This is an extremely interesting paper, thanks. I had had a chance to
look at it indeed a couple of days before you posted, so I have some
thoughts on it already. The main contribution is the postulated
mechanism for 6-bit encoding in CaMKII. But to say that this solves
the problem of memory is an yet unjustified stretch. We know
experimentally that removal of up to 90% of the brain of rats doesn't
seem to impact their memory (they can still negotiate a maze they had
learned before surgical removal of parts of their brains). It's as if
memory was nowhere and everywhere, which is what motivated Karl
Pribram to think of memory as holographic. Such lack of localization
is not tackled by the mechanism proposed in this new paper, since the
local storage of specific bits in CaMKII would seem to require
precisely the opposite. Another thing to keep in mind is that it is
currently unknown, even in principle, how synaptic activity leads to
conscious experience (in principle, it could all happen purely
electrochemically, 'in the dark,' like much of the brain's activity).
So even if we had a complete and proven model of how synaptic activity
patterns can be stored (and we are far from it, despite this new paper
representing some potential progress), it would still leave open the
mechanism for how that can be translated into the experience of
memories.
Stuart Hameroff is the lead researcher of this new study. So it's a
matter of time before he links this to his 'quantum consciousness'
theories, which are non-local. This could make his proposed mechanism
of memory more plausible, but at the cost of deviating from strict
materialism as understood and adopted by most neuroscientists today.
In other words, the explanation would sit somewhere between the
material and non-material, with a fair dose of quantum "woo"... :-)
For my own philosophy, it is not important whether memories have
physical correlates (i.e. bits) stored in the brain or not. In fact,
in my first book, "Rationalist Spirituality," I explicitly assume that
memory correlates (which I called 'symbols') are indeed stored in the
brain like bits in a computer. I do think, though, that finding these
mechanisms of storage, in view of the empirical data at hand, has
proven extremely elusive and won't be solved suddenly.