Quantum Biology and the Hidden Nature of Nature

[testingus]

Just came across this youtube video discussing Quantum Biology



It features Paul Davies and Seth Llyod. It debates whether biology makes use of non-trivial quantum effects to improve performance and/or perform functions that cannot be accounted for classically. I'm posting it here and asking the following, Do you think biology makes use of non-trivial quantum effects? If so, why? If not, why not?

 

[Pythagorean]

I'm posting it here and asking the following, Do you think biology makes use of non-trivial quantum effects? If so, why? If not, why not?

It's a broad question, and I think it should be answered on a case-by-case basis. QM has been shown to be significant in at least one process of one biological system (photosynthesis in plants). I think I also remember something about bird navigation. I haven't watched the video you posted and I see a leaf on it, so maybe that's discussed.

 

[Bystander]

Photosynthesis as Pythagorean has noted, vision/eyes/rhodopsin(+other dyes), bioluminescence.

 

[Ygggdrasil]

There are a few well documented cases where biological processes take advantage of quantum effects to function. Here's a link to a news feature from Nature describing a few examples (with citations to the relevant literature).

 

[.Scott]

It debates whether biology makes use of non-trivial quantum effects to improve performance and/or perform functions that cannot be accounted for classically. I'm posting it here and asking the following, Do you think biology makes use of non-trivial quantum effects? If so, why? If not, why not?

I find it aggravating that the most in-your-face phenomenon in Physics is off limits to Physics. It's what Seth Lloyd referred to as the "C word". There's only one known physical condition where lots of information can exist in a single state - QM super positioning.

So my answer is an emphatic "Yes".

 

[testingus]

There's only one known physical condition where lots of information can exist in a single state - QM super positioning.

What about Tononi's Integrated Information? Is this not a classical description of state with a large amount of information?

http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000091#pcbi-1000091-g018

 

[.Scott]

What about Tononi's Integrated Information? Is this not a classical description of state with a large amount of information?

http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000091#pcbi-1000091-g018

That article is using "Hopfield networks" or "Hopfield-like networks" to get a the "integrated information" into a small number of neurons. (The article also addresses other structures.) The authors (Balduzzi and Tononi) recognize that, for consciousness, the information needs to be integrated - and work with ways of describing that integration. Their paper reinforces the critical step in recognizing that information doesn't just add up on its own, that there needs to be some systematic way of causing this information to build up. But they only go as far as the neuron - or small network of neurons. Once at that level, you still haven't merged the information into a single state - only a collection of independent states that fit within a neuron or network.

 

[Pythagorean]

As far as I know, in the context of biology, the only useful aspect of information density from QM is that it allows for classical matter. The organism itself doesn't exploit the information density.

Even in the case of photosynthesis, the advantage seems to be the speed.

 

[.Scott]

As far as I know, in the context of biology, the only useful aspect of information density from QM is that it allows for classical matter. The organism itself doesn't exploit the information density.

Even in the case of photosynthesis, the advantage seems to be the speed.

If you can tell me that you are conscious of more than a couple of bytes of information at a time, then that information density has affected your behavior.

 

[Pythagorean]

If you can tell me that you are conscious of more than a couple of bytes of information at a time, then that information density has affected your behavior.

"More than a couple bytes" can easily be handled by the relevant fraction of 86 billion classical neurons each with a 7000 synapses, and each modulated by a number of second-messenger and transcription processes. Nobody denies that the classical mechanisms have QM under the hood, but this discussion of "Quantum Biology" is more about biological mechanisms that can only be handled by QM perspective and not the classical perspective (i.e. "non-trivial" QM as the OP put it).

 

[testingus]

"More than a couple bytes" can easily be handled by the relevant fraction of 86 billion classical neurons each with a 7000 synapses, and each modulated by a number of second-messenger and transcription processes. Nobody denies that the classical mechanisms have QM under the hood, but this discussion of "Quantum Biology" is more about biological mechanisms that can only be handled by QM perspective and not the classical perspective (i.e. "non-trivial" QM as the OP put it).

But are the many bytes of the neurons/synapses merged into a single coherent state? I think what .Scott is trying to state is that even though IIT gives a measure of integrated information, it does not provide a mechanism for which the info is integrated (correct me if I'm wrong). Fundamentally then wouldn't you need a QM non-trivial process to merge/integrate the information into a single whole state?

 

[Pythagorean]

But are the many bytes of the neurons/synapses merged into a single coherent state? I think what .Scott is trying to state is that even though IIT gives a measure of integrated information, it does not provide a mechanism for which the info is integrated (correct me if I'm wrong). Fundamentally then wouldn't you need a QM non-trivial process to merge/integrate the information into a single whole state?

But the evidence already shows that there's too much heat and to large of spatiotemporal scales in the brain to support a QM-based integration. I'm not sure either QM or classical physics is enough, but the evidence (neural correlates consciousness with respect to dynamics) implies that it will be an extension of classical mechanics, not a reduction to QM (the mountain of understanding we have of brain function is based on classical models and information theory, compared to a very tiny minority of very speculative propositions with regards to QM). What Tonini shows is that information integration (of classical components) is consistent with our expectations. You're right that it doesn't give a mechanisms, but the "dependent variable" is the dynamics and interaction of classical components.

 

[testingus]

But the evidence already shows that there's too much heat and to large of spatiotemporal scales in the brain to support a QM-based integration. I'm not sure either QM or classical physics is enough, but the evidence (neural correlates consciousness with respect to dynamics) implies that it will be an extension of classical mechanics, not a reduction to QM (the mountain of understanding we have of brain function is based on classical models and information theory, compared to a very tiny minority of very speculative propositions with regards to QM). What Tonini shows is that information integration (of classical components) is consistent with our expectations. You're right that it doesn't give a mechanisms, but the "dependent variable" is the dynamics and interaction of classical components.

I thought the "too warm and wet" argument against QM processes in biology had been put to rest? The work in the field of photosynthesis (Greg Engel, Greg Scholes) has shown that thermal noise can assist quantum coherent energy transfer in these systems. The spatiotemporal scale concern is one that is actively under investigation.

 

[.Scott]

But the evidence already shows that there's too much heat and to large of spatiotemporal scales in the brain to support a QM-based integration. I'm not sure either QM or classical physics is enough, but the evidence (neural correlates consciousness with respect to dynamics) implies that it will be an extension of classical mechanics, not a reduction to QM...

QM and relativity is the extension to classical mechanics - there is no other. So if we need an extension, it will be to QM.
Besides, the wet and warm QM is merely a technological issue. But putting significant amounts of inform into a single state is fundementally missing from classical electronics/mechanics - eventually, HUP (QM) gets in the way.

... (the mountain of understanding we have of brain function is based on classical models and information theory, compared to a very tiny minority of very speculative propositions with regards to QM).

We haven't even started to look for significant QM mechanisms. And, as can be seen with the photosynthesis debate, it's a difficult detect and demonstrate. It's a far cry from hooking up an osciliscope to an electrode and probing the brain of a guinea pig. When we finally find the mechanism, it will be because we already know exactly what the mechanism does and what it's overall contribution is to thought. At that point, we will know exactly what we are looking for and roughly where to find it.

What Tonini shows is that information integration (of classical components) is consistent with our expectations. You're right that it doesn't give a mechanisms, but the "dependent variable" is the dynamics and interaction of classical components.

Tonini's paper is useful. It shows how the information capacity of a neural net system increases based on how the neurons are designed and interconnected. But having 10 bytes classically distributed among 1000 neurons is far from having a single state. It's like writing 10 bytes worth of information onto a piece of paper and expecting the paper to be conscious of it.

 

[Pythagorean]

@testingus

As far as I know, this is only true for very local events, such as the photosynthesis example.

@.Scott

There's really no reason to require that all consciousness must relate to one physical state as we describe it on paper. Consciousness can be a conflicting and dissociative amalgam of experience. There are many examples of this in the field of clinical neuropsychology, as Tononi pointed out.

 

[Almeisan]

I don't really get it. Nature does what it does and nature does it in a way that we needed QM to explain it.

No chemical bond can be explained classically. Not one thing that it does it does without QM.

Obviously life would not be able to exist if nature had to obey classical mechanics. That's a truism.
This lecture seems to try to explain mysteries with novel (interpretations of) QM.
Cheat time and cheat distance? Does QM do that?
Or are they talking about biology being able to change the Planck scale?

Do we know how biological systems work in terms of particle physics? No we do not. Right now it is already difficult enough to get inferences about the shapes of proteins. Let alone how they operate on a scale were quantum 'fuzziness' dictates and classical thinking breaks down completely.
The lecture talks about 'we have evidence quantum mechanics' plays a role in photosynthesis. And it's physics professors saying that?

What do they think biology is? Do they think that once one of 'their' photons hits a plant leaf it suddenly changes to a world where Calc I is all you need to explain what happens 'because biology isn't a real science and they don't do calculations'?

Would be a truly amazing discovery when it is shown that biology can 'turn off' QM and catch photons or transfer electrons 'as if they behaved as particles'.

 

[billschnieder]

I don't really get it. Nature does what it does and nature does it in a way that we needed QM to explain it.

Nature doesn't care what theory humans have developed to attempt to explain it. QM is just another man-made theory, just like classical mechanics, albeit a very successful one. It doesn't make any sense to say nature "turns off QM", QM is a theory, a series of postulates and a mathematical set of rules for manipulating information to enable humans ask and answer questions about nature. It is not a physical thing out there which nature can use or "turn-off".

 

[Almeisan]

Of course it doesn't make sense. Did you read my post?

 

[Pythagorean]

We don't need QM to explain the trajectory of a cannon ball. You might even argue that a QM treatment wouldn't be successful. Likewise, the dynamics of an RLC circuit can be predicted more easily with classical electrodynamics.

The neuron and its coupling in a network can be described with classical electrochemistry (the Nernst potential) and a classical leaky capacitor. Since the electrical activity of neurons is what we correlate with consciousness, cognition, and behavior, looking for uniquely QM effects is kind of begging the question.

 

[.Scott]

The neuron and its coupling in a network can be described with classical electrochemistry (the Nernst potential) and a classical leaky capacitor. Since the electrical activity of neurons is what we correlate with consciousness, cognition, and behavior, looking for uniquely QM effects is kind of begging the question.

Since we don't know exactly what the neuron does, I wouldn't be so quick in presuming how it does it.

Most recently, here is an article that challenges the notion of how neurons store long-term memories:
http://elifesciences.org/content/early/2014/11/17/eLife.03896

 

[testingus]

Since the electrical activity of neurons is what we correlate with consciousness, cognition, and behavior, looking for uniquely QM effects is kind of begging the question.

It's not the simple electrical activity of neurons alone, but rather the coherent neuronal oscillations that correlate with all basic cognitive functions and consciousness. These oscillations are mediated by local and long-range neuronal communication and affect synaptic plasticity. It is still unclear how these very fast and complex changes can be explained based exclusively on well-established synaptic connections. This coherence could be a basic principle in biology that may extend down to quantum level coherence that is magnified. See the attached articles.

Plankar, M., Brežan, S., & Jerman, I. (2013). The principle of coherence in multi-level brain information processing. Progress in biophysics and molecular biology, 111(1), 8-29.
Craddock, T. J., Priel, A., & Tuszynski, J. A. (2014). Keeping time: Could quantum beating in microtubules be the basis for the neural synchrony related to consciousness?. Journal of integrative neuroscience, 13(02), 293-311.

Since we don't know exactly what the neuron does, I wouldn't be so quick in presuming how it does it.

I second this.

 

[Pythagorean]

It's not the simple electrical activity of neurons alone, but rather the coherent neuronal oscillations that correlate with all basic cognitive functions and consciousness. These oscillations are mediated by local and long-range neuronal communication and affect synaptic plasticity. It is still unclear how these very fast and complex changes can be explained based exclusively on well-established synaptic connections. This coherence could be a basic principle in biology that may extend down to quantum level coherence that is magnified. See the attached articles.

You say "coherent neuronal oscillations", I say "electrical activity of neurons" - but we are essentially talking about the same thing. Chemical synapses aren't the only thing involved - hormones from the bloodstream, genetic processes affecting receptor density and distribution, electrical synapses which allow for much faster coupling than chemical synapses, second messenger systems that change the electrical properties of the neuron by increasing the conductance through a receptor or changing the kinetics of the votage response of the receptor.

Microtubule invasion in dendrites is brought about by local electrical activity, and allows for the transport of synaptic components. Of course, microtubules are known as a structural component to cells themselves, and that's what they do in the synapse [1]. Like actin, they can increase and decrease the area of the synapse, but they are quite slow compared to the electrical dynamics of the brain. If you're looking for "very fast" it's going to be the electrical activity. Much of the "very complex" lies in the complicated electrical interference in the dendritic trees. I assume you recognize that the Craddock paper is a speculation, not an experimental result. There are similarly speculative approaches that require no new physics [2], making them more reasonable via Occam's razor. In fact, the two prominent theories are both classically based. The problem is not that classical physics can't account for the speed or complexity of neural oscillations, it's that we don't know the details of the interaction. Much like you can drop a handful of tic-tacs over and over and never predict the pattern - but you don't need any new physics, it's just the specific trajectories of each tic-tac that is lacking. This is somewhat analogous to the problem with large scale integration:

For some authors, the hierarchical organization of the brain suggests that the associative areas that mediate between sensory and motor areas provide the basis for integration (see Ref. 14 for an example). By contrast, we and others have argued that networks of reciprocal interactions are the key for integration4, 10. Among various modes of reciprocal interactions, we favour phase synchronization between the participating neuronal groups, which is certainly the most studied mechanism. Note that the terms synchrony and phase have been used in the literature with widely different connotations; here we adhere to the meaning derived from dynamic-systems analysis (Box 2 ).

Notice that the point of contention is about what neural structures are doing the integration (and in what order and direction), not really "how" the neurons are doing it.

[1] http://www.ncbi.nlm.nih.gov/pubmed/19052200
[2] http://www.ncbi.nlm.nih.gov/pubmed/11283746

 

[testingus]

You say "coherent neuronal oscillations", I say "electrical activity of neurons" - but we are essentially talking about the same thing.

I hear what you are saying, but this is like saying that "Beethoven's 5th" and "noise" are essentially the same thing because they are both sound. I agree that the electrical activity is important, but the coherent synchrony is the key.

Much of the "very complex" lies in the complicated electrical interference in the dendritic trees.

No argument here.

There are similarly speculative approaches that require no new physics [2], making them more reasonable via Occam's razor.

In regards to this, from the Varela paper you quote:

Although the mechanisms involved in large-scale integration are still largely unknown, we argue that the most plausible candidate is the formation of dynamic links mediated by synchrony over multiple frequency bands.

What generates the synchrony?

 

[Pythagorean]

Synchrony has several mechanisms. Entrainment by a cell with multiple long range outputs (as is typical of glutamatergic and 5ht neurons) can lead to spatially distant oscillations. The primary function of gap junctions (the electrical synapsed I mentioned earlier) is synchrony. There are also GABAergic inhibitory neurons connected in parallel across excitatory neurons that synchronize information streams by a synchronized inhibition. There's a review that covers some of this and references the literature as well as demonstrates that even in long distance axonal synchronization, the lag can fade away [1].

[1] https://ifisc.uib-csic.es/presentations/presentation-detail.php?indice=207

 

[testingus]

Synchrony has several mechanisms. Entrainment by a cell with multiple long range outputs (as is typical of glutamatergic and 5ht neurons) can lead to spatially distant oscillations. The primary function of gap junctions (the electrical synapsed I mentioned earlier) is synchrony. There are also GABAergic inhibitory neurons connected in parallel across excitatory neurons that synchronize information streams by a synchronized inhibition. There's a review that covers some of this and references the literature as well as demonstrates that even in long distance axonal synchronization, the lag can fade away [1].

[1] https://ifisc.uib-csic.es/presentations/presentation-detail.php?indice=207

Thanks Pythagorean, but the cited material is presentation and not a publication which can be read. Is it related to the work by Gollo et al. [1] ? If so then I have to question the results. The methods used solve differential equation based models of neurons (Hodgkin-Huxley) using Runge-Kutta based methods. This is an iterative method that implicitly assumes sychronized updating a priori, so I don't find the results surprising. You wouldn't get these results with an asynchronous updating method. This comes back to the question of where the synchronization comes from.

[1] Gollo LL, Mirasso C, Sporns O, Breakspear M (2014) Mechanisms of Zero-Lag Synchronization in Cortical Motifs. PLoS Comput Biol 10(4): e1003548. doi:10.1371/journal.pcbi.1003548

 

[Q_Goest]

You say "coherent neuronal oscillations", I say "electrical activity of neurons" - but we are essentially talking about the same thing. Chemical synapses aren't the only thing involved - hormones from the bloodstream, genetic processes affecting receptor density and distribution, electrical synapses which allow for much faster coupling than chemical synapses, second messenger systems that change the electrical properties of the neuron by increasing the conductance through a receptor or changing the kinetics of the votage response of the receptor.

I like this point. There seem to be a number of different perspectives on or descriptions of the activity between neurons. Those phenomena produced from the interactions should in principal, arise from the local interactions between neurons as described by Hodgkin, Rall or whoever, regardless of which way you want to look at or describe the interactions. If the interactions between neurons are classical in nature, they are separable, so phenomena such as “coherent neuronal oscillations” should emerge regardless of which description you choose of those interactions.

Self organization is something which arises from local interactions. It’s also defined by “weak emergence”. Assuming neurons don’t exploit any of the special features of quantum mechanics, then downward causation can’t influence the brain as a whole. Since there’s no reason to accept that the brain as a whole is some kind of quantum computer, there’s no reason to accept that local interactions are insufficient to produce all of the phenomena produced. The interactions between neurons are 'calculated' using software like "Neuron" and "Genesis" and if you look at those software, you will find they are consistent with weak emergence and self organization so they should produce all the same phenomena (in principal) as actual neurons regardless of how you want to describe those interactions.

 

[Pythagorean]

Thanks Pythagorean, but the cited material is presentation and not a publication which can be read. Is it related to the work by Gollo et al. [1] ? If so then I have to question the results. The methods used solve differential equation based models of neurons (Hodgkin-Huxley) using Runge-Kutta based methods. This is an iterative method that implicitly assumes sychronized updating a priori, so I don't find the results surprising. You wouldn't get these results with an asynchronous updating method. This comes back to the question of where the synchronization comes from.

[1] Gollo LL, Mirasso C, Sporns O, Breakspear M (2014) Mechanisms of Zero-Lag Synchronization in Cortical Motifs. PLoS Comput Biol 10(4): e1003548. doi:10.1371/journal.pcbi.1003548

Step sizes for the Runge Kutta algorithm are much much smaller than the mechanisms of synchrony so that they can be taken to be continuous; this is a rather basic ptinciple in numerical analysis. You can test the stability of your functions and choose an appropriate algorithm if RK doesn't work. But besides that, there is also experimental evidence. Synchronization via typical neuronal mechanisms isn't exactly controversial. None of the mechanisms are impossible in the large scale, we just don't have the fMRI resolution to determine what mechanisms are playing what role.

@Q_Goest

Hodgkin Huxley would be appropriate for certain types of neurons for a short time scale, but you'd have to make adaptations to the model to consider hormonal and genetic changes. In fact, the HH model doesn't even include a synapse. It is only an axon membrane. Of course, it still does all imply weak emergence as, you say. All these adaptations have been made in various forms. My own research includes a lot of second messenger modulation on kinetics.

I'm not sure how well we can compare simulations to a real brain. The brain is doing a lot of fantastic things, the computer is flipping registers around and generating an abstract symbol that we interpret in a relevant way. It's just a very sophisticated note pad and pen. To generate consciousness then, our best bet would be to physically reproduce the brain as best as possible.

It would be nice to see the results of some of the popular thought experiments. What happens if you replace a conscious human's brain one neuron at a time? What happens if you separate the neurons spatially while keeping the information processing the same?

My prediction for the first is, of course, nothing. The second is interesting: if separating the components destroyes consciousness, I'd be inclined towards an electromagnetic field theory of consciousness. If not, then we have to stick to the information theory approach.

 

[.Scott]

Since there’s no reason to accept that the brain as a whole is some kind of quantum computer, there’s no reason to accept that local interactions are insufficient to produce all of the phenomena produced.

There is a reason to accept that the brain employs quantum superpositioning to process information. It's easy to explain - but perhaps not so easy to understand. In order to be conscious of something, the information describing that something has to be in one state - not divided into discrete symbols. Classical information processing doesn't provide a mechanism for that.

 

[Pythagorean]

There is a reason to accept that the brain employs quantum superpositioning to process information. It's easy to explain - but perhaps not so easy to understand. In order to be conscious of something, the information describing that something has to be in one state - not divided into discrete symbols. Classical information processing doesn't provide a mechanism for that.

That's kind of begging the question isn't it? What is the reasoning that "In order to be conscious of something, the information describing that something has to be in one state"?

 

[.Scott]

That's kind of begging the question isn't it? What is the reasoning that "In order to be conscious of something, the information describing that something has to be in one state"?

Obviously the character string "tree" is not enough (or the right kind of) information to sustain the conscious experience of a tree - but let's use it as an example anyway. If you had four neurons, each with one of the letters, which neuron would have the experience of "tree"? In order for "you" to be conscious of "tree", there has to be one place that has all the information.

If you write "tree" on a piece of paper, obviously there is no consciousness of "tree" because the information is divided. You can make the paper as small as a neuron, but it won't help. You can process the information - reencode it, photocopy it, etc - but none of that will create a situation where the entire "tree" is in one state - and so there is no place for the "tree" experience to exists.