of sensory perception; they just didn't have the same sense of self. The terminology is a little confusing. I don't think he's nuts, though some of his chronology and analysis of ancient cultures may be off - the main point, like you say, is that the self and selving processes are affected by the culture that people grow up in - software programmed in at an early age.
With enough data and sophisticated machine learning techniques, we can probably detect anything, as in your example. The way that it worked in that study was that they'd scan people's brains and ask them to think about, say, skiing, and the machine looks at the pattern of static and then detects when the brain is in a statistically similar state. But you need enough data, and with stream entry you'd have to rely on data from a lot of different people - individual neurological differences might make it complicated. But I agree, if we manage to catch enough people entering the stream and scan their brains, we may be able to come up with some pattern of static and statistically match it to people.
My point wasn't about detection, though, but about being able to easily describe what happens on a neurological level during stream entry. Stream entry is a "mind" event, not a brain event like a stroke. While there's probably some complicated pattern that is associated with stream entry, I doubt it can be simply described by "X, Y, and Z is what happens in the brain."
I'm not clear on what you mean by "some brain events are more significant than others." There's significance on a neurological level, and there's significance on the level of meaning and experience. A brain event that's very significant neurologically may have only minor significance to the person's life and vice versa. While stream entry is a very significant event in a person's life, it's not clear to me that there's anything especially significant about it neurologically. I see it more as a complex, high-level organizational shift - in other words, I doubt you'd see anything special or noticable without a statistical analysis.