I don't believe it's 'just' chemistry and responses to stimuli,
  no. The model we currently use is somewhat flawed, I believe. We
  study these things "as if" we're isolated beings in a
  laboratory. But we're not. We do tests on humans "as if" that is
  true, and that causes the results to confirm our preconceived
  notions of the self.

  But we're social creatures, as much as we deny it. Perhaps when
  we have enough "stuff" in us already we can make it. But what
  baby survives without socializing?

  Is that responses to stimuli? Maybe. But at some point, if it's
  a machinery, the machinery simply runs at that point. We're
  here. You're here. I'm here. There's a you there, and a me here.

  Can a state snapshot be taken of a brain to be able to predict
  the next response?

  Theoretically maybe. But the thing is: We're not static. An
  unmoving brain is a dead brain. We don't respond to stimuli the
  same way each time, because we're in a different state each
  time.

  We're not the same person moment to moment.

  This constant change pulls the rug out from under the chemistry
  + responses to stimuli because the subject of study is
  continually in flux and is not isolated from the environment. It
  can't be.

  We're always in a here/not here state.