You might not realise it but this sentence you are reading is changing your gene expression.

Any man who can write a sentence like that and mean it is awesome in my book. Anthony Bell is such a man. He is a theoretical neuroscience researcher working on unsupervised machine learning and processing. In my opinion, that title goes up there with planet hunter and dolphin linguistics expert as being among the coolest jobs in science. Around ten years ago, he wrote a paper about Levels, Loops, and the Future of Artificial Intelligence. He readily admits that it was in some ways written negatively but there are still plenty of great points to be taken from it. (And if you want to see what exactly it is that theoretical neuroscientists do, check out this library of videos.)

I think the reason I like Bell’s paper so much is that it touches on a lot of points quickly and easily. It starts with an overview of the optimistic sci-fi-ish scenario of the future given by people like Raymond Kurzweil. But he then goes on to say this largely assumes a perfect functionalist (computer-like) account of the brain can be readily given. But as far as objects go, not too many are deterministic and discrete in time and space in the same way a computer is. As he mentions: “The whole state of the machine [a computer] at the digital level may be written down. No natural objects seem to be of this nature. The computer is really a physical instantiation of a model. We know a model can computer, but can it live or think?”

The problem with the perfect functionalist account is the question of what level do you model at. How do you “write down” the state of a brain? What exactly is corresponding to the logic gates of a silicon chip?

If we wrote down the sequence of all spikes of all neurons, would that be enough to specify the ‘neural computation’? Do molecular and biophysical processes exist to implement a ’spiking computer’ at the neuron level?

Bell thinks not.

While no specific physical processes below the gate-level of a computer interfere with the model-like operation of the computer (unless something goes wrong), this cannot be said at the neuron level of the brain. Molecular and biophysical processes control the sensitivity of neurons to incoming spikes (both synaptic efficiency and post-synaptic responsivity), the excitability of the neuron to produce spikes, the patterns of spikes it can produce and the likelihood of new synapses forming (dynamic rewiring), to list only four of the most obvious interferences from the subneural level. Furthermore, transneural volume effects such as local electric fields and the transmembrane diffusion of nitric oxide have been seen to influence, respectively, coherent neural firing, and the delivery of energy (blood flow) to cells, the latter of which directly correlates with neural activity.

Essentially, he says that it would be easy to believe that the neural level is the only level needed to copy down in the brain. But it is just not the case. There are too many complications with viewing the brain in this way.

So what are the other possibilities? Groups of neurons, spiking together could smooth out the noise and other interference of the neural level. Perhaps that is a good place to model intelligence. But this level is also flawed. For there is a large “failure [by researchers] to appreciate that noisiness is in the eye of the beholder… in the case where a stimulus is presented and that part of the neural response which does not correlate with the stimulus is regarded as noise, we have a situation almost as bad as thinking French people are stupid because they produce strange noises in response to questioning.” Ha! Did I mention he’s funny, too?

There’s also the molecular level. A description of all known molecules in the brain could (potentially) open doors. But

The location of the molecules are important. Testing of enzyme reactions in bulk phase (solutions in test-tubes) is partly responsible for an impression that in the cell, molecules largely jitter around with Brownian motion and sometimes bump into each other and react. What turns out to be more likely is that most reactions take place locally in membrane-associated protein complexes, and the product of one reaction is passed directly on as substrate for the next. Evidence for this detailed spatial organization, called metabolic channeling, is accumulating.

And even if we could write down the position of every molecule:

There are sub-molecular interferences that violate the separateness of the `molecular machine’ level, and they are quantum effects. Two examples of this are electron transfer in photosynthesis and the energetics of enzyme interactions. In both cases, quantum coherences are necessary to explain the efficiency of the reactions. But we don’t even need to go as far down as quantum effects, because proteins do not end at the edges of the black and red balls of which ball-and-stick molecular models are constructed. Their electrical fields extend into the surrounding water molecules, orientating them to form what is called structured water. Structured water is also important in determining how enzyme reactions occur, and how ion channels are selective to certain ions. To argue that one piece of structured water or one quantum coherence is a necessary detail in the functional description of the brain would clearly be ludicrous. But if, in every cell, molecules derive systematic functionality from these submolecular processes, if these processes are used all the time, all over the brain, to reflect, record and propagate spatio-temporal correlations of molecular fluctuations, to enhance or diminish the probabilities and specificities of reactions, then we have a situation qualitatively different from the logic gate. The variables lying beneath the level of a molecular `gate’ can affect the behaviour of the gate, so the functionalist is again frustrated, and the notion of the brain as a molecular `computer’ can be viewed as no more than an analogy, and an inaccurate one.

These are all points that I tried to make (more clumsily, as I am not a neuroscientist) in various posts about mind uploading, the analogy of the brain as a computer, and dualism. These are not opinions. They are scientific facts. “A computer is an intrinsically dualistic entity, with its physical set-up designed not to interfere with its logical set-up, which executes the computation. In empirical investigation, we find that the brain is not a dualistic entity. Computer and program may be two, but mind and brain are one. The brain is thus not a machine, meaning it is not a finite model (or computer) instantiated physically in such a way that the physical instantiation does not interfere with the execution of the model (or program).”

Basically, the processing of the brain interferes with its physical structure; the two are one in the same. A computer does not do this. If we could build computer chips that were fine with thousands of their logic gates dying every day or were able to change the system of connections within them, we might be able to recreate a brain. We can see that the substrate does matter. I still think that intelligence can exist in silicon but I am more and more convinced that there is not way to separate the “software” of a human brain from its “hardware”. There doesn’t seem to be a real distinction between the two.

The other major thrust of Bell’s paper furthers this claim. He talks about the feedback loops that exist everywhere in nature (and especially in our brains). As he says:

feed-forward processing in the nervous system is the exception rather than the rule, and often what looks feed-forward contains complicated feedback systems at a different level of analysis. For example, the spikes of a cortical neuron have now been seen to extend far into the dendritic tree, affecting, through voltage- dependent channels, the integration of signals from synapses. This destroys the illusion that the neuron works like a directional `neural network’ neuron, performing a weighted sum of its input signals.

His claim is that we need to stop thinking of processes in the brain as feed-forward situations. It is very tempting to say that X causes Y but:

X may rise, causing Y to rise, but then increased Y usually causes X to diminish, directly or through some other variables Z. These cycles of positive and negative feedback are universal in biology and cause equilibrium values of X and Y, or stereotypical dynamic behaviour to occur. A neural spike is one example of a transient dynamic caused by positive and negative feedback, where X is the sodium current and Y the potassium current.

Feedback loops can sometimes be quite wide (X causes Y through Z causing W causing V causing U etc…) and all of the subsequent steps might have effects themselves. As Bell says, this is why it is so hard to develop drugs that specifically pinpoint one brain process (like happiness or depression) with a single chemical. All of those “side-effects” of the chemical are cyclic processes that are happening in the brain and affect the brain process at hand.

I know that I am going into a deep level of detail with this paper, but I just find it so fascinating. One point that Bell touches on relates to areas of the Singularity I haven’t paid much attention to (namely genetic engineering and molecular technology) because I feel that they aren’t as important and there are plenty of other commentators discussing their problems already. In particular with genetic engineering, Bell takes issue with the idea that our genome is somehow a master controller of our selves.

Arguments pointing back to the genome as the causal factor behind animal behaviour and intelligence are so universal in our culture, that to allow the genome special status outside feedback cycles would be to endorse a control-node mysticism rivaled in shape and form only by that of the monotheistic Anglican bishops who debated so famously with T. H. Huxley. (When science became a greater authority on human origins than the church, the transition hid the fact that it was a change of government without a change in policy.)

The central dogma of molecular biology is wrong! Sequences of DNA code for strings of amino acids – true – but how these amino acids are assembled into functioning proteins and which parts of the DNA are read in the first place are both controlled by proteins, and depend on the state of the cell and its type. It’s as if there was a bookish town (a cell) with a central library (the genome) and people (proteins) who came in to read short sections here and there, share with each other what they had read, and use the knowledge to build and change the town. Who is controlling here – the townsfolk or the library? (Answer: neither.)

DNA is not some master controller with perfect knowledge of the outside world. It is a part of a whole interactive reality. It is shaped by that reality just as much as it shapes it. “The organism and its genes are caught in a cyclic dynamic, and if the organism decides to spend its afternoon in a (real) library, instead of attempting to father children, then you can be sure that the pattern of gene expression will alter accordingly.”

It is the same with neurons and information. In one sense, the neurons give rise to your intelligence. But your intelligence also affects the neurons. What you choose to read or discuss prunes and alters the connections in your brain. Who is doing the controlling here? Where does the “information” arise? The answer is no one and nowhere.

The paper ends with the conclusion that in the future:

There will be no nanotechnological robots running around inside our bodies, at least none that are any more wizardly than the non-machine-like molecular complexes that already exist. There will be no `control node’ drugs that can pin us on the right end of the sadness/happiness spectrum, and thankfully we can drop this one-dimensional view of the human emotions. There will be no people living without brains, as digital patterns in the Internet. There will be no spiritual machines, models so advanced that they can deduce things that we find mysterious. There will be no machines with minds.

I honestly don’t know about the last sentence. I think this is what Bell might have been refering to when he said that the paper had a negative edge to it. I still do believe that there can be minds in a machine, some instance of intelligence brought about by complex programs. But it does seem that maybe this idea of reverse-engineering the brain is a lost cause. The brain already does a much better job than we could recreate. If and when we do create AI, it will come from a different evolutionary background. It will be on a different substrate, with different rules. It be so foreign, so alien to our way of thinking that I think we might not recognize it at first. It’s possible that there will be long academic debates as to whether or not we’ve actually made it at all. It will not neccesarily be more intelligent or less intelligent. It will simply be.