A Physicist Thinks About Autism

I’ve been thinking quite a lot about autism for the last little while, because I am an autism Dad. My younger son has autism, severe developmental delay, attention deficit, hyperactivity and epilepsy, a heavy burden for anyone. It is amazing that he is still full of sunshine and smiles.My wife and I are also professional Physicists, and so as practicing scientists have been looking around at the medical literature on autism. We can see that there is a spectrum of different symptoms, all lumped together as “The Autism Spectrum”. Now to a Physicist, when spectrum is mentioned, we usually think of the electromagnetic spectrum (visible light, infra-red, radio waves and so on). The electromagnetic spectrum is well understood as a product of Maxwell’s equations, and so the properties of the different types of radiation can be deduced. The autism spectrum is a lumping together of many observations, and so appears to be a term simply to mean a collection of things which have similar characteristics. From looking at various drug trials, it is fairly reasonable to assume that there are several (or indeed many) different medical conditions, all of which produce similar symptoms to varying degrees, and so in the future, what we know call “Autism”, will actually get split into several or many distinct diagnoses. Some of these conditions will give rise to “higher functioning” Autism, and some to more severe kinds, such as our son has.

Spherical Cow by Andrew Robinson 2015

Now physicists are known as Reductionists, we like to simplify things down to the simplest possible model of what is going on, and see whether we can explain things using that model. If that fails then we try more complicated models. This seems very logical to trained Physicists, but can lead to some assumptions that look from the outside as rather silly. After all, Physicists are the people who devised that marvellous construct, “The Spherical Cow”. Even Physicists realize that this is a gross distortion of reality, but nevertheless using simplifying concepts can give us insight into physics if used correctly. It has the added bonus of giving us a few jokes too. Do remember that I am not an expert on the brain, by any stretch of the imagination, and what I am doing here is trying to use a very simple spherical cow model to talk about important features of the hugely complex brain (Real neuroscientist please have mercy).

So here is my “Spherical Cow” simplified model of the brain. Let’s assume that the brain functions properly (loosely defined) when certain neurochemicals are in a certain balance. There will be different balances across the human population, to be sure, but we assume for the general population, there are a set of “normal” states and a set of “minority” states where there is autism, (or indeed other mental illnesses). Since there are a large number of states, it looks like a continuous spectrum, but if you are able to look closely enough you will see the underlying state structure. Physicists are used to this concept because we routinely consider individual atoms or particles, where we have to apply Quantum Physics, but when we put large numbers of atoms together, the quantum states blur into a continuum to give us Newtonian Physics. Now imaging that every state can interact with every other state and in fact the states change by well-defined rate equations. Chemists and biochemists will recognize this as reaction dynamics and reaction equilibrium. In reality, there may be other outside events which are influencing the states of the brain, but that is another issue.

So let’s start with a three state brain. There are 6 possible reaction paths to convert between the states. You can consider these as reaction rates, or probabilities that state A will convert to state B in a given time. It turns out that the number of pathways increases rapidly as you add extra states, so that for N states, you need N x (N-1) pathways.

Three states lead to six reaction pathways

I actually adapted a piece of software, originally written to simulate activity in a three-quantum level laser, to do this. The laser model is actually a bit easier than the one we are discussing now. For the computing people, it’s a short code written in Python and is available on request.

Now one of the interesting things about this three level brain is that you can find combinations of states which become stable over time. Here is one of these.

There is only a trace of the blue state at the start, but eventually it dominates

In this case, we start with equal amounts of the green and the red states, and a tiny amount of the blue. As the system runs the levels of red and green eventually drop to zero, while the level of blue increases to a maximum level. This equilibrium position has only blue in it.

But now watch what happens if we use exactly the same situation, but start with no blue at all:

With no blue to start, the red state can now maximize.

You can see that blue never appears in this equilibrium, and so in its absence, the red state now takes over and becomes the dominant state.

So we can say that using the same model, the starting conditions can change the resulting equilibrium. Now imagine a baby developing in the womb. A change in one vital neurochemical can produce a different final state from normal. Sound familiar?

So now try and imagine the level of complexity of say, a 100 different neurochemicals, all of which may be present in a continuum of concentrations. That means we can see how difficult it is to track down the causes of the conditions underlying the Autism Spectrum. We will not truly understand it until we have a much better understanding of the neurochemicals, their interactions with each other and with other parts of the human body. But we must push for this understanding. Only then will we be able to make a better job of predicting which medication will be effective in those circumstances. Those of us who have autism children know that the medical approach is to start at the top of the list of possible medications and work down until you hit something that works. We will only do better when we classify the conditions much better, and naturally have better diagnostic tools to do so. I have no doubt that this can be achieved, given enough time and enough resources, but it is one of the big challenges for the 21st Century.

There are even some chemical reactions which show an oscillation between two equilibrium states over a period of time. This sounds remarkably like some types of neurochemical imbalances which come and go leading to mood swings or personality changes.

Now there is a lot of controversy about the stance of the US-based organization “Autism Speaks” which advocates for “A Cure for Autism”. Many people with higher functioning autism are very, very unhappy with this, as they are comfortable within themselves and don’t like being classified as needing curing. I agree with them. But then I also have a low functioning autistic son, and his quality of life would certainly be enhanced by some mitigation of his symptoms. So I see research into this area as being necessary. But there is a huge ethical question on treatment. I would certainly allow him to have medication to try and mitigate his symptoms. But higher functioning individuals with autism must have the right to choose.

Anyhow, this may be a moot point. It is entirely possible that once the brain has reached a stable equilibrium it might be difficult or impossible to push it into a different (more normal state) without serious side effects.

So there you have an oversimplified spherical cow model of activity in the brain. I hope that it makes sense to you, and I hope that it gives you some insight into the complexity of autism research. I salute all of those scientists and physicians working in this area. Thank you for your efforts. They are much appreciated.

Comments are always welcome.

Like what you read? Give Andrew Robinson a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.