Thursday, November 03, 2016

Matt Fisher: brains may use phosphorus nuclear spins to act as quantum computers

First, off-topic: On Sunday, during a floorball match, I was encouraged to stand on my left foot improperly. A bone [left, base] got really broken, X-rays showed today. (That's not the first broken bone: I broke a rib during the same sport in 2009.)

This modern replacement for an orthopedic cast is thinner, more efficient, and also more resilient. So I can do something resembling walking without crutches. A disadvantage is the extra $30. Compassionate readers may help their handicapped humble correspondent via PayPal (thanks a lot to Umesh who did!).

Matthew P.A. Fisher is a top condensed matter physicist in Santa Barbara; check his impressive trace in the literature. In the article
A New Spin on the Quantum Brain
in the Quanta Magazine, Jennifer Ouellette describes a provoking or exciting paper (arXiv) in Annals of Physics (Google Scholar) a year ago, the same Brian-Greene-led journal that published some deceitful stuff by Joy Christian.

The question is whether the human brains use some tricks of quantum computation. That could explain why we still seem smarter and more creative (or better in recognition and similar tasks) than computers with existing programs – even though the computers seem to have much more computing power. Most people in neuroscience say (and I tend to believe) that no such enhancement is needed to explain why our brains are pretty good. But a priori, it's possible that the intrinsically quantum operations with the quantum entanglement are relevant in the brain.

As soon as you start to answer the question quantitatively, however, you're led to the sad conclusion that the quantum computation is impossible in the warm and wet environment of the biological brain. For (man-made) quantum computers to work, the decoherence – the disappearance of the information about the relative quantum phases – due to the interactions must be dramatically suppressed. And the wet and warm environment of the brain seems to induce too many interactions that make the decoherence way too rapid.

One wants to be more careful, however. In the previous paragraph, you were probably imagining some order-of-magnitude calculation that is using some typical size of the hypothetical qubit in the brain, a single number describing the strength of its interactions with other things, and you got a single time scale estimating how quickly the relative phase disappears. And it disappears very quickly.

However, as Matt Fisher appreciated, this way to "debunk" the quantum computer inside the brain may be too sloppy because there are different components in the brain, different degrees of freedom, and they interact with interactions of vastly different strengths. Maybe there are some qubits in the brain whose coherence is protected because their interactions with the rest of the brain are weak enough.

The Quanta Magazine article tells us that Fisher started to think about these neuroscience matters after he was depressed for a long time and took some drugs. He realized that there are various papers claiming that "quantum mechanics at longer distance scales" is relevant in photosynthesis and other quantum biological processes. Also, he saw some papers indicating that the biological behavior depends on something non-chemical about the cells, the nuclear spins. In particular, there were some experiments showing the dependence on whether you have lithium-6 or lithium-7 somewhere.

Fisher has actually spent several years trying to make his theory that nuclear spins carry the quantum bits as convincing as possible. For some reasons, he decided that the phosphorus nuclei are the best to carry the nuclear spins and the protection of the coherence is even better when they're incorporated into the "Posner molecule" \({\rm Ca}_9({\rm PO}_4)_6\). And he's been really trying to combine the molecules into even larger systems to further suppress the decoherence.

The whole strategy is rather clever. Note that the nuclear spins interact really weakly. One known manifestation of this fact is that the hyperfine structure of the spectral lines – which is due to the interactions of the nuclear spins – is finer than fine. One may imagine that some nuclear spins interact really weakly – and decohere really slowly – and they may be combined in ways that makes the decoherence even slower, perhaps much slower. Moderate skeptics say that Fisher may get up to seconds; Fisher himself believes that the quantum coherence may last up to hours or days or longer.

This slowdown has another face that doesn't seem to be discussed there. When the interactions are too weak, it's also harder to measure the qubit. So I suppose that there would have to be several stages how the molecules are being used. They would have to be shielded using Fisher's techniques during the "quantum computation" and then the shield would be opened and the measurement would become possible.

It's sort of plausible. Fisher says that the nuclear spins are the only exception that could enable brains act as quantum computers. He may be right. At least, it's almost certainly the only seemingly defensible exception that we know at this moment. But is it right?

If it's right (and I would give it some 5% probability – a somewhat random number: probably not but not "insanely unlikely"), there should be explanations for several more questions. How is it being used and how complex quantum algorithms are running inside our brain? And also, how did the quantum computation evolve during Darwin's evolution? And when did it evolve? What were the first organisms (or what are the simplest animals or plants) that use this quantum computation involving nuclear spins in some way?

One reason why I am somewhat skeptical is that I don't believe that there was an evolutionary pressure to develop complex mechanisms to "shield qubits really well" and to "allow the system to switch between shielding and measuring", as explained above. You know, if monkeys 10 million years ago had to compete for bananas and the winner would be the monkey who cracks an encryption code of a bank by factorizing a large integer by Shor's algorithm, it would make sense that a mutation of the monkeys that has a quantum computer with Shor's algorithm in the brain would beat the less sophisticated monkey competitors.

For some reason, I believe that the monkeys or our more distant ancestors were not competing in their ability to crack cryptographic codes. They were mostly competing in tasks for which simple classical computers inside their brains were more than enough. I may be wrong, of course, but I would like to be shown why I am wrong if I am wrong. If I am right, then it's unlikely that the quantum computer would have evolved inside the brains. Like in many creationist-evolutionist debates, however, the pieces of the quantum computer could have evolved for some other reasons that make sense. And it could have been easy to combine the pieces quickly.

At any rate, it would be good if some competent people were looking into this highly intriguing Fisher's idea. If it has a chance to be true, it could help us both to understand "ourselves" as well as construct viable man-made quantum computers.

No comments:

Post a Comment