[ExI] quantum brains
stathisp at gmail.com
Sun Jan 24 09:12:21 UTC 2010
2010/1/24 The Avantguardian <avantguardian2020 at yahoo.com>:
>And of course the microtubules are a bizarre non-sequitor brought in by Hammeroff because like most molecular biologists, he thinks the protein or gene he is specialized in is the most important molecule in the world. So let's throw out the gravity, since astronauts in zero G don't get stupid or lose consciousness. And let's assume that microtubules are a structural component of the cytoskeleton and leave it at that. After all neurons are not particularly mobile for cells and if microtubules were the mechanism of consciousness, peoples would think with their muscles instead of
> their brains because myocytes have more microtubules than neurons. And let's put off any conclusions regarding the computability or uncomputability of the brain until we have better defined the problem space and have more data. Let's just entertain the naked hypothesis that brains may utilize QM for some part of their function and see where it leads.
Yes, but I am specifically interested in the question of brain
computability, whatever the underlying mechanism. If brains are not
computable then that has profound implications for philosophy of mind;
if brains use quantum level events but are still computable then that
is scientifically interesting but it doesn't make much difference
> Perhaps. But there are other parallels that don't apply to classical chaotic systems. Things like the mind:body -> quantum wavefunction:quantum information dualites that Serafino mentioned in his post to which one could add the more cliche wave:particle duality.
> Then there is the way that excitory post-synaptic potentials (EPSP) and inhibitory post-synaptic potentials (IPSP) can sum over time and space to trigger neuronal depolarization that is very reminiscent of constructive and destructive interference. Although by my cursory swim through the literature, reports are all over the board with the quantitative measurements of these things, the mean voltages of these signals are approximately 5 millivolts and the mean current is about 25 picoamps. The average duration of these signals are about 20 milliseconds. Multiply all these rough figures together and you get approximately 2.5 femtojoules or about 15,600 eV which is admittedly too much energy for quantum effects.
> However there are EPSP that fire spontaneously at a much smaller voltage called miniature EPSPs or mEPSPs that are in range of about 400 microvolts, 10 picoamps, and last about 1 millisecond. These are thought to be caused by single vessicles of neurotransmitters being randomly being released into the synaptic cleft. These things have an energy of about 25 eV which for comparison is not much higher than the ground state of the hydrogen atom at -13.6 eV. Moreover these things happen quite frequently.
> Now it seems to me that these mEPSPs in the brain are very similar to the quantum fluctuations in normal matter. Like the fluctuations in a nebula of hydrogen gas that could trigger the condensation of the gas into a protostar. Since EPSP are additive, one or more of these things could push a subthreshold normal EPSP over the threshold causing a recipient neuron to depolarize and initiate an action potential. So this is a potential mechanism for leaps of intuition, hunches, imagination, and creativity. And it seems a much more testable hypothesis than quantum microgravity yanking on microtubules ala Penrose and Hammeroff. Now the interesting question is are these things capable of the more bizzare quantum behavior like uncertainty, entanglement, and wave-particle duality? e.g. could mEPSPs interfere with themselves? Or might they exhibit particle properties and be called "psions"?
It seems quite possible that quantum fluctuations might be amplified
by the brain so that they have macroscopic effects on behaviour, but I
don't see how this would do anything other than function as a random
number generator. Essentially the neuron is just a black box which
will fire or not fire depending on the inputs it receives and on its
internal state. If it contains an RNG that would affect its propensity
to fire, but modelling that should be trivial compared to the
difficulty of modelling the complex interconnections between neurons
and the deterministic component of each neuron's internal state: if
you get that right, then the neuron model will function like a real
neuron and the brain will function like a real brain, including
consciousness. Penrose thinks it is impossible to model these things
because neurons do something essentially non-algorithmic. Ordinary
quantum mechanics doesn't cut it here since it is computable with the
exception of its randomness (unless you cheat and use a branching
world algorithm), which can be easily approximated with a pseudorandom
number generator. So Penrose postulates that there is a
non-algorithmic theory of quantum gravity which allows neurons to act
as hypercomputers. This is how far he is forced to go in order to
maintain his view that computers could never have true intelligence or
More information about the extropy-chat