I Prof Mike Wiest's proposed link between quantum and consciousness

[nomadreid]

TL;DR Summary

How valid are Prof. Mike Wiest's (Wellesley College) speculations, based on his experiments (link given), of the link between consciousness and quantum processes?

I am not sure if this belongs in the biology section, but it appears more of a quantum physics question.

Mike Wiest, Associate Professor of Neuroscience at Wellesley College in the US. In 2024 he published the results of an experiment on anaesthesia which purported to point to a role of quantum processes in consciousness; here is a popular exposition:
https://neurosciencenews.com/quantum-process-consciousness-27624/

As my expertise in neuroscience doesn't reach up to an ant's ear, and my quantum physics background at an elementary level, I have no idea how valid the experiment or his tentative conclusions about a link between consciousness and quantum processes are. I did cringe a little at the mention of these processes taking place in microtubules in the brain, as these figured prominently in the discredited "Orchestrated Objective Reduction" theory of Penrose and Hameroff. (I must admit that, knowing more mathematics than neuroscience, it was Penrose's misunderstanding of Gödel's results that led me to the refutations of the "cascade" theory which tried, much less successfully than one of this year's Nobel winners, to link quantum effects to macroprocesses.) But of course guilt by association ("microtubules") is not valid, and so I turn to the more knowledgable people on this forum.

I could not find replication of his results, but I probably am not looking in the right place, or perhaps a year is too little time. Even if the results are replicated, I am not knowledgable about all the possible confounding factors which might mean that the observed effects are due to classical rather than quantum effects.

Thank you in advance for shedding any light on this.

 

[PeterDonis]

I have no idea how valid the experiment [is]

As far as his claimed effect of microtubule-binding drugs on anesthesia, I have no idea. That sort of discussion belongs in the appropriate biology forum, as the experiment itself has nothing to do with quantum physics. It would be evaluated using the same kind of standards as any other experiment claiming to show a biological effect of a drug.

or his tentative conclusions about a link between consciousness and quantum processes

Those aren't even his "tentative conclusions". He's just saying "somebody else came up with this microtubules are involved in consciousness thing, and I did an experiment about microtubules and anesthesia". Again, the experiment itself has nothing to do with quantum physics.

 

[nomadreid]

Thanks very much, PeterDonis. Hm, Indeed I appear to have read it too quickly, having been led astray by the summary by the journal at the beginning (which deceptively makes it sound as if there was a quantum aspect to the experiment), but I see from the quote by the author that “Since we don’t know of another (i.e., classical) way that anesthetic binding to microtubules would generally reduce brain activity and cause unconsciousness,” Wiest says, “this finding supports the quantum model of consciousness.”, that the link to quantum processes is pure speculation on his part. So, oops, sorry for posting in the wrong forum.

 

[mitchell porter]

I worked for Hameroff for a year. I consider their theory a great contribution, because of how mind-opening it is.

In my opinion, if quantum information is genuinely part of cognition or consciousness, the microtubule is an excellent candidate for where it could be happening. There needs to be some protection against room-temperature decoherence, and the cylindrical topology of the microtubule suggests topologically stabilized quantum states.

However, it is very difficult to directly prove or disprove quantum effects in living matter, because their contribution may be quite subtle. You can see this in the debate about whether there's a quantum speedup in the transfer of energy through the photosynthetic complex. The same problem arises in quantum computing paradigms that are built on top of a classical computational process, like adiabatic quantum computing.

The standard view of the microtubule is that its role is structural. An amoeba extends a pseudopod by growing out its microtubules, the microtubules move the chromosomes around during cell division, they are also the highways along which certain enzymes are carried by the walking molecule kinesin.

Recall that the basic idea of how computation occurs in the nervous system, is through a combination of polarization waves traveling along axons, and neurotransmitters being spat across the synaptic gap from one neuron to the next. Whether these events occur depends on a lot of conditions in and around the cell, and the microtubules can play a very indirect part by affecting those background conditions.

But if you were looking for a way that microtubules could be directly involved in information processing, the way that the axon membrane and the synaptic vesicles are, I would start with the microtubules in the neuronal axon, which are oriented along the axon. There might be electromagnetic interactions between the axon electrical potential and mobile electrons within the microtubules; there could be chemical interactions between ions and proteins that cross the axon membrane, and the tubulins in the microtubules.

I believe there are a number of experimental results indicating some interaction between anaesthetics and microtubules. I don't keep up with that side of things - in fact, I'm prepared for actual biologists to correct my word-pictures above, in any number of ways - and I can't tell you how Mike Wiest's results impact the existing body of theory and knowledge in anesthesiology (as opposed to Orch OR theory, specifically). But I would expect that the observations themselves are valid.

Here's an article by Wiest himself: "Old theory, new evidence"

 

[pines-demon]

I can see why searching for quantum phenomena in the brain would be encouraged or at least interesting to look into it. But I cannot even fanthom how quantum phenomena in the brain could have any link to consciousness, and even Penrose is very handwavy about this.

 

[sbrothy]

On the face of it it almost sounds like it belongs in my "philosophy thread". Whenever I hear (read) "quantum" and "consciousness" in the same sentence alarms go off.

 

[AlexB23]

On the face of it it almost sounds like it belongs in my "philosophy thread". Whenever I hear (read) "quantum" and "consciousness" in the same sentence alarms go off.

Agreed 100%. The article sounds like bologna sandwich.

 

[sbrothy]

Why thank you very much for that backhanded compliment on my "philosophy thread"!

EDITED.

 

[AlexB23]

Why thank you very much for that backhanded compliment on my "philosophy thread"!

EDITED.

Haha, you are funny. But this article was posted by OP, @nomadreid

 

[mitchell porter]

I can see why searching for quantum phenomena in the brain would be encouraged or at least interesting to look into it. But I cannot even fanthom how quantum phenomena in the brain could have any link to consciousness, and even Penrose is very handwavy about this.

There is no consensus about how anything physical, classical or quantum, manages to relate to consciousness; and yet there needs to be some kind of relationship, since the brain is physical. Also, Penrose himself never said anything as simple as "quantum mechanics explains consciousness".

His particular motivation was to find a physical/computational mechanism capable of transcending the limits implied by Godel's incompleteness theorem, since (it seemed to him, and others) that the brains of human mathematicians must be doing something beyond Turing-equivalent computation, in order to perform meta-mathematical reasoning.

But the whole of physics seemed to be Turing-computable, except, potentially, quantum gravity. He combined this with his thoughts as a physicist on how wavefunction collapse seemed to be an information-creating process complementing the information-destroying process of Hawking radiation; and arrived at his paradigm that (1) some exotic trans-Turing dynamics governs the spontaneous collapse of quantum superpositions in which the proper time of the branches becomes sufficiently out-of-sync (2) this is the physical process which, in the human brain, implements certain forms of cognition.

And then he hooked up with Hameroff to suggest to the world that neural microtubules are the specific biological locus where this occurs.

So I want to emphasize that Penrose has never proposed a conceptual resolution to the philosophical problem of how consciousness relates to matter. In philosophy of mind, there are all these concepts like emergence, epiphenomenalism, functionalism, property dualism, mind-brain identity, and so forth, which try to answer that question. Penrose and Hameroff are making a (highly exotic) physical hypothesis about the nature of the brain processes associated with specific forms of cognition.

Not many people agree with Penrose's motivating argument. The question of the powers of meta-mathematical reasoning, and what's actually going on there, again leads to philosophy of mind, in this case the semantics of cognitive states, and how brain states and computer states manage to "be about" something - how they manage to have meaning. As with consciousness, again there is no intellectual consensus here, just the clash of competing schools of thought. The human race has not definitively figured out how to think about these topics.

On Hameroff's side, the key novel hypothesis is just that microtubules perform informational functions that are directly relevant to cognition and consciousness. That hypothesis could be true classically, or it could be true in a non-exotic quantum-mechanical way. The Hameroff-Penrose theory ("exotic" quantum dynamics in the microtubules) just happens to be the most extreme version of this hypothesis. But in principle, Wiest's work could also be used in defense of the weaker versions of the hypothesis.

 

[Nugatory]

and yet there needs to be some kind of relationship, since the brain is physical.

It is not impossible that more years (decades?) of AI development will lead to conscious physical systems In which the underlying physics is well understood. We may be perplexed by the emergent behavior of a system of, say, $10^{15}$ transistors, but we would not doubt that we understand the physics at every step.

But somewhere in there we will have to agree about what it means that a physical system is "conscious". I know that I am, but I am not 100% sure of the connection between my consciousness and the physical object inside my skull.

 

[PeterDonis]

His particular motivation was to find a physical/computational mechanism capable of transcending the limits implied by Godel's incompleteness theorem, since (it seemed to him, and others) that the brains of human mathematicians must be doing something beyond Turing-equivalent computation, in order to perform meta-mathematical reasoning.

It should be noted that there are plenty of others who disagree with this point of view. One of the most notable arguments against it was due to Daniel Dennett, the philosopher, who constructed a thought experiment scenario in which a Turing machine can actually prove its own Godel sentence! He did this by taking advantage of the fact that one Turing machine can emulate another. (I personally am with Dennett and the other skeptics on this one--I think claims that humans somehow are able to transcend Godel's theorem are based on a misunderstanding of what the theorem actually says, combined with overstating what humans can actually do.)

 

[pines-demon]

His particular motivation was to find a physical/computational mechanism capable of transcending the limits implied by Godel's incompleteness theorem, since (it seemed to him, and others) that the brains of human mathematicians must be doing something beyond Turing-equivalent computation, in order to perform meta-mathematical reasoning.

But the whole of physics seemed to be Turing-computable, except, potentially, quantum gravity. He combined this with his thoughts as a physicist on how wavefunction collapse seemed to be an information-creating process complementing the information-destroying process of Hawking radiation; and arrived at his paradigm that (1) some exotic trans-Turing dynamics governs the spontaneous collapse of quantum superpositions in which the proper time of the branches becomes sufficiently out-of-sync (2) this is the physical process which, in the human brain, implements certain forms of cognition.

This is what I find hand-wavy. It is well know that a classical computer can simulate anything a quantum computer can do (just in more steps) so anything quantum is computable. UNLESS according to objective collapse interpretations like Penrose's, the collapse is truly random. But then, it does not give you nondeterministic Turing machines (that can do uncomputable stuff) this only gives you probabilistic Turing machine the most boring sense of non-computable (you could build one out of coin flips).

And even accepting that quantum computing can do some uncomputable stuff, then what? I have never found what would that have anything to do with whatever micro-tubules are doing or consciousness. The only thing that makes me feel comfortable is that you say that Penrose has no answer to this.

 

[mitchell porter]

This is what I find hand-wavy. It is well know that a classical computer can simulate anything a quantum computer can do (just in more steps) so anything quantum is computable. UNLESS according to objective collapse interpretations like Penrose's, the collapse is truly random. But then, it does not give you nondeterministic Turing machines (that can do uncomputable stuff) this only gives you probabilistic Turing machine the most boring sense of non-computable (you could build one out of coin flips).

And even accepting that quantum computing can do some uncomputable stuff, then what? I have never found what would that have anything to do with whatever micro-tubules are doing or consciousness. The only thing that makes me feel comfortable is that you say that Penrose has no answer to this.

Noncomputability is a different thing from nondeterminism. It's actually related to undecidability, as is seen in the halting problem. For a given computer program, the halting problem is the question whether that program would run forever, or eventually halt. It was an early discovery of computer science that there cannot be a computer program which correctly answers this question for all possible programs (that is, an arbitrary program is supplied as input data, and the output is whether or not it halts when run). The proof is by contradiction: if there was such a perfect halting oracle, you could build a bigger program out of it that should both halt and not halt. It is the same kind of argument as Godel's incompleteness theorem, but in the context of computer programs rather than axiomatizations of arithmetic.

However, although program halting is in general Turing-undecidable, one supposes that for each possible program, there is some fact as to whether or not it can halt. There is, therefore, a definite binary function which takes programs as inputs, and produces "yes, it halts" or "no, it doesn't halt" as outputs; but there is no program which computes that function. It is a deterministic noncomputable function.

Furthermore, there are actually infinite hierarchies of noncomputable functions. You can take that uncomputable halting function, and define it as a new computational primitive alongside the usual logic gates, something that you can call inside a computer program. Now there is an expanded set of possible programs, in which the usual Turing computation is expanded by the ability to call on a halting oracle. This expanded set of programs is able to "compute" answers to the halting problem, because they can just make a function call to the halting oracle. But actually there is now an expanded halting problem, for programs that can access a first-level halting oracle, and again, they can't answer it. It would require a second-level halting oracle to answer this expanded halting problem.

And so it goes, with the repeated addition of halting oracles to your set of computational primitives, being analogous to the repeated addition of consistency axioms to an incomplete axiomatization of arithmetic (and also to the construction of transfinite ordinals). You keep expanding the range of what you can compute or prove, but there continue to be functions that you can't compute and facts that you can't prove.

All this might be regarded as an arcanum of interest only to academic specialists, except that concrete problems occasionally turn out to be undecidable. The first such was Hilbert's tenth problem: does an arbitrary Diophantine equation have solutions? There is no program that can answer that question for all possible Diophantine equations. We can deduce the answer for certain classes of Diophantine equation, but not for all... An example of interest to physicists is the existence of a mass gap in many-body spin systems; it was shown ten years ago that this too is undecidable. Joel Hamkins here gives a list of noncomputable functions arising from undecidable problems.

The physical example is interesting, because it suggests that a physical process could implement an oracle function: maybe you can answer the mass gap question for a specific spin system, by simply building it and watching what happens! Furthermore, there is some prospect that quantum gravity can do this, because you have 4-manifolds showing up in gravitational path integrals, and there are undecidable properties of 4-manifolds.

@PeterDonis in #12 has already stated the skeptical position regarding the so-called Lucas-Penrose argument, that mathematical reasoning can't be solely a product of Turing computation, that it must have access to oracle functions. Dennett and others argue that you can account for the demonstrated abilities of mathematicians without resorting to that. But we could say that if you did think it was necessary to suppose that human cognition can access trans-Turing computational primitives, then a role for quantum gravity in the brain might just be able to do it. :-)

I feel I should add something about the status of this idea - "noncomputable quantum gravity governing quantum state reduction" - solely as a research program in physics. The idea that quantum gravity might break quantum linearity goes back to the information loss problem associated with black hole evaporation. These days, in string theory at least, they believe that there is no information loss. Penrose, however, has a complex of alternative ideas that he still pursues under the banner of twistor theory. The most recent work in this vein, that I know of - "Quantum state reduction, and Newtonian twistor theory" - dates from 2022. That work is probably the most advanced version of Penrose's research program yet published, and it contains nothing about noncomputability or undecidability, they are still just trying to forge the link between gravity and quantum dynamics. (There are other people working in that area too; Sabine Hossenfelder put out a paper on that topic just this week, but I haven't read it yet.)

 

[pines-demon]

Noncomputability is a different thing from nondeterminism. It's actually related to undecidability, as is seen in the halting problem.

Just for sake of everything, I am not equating noncomputablity with nondeterminism, I am saying that quantum computing does not give you noncomputability, and does not give you nondeterminism either.

The physical example is interesting, because it suggests that a physical process could implement an oracle function: maybe you can answer the mass gap question for a specific spin system, by simply building it and watching what happens!

You would have to argue what is uncomputable in quantum mechanics for that to work, under current quantum mechanics there is nothing that is uncomputable.

Furthermore, there is some prospect that quantum gravity can do this, because you have 4-manifolds showing up in gravitational path integrals, and there are undecidable properties of 4-manifolds.

I feel I should add something about the status of this idea - "noncomputable quantum gravity governing quantum state reduction" - solely as a research program in physics. The idea that quantum gravity might break quantum linearity goes back to the information loss problem associated with black hole evaporation. These days, in string theory at least, they believe that there is no information loss. Penrose, however, has a complex of alternative ideas that he still pursues under the banner of twistor theory. The most recent work in this vein, that I know of - "Quantum state reduction, and Newtonian twistor theory" - dates from 2022. That work is probably the most advanced version of Penrose's research program yet published, and it contains nothing about noncomputability or undecidability, they are still just trying to forge the link between gravity and quantum dynamics.

Why would quantum gravity be needed to explain the brain??? That makes the whole argument weirder. Quantum gravity is an open problem and I refuse to speculate on that without specifics. @PeterDonis will probably agree.

 

[PeterDonis]

Why would quantum gravity be needed to explain the brain?

I don't think the research program referred to is to use quantum gravity to explain the brain specifically. I think it is to use quantum gravity to explain why measurements have single outcomes, or more generally why decoherence leads to single outcomes. That requires some kind of nonlinearity.

 

[pines-demon]

I don't think the research program referred to is to use quantum gravity to explain the brain specifically. I think it is to use quantum gravity to explain why measurements have single outcomes, or more generally why decoherence leads to single outcomes. That requires some kind of nonlinearity.

Oh right! It was the "quantum gravity" terminology that threw me off. Proving or disproving gravitational-based collapse does not mean that it reveals anything about quantum gravity (unification of QFT/GR) right? Maybe in a sense it does but I would keep the terms separate.

 

[pines-demon]

I feel I should add something about the status of this idea - "noncomputable quantum gravity governing quantum state reduction" - solely as a research program in physics. The idea that quantum gravity might break quantum linearity goes back to the information loss problem associated with black hole evaporation. These days, in string theory at least, they believe that there is no information loss. Penrose, however, has a complex of alternative ideas that he still pursues under the banner of twistor theory. The most recent work in this vein, that I know of - "Quantum state reduction, and Newtonian twistor theory" - dates from 2022. That work is probably the most advanced version of Penrose's research program yet published, and it contains nothing about noncomputability or undecidability, they are still just trying to forge the link between gravity and quantum dynamics.

Retaking this point. So basically the only idea here is that gravitational-based collapse is maybe not like the quantum mechanics we know (beyond Standard Model) so that it could allow noncomputability? I feel like this is like looking for the God of the gaps but replacing God by consciousness.

 

[PeterDonis]

Gravitational-based collapse does not mean that it reveals anything about quantum gravity (unification of QFT/GR) right?

We won't know that for sure until we have a confirmed theory of quantum gravity.