Last time we finished on synapses and today we are going to finish our introduction to neuroscience. You see, all this is just the tip of the iceberg. The whole idea of the series is to learn at least something about our “neural network” so that we can draw inspiration from it and maybe even recreate it.
So today you’ll learn about different kinds of neurotransmitters, regulations in our bodies, experiments on frogs, and axon’s ability to regrow.
Neurotransmitters are chemical messengers that transmit a message from a nerve cell across the synapse to a target cell. The target can be another nerve cell, or a muscle cell, or a gland cell (a group of cells that synthesizes substances). They are chemicals made by the nerve cell specifically to transmit the message. Neurotransmitters can be peptides but they’re largely amino-acids or nucleotides. And there can be many types of neurotransmitters per single synapse.
In the previous article, we said that there are excitatory and inhibitory synapses but synapses are defined by the neurotransmitters they produce. And so there are excitatory and inhibitory neurotransmitters.
Examples of excitatory neurotransmitters are glutamate (involved in being alert) and adenine (involved in being sleepy). And, for instance, one of the things that caffeine does is counteracts the effects of adenine.
As for inhibitory ones, there are, for example, GABA (gamma-aminobutyric acid) and glycine which are the two major central neural system neurotransmitters. GABA, among other functions, is an anti-anxiety neurotransmitter and so increased GABA activity can have a sedative effect. And glycine participates in the processing of motor and sensory information that permits movement, vision, and audition.
There are also neurotransmitters that both excitatory and inhibitory. Such as acetylcholine, which is involved in neuromuscular connection and also found in the central neural system, serotonin and dopamine, which both kinda make you feel good, we will talk about dopamine in the next article.
You may already have guessed but I’ll note it explicitly, neurotransmitters move the action potential of neurons. Inhibitory ones bring the action potential value (I really can’t stop thinking about it in math terms) away from the firing threshold and excitatory vice versa.
If something sends those neurotransmitters (basic signals), then some other thing must receive it. The way neurotransmitters affect what they’re going to do is by binding the appropriate receptors. There are many types of receptors and there can be more than one type of receptor per synapse. But we’re sticking to the basics so here are two classes you should definitely know about.
The first one is the ionotropic receptor (ion channels), they’re the receptors that bind the neurotransmitter and open up. And they have a very rapid response (in the millisecond range).
The second class is the metabotropic receptor, they’re not ion channels and they actually work by binding the neurotransmitters which activate a second messenger system, which in turn, later on, opens an ion channel. And their response is much slower (often much greater than a second).
Once again, one neurotransmitter can activate more than one type of receptors. This was implicit in the notion that neurotransmitters could be both inhibitory and excitatory. For instance, serotonin binds up to no fewer than 14 receptors. And one neuron can express a whole bunch of serotonin receptors.
So as you see, our bodies tend to have many-to-many connections. Imagine the scale (millions and billions of neurons, synapses, dendrites, neurotransmitters, and receptors) and you’ll probably get why can’t we model our brain whatsoever.
We’ve already said that our bodies tend to have more chemical synapses rather than electrical synapses because the chemical ones are highly regulatable. And so now we’re talking specifically about these chemical synapses and how they fine-tune responses of our nervous system to the smallest stimuli both from within our body and outside.
There are a couple of ways to regulate synapses, at least that we know of, but they all have to do with changing the number of neurotransmitters.
The first is degradation which means that when a neurotransmitter is released, in some cases, it’s degraded by specific enzymes, for example, there’s a particular enzyme that breaks down acetylcholine. And note how gentle the system is, if that enzyme is inhibited, you go into respiratory shock, you cannot breathe anymore because you have to activate and inactivate the muscles via the nerves, as you breathe. And acetylcholine does activate your muscles.
Another way to change the number of neurotransmitters is re-uptake (reabsorption) by a presynaptic cell. When the neurotransmitter is released it can get taken up by the presynaptic cell and that’s a kinda prudent way of doing things because the cell doesn’t need to keep synthesizing neurotransmitters anymore. It’s exactly the case for serotonin and dopamine.
And in some cases you can regulate the synthesis, the number of neurotransmitters that’s being made. A big class of neurotransmitters regulated in this way are the endorphins which are the natural opiates of the body, basically, pain regulators.
All of these processes are regulatable both for modulating normal synapses and also in all cases for medication targets. Long story short, our body changes the number of neurotransmitters as a response to stimuli but also drugs change their amount.
Let’s take a nerve gas for instance. During World War Two the US used a gas called Sarin against Japanese troops hiding underground. This gas inhibits acetylcholinesterase which is supposed to breakdown acetylcholine in a synaptic cleft. Therefore receptors get repeated stimulation with acetylcholine which leads to respiratory paralysis and death.
However, it can be countered by atropine, a competitive inhibitor of acetylcholine. It binds to the receptors and prevents acetylcholine from binding thus saving from death. This is actually how US soldiers protected themselves from the gas, they had vials of atropine. So in case of a nerve gas attack, if you inject yourself with atropine, you’ll stop the acetylcholine from working and survive, however, you’ll probably be kind of floppy.
The other way to modulate how often the post-synaptic cell is stimulated is by modulating the receptors. Intuitively, if you have more receptors, a neurotransmitter has more place to bind and thus a greater signal can be sent.
So there are three ways to do it, you can change the number of receptors, the affinity for a neurotransmitter, and the receptor’s responsiveness. And all of these three things have something do to with learning, memory, and addiction, these changes are very slow, they occur over minutes, days and weeks.
One of the outcomes of changing all of these parameters about the receptors is that over a long period of time you really change the way a synapse works. This is essentially what practice does. You tweak how a synapse works through these parameters by repeatedly simulating that synapse. For instance, when you play the guitar you little by little alter the synapses that allow you to engage in this activity.
And so there are two possible outcomes in which synaptic response changes due to repeated synaptic stimulation. It’s either increased or decreased response. The increased response takes place at excitatory synapses, it means that the action potential is more likely to happen and this process is known as long-term potentiation. The decreased response, on the other hand, happens at inhibitory synapses, and the process is known as long term depression.
It’s believed to be the way memory works. In terms of reinforcement learning, the connections that lead you to a higher reward are getting stronger and vice versa, the connections you make mistakes with are getting weaker. Just like in artificial neural networks, the lower the weight of a neuron, the less impact on the outcome it does.
Another concept we should talk about is neural circuits. Neurons in our brains don’t just act on their own as single units but form groups of interest, so to speak.
Just look and these brain scans, we apply somewhat different regions of our brain for each and every task. Tens of thousands of neurons are used to process what and how you see and say things. And it’s only the brain’s activity we’ve measured on the surface, also known as the neural cortex, imagine what happens deeper.
However, the real questions here are how the circuits formed and how neurons know where to connect. Let’s think logically, it’s either a random process, which sounds a bit dumb or a guided process, which sounds much better.
One could argue that evolution is all about mutations, and mutations are random, so why is stupid to consider that throughout billions of years we randomly developed all the neural circuits? The thing is we are constantly developing new connections inside us, as we’ve already said, a brain can significantly change due to learning and perfecting new skills. And not only that, there are many vital processes that must be restored in case of destruction and so how would a random process help to do it, just randomly do stuff until it finally works?
Long story short, both the random process (survive if a connection was made) and the guided process (told where to go) of setting paths and connections inside our bodies after many years of research turned out to be correct.
Sperry’s frog retina rotation experiment
One of the most notable experiments in this field was conducted by Roger Wolcott Sperry in the 1960s. He took a frog, cut out its eye, rotated it 180°, and placed it back to see if and how neural connections restore. Pure frogs, right?
There are two regions of frog’s retina, and ours as well actually, nasal, located close to a nose, and temporal, on the other side. Those regions are respectively connected to two regions of optic tectum which, among many other functions, is responsible for directing eye movements. During the experiment those connections were cut and, spoiler restored themselves after some time.
One of the possible outcomes is that axons, whose jobs are to lay those paths, randomly stumble upon some region of a brain and connect to it but turns out axons “know” where to grow. That’s why the frog was able to restore its vision and eye movement and Parry got a Nobel for this experiment. So, I guess, everyone won.
Not only do the guidance of neurons (axons) occur during the development and repair of our bodies but scientists also think that it happens during learning new skills as well. So some of us might literally have a more complicated system than the others. And it nicely fits into our ongoing metaphor about computer systems and human brains, a system which is capable of many functions is usually more sophisticated and advanced than less developed analogs, so constantly learning a new skill and widening your horizons may be a very good idea.
It’s been discovered that a neuroblast, basically a forming neuron, during its development sends out processes that are called neurites which initially look the same but later on one of them becomes an axon and the rest become dendrites. This is when the pioneer axon (the first who finds the target) is formed and the others follow the same path which creates a bundle of axons, or fascicles, and together they make a nerve.
But how does the pioneer axon seeks the path? On the very tip of the axon, there is a thing called the growth cone which is similar to a detector of some kind.
Microtubules both stabilize the axon and also transport substances to and from the cell body like little railway tracks. At the very tip, they interdigitate with these finger-like protrusions which are very dynamic, I’ll show you a video soon, and they protrude because of polymerized actin (F-actin).
There are receptors on the surface of the growth cone that is sampling what ligands are in the environment and if the ligands are favorable, they will stabilize those protrusions and make more of them take place.
The growth cone is paramount to axon outgrowth, it’s like a navigation system which catches signals that attract axons towards them or signals that repel axons. F-actin increase in the environment creates attractive signals and make the growth cone extend. As for repulsive signals, the growth cone collapses due to F-actin becoming G-actin or unpolymerized actin.
The ligands are actually the guidance signals when they bound to a receptor and it leads to signal transduction, hence axon outgrowth.
It reminds me of the search algorithms we implement in computer science. It goes one direction, realizes that the heuristic isn’t working out there, switches to a new road, and repeats it until the path is found.
Neuroscience is an amazing, mind-boggling, and incredibly complex field of study. I’ve tried to tell you as much as I could about our brain but I am no expert, so there might be inaccuracies. However, I urge you to explore the subject and draw inspiration from it.
Next time we’ll be talking about dopamine, the reward system of our bodies, and reward functions in reinforcement learning, so stay tuned.
 Hazel Sive, Tyler Jacks, and Diviya Sinha. 7.013 Introductory Biology. Spring 2013. Massachusetts Institute of Technology: MIT OpenCourseWare, https://ocw.mit.edu. License: Creative Commons BY-NC-SA.
 Richard S. Sutton and Andrew G. Barto. 2018. Reinforcement Learning: An Introduction. A Bradford Book, Cambridge, MA, USA.
 Peter Dayan and L. F. Abbott. 2005. Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems. The MIT Press.