The inner workings of the human mind

Our mind is an extraordinary machine. In an engineering perspective, it’s the holy grail of efficiency and effectiveness. For many years, cognitive scientists have wondered about what makes us ‘intelligent’. Does the brain’s giant neural network implement the ultimate learning algorithm?

Have you ever wondered how the mind works? How this mysterious entity allows us perceive the world, understand it and take action to achieve our goals? Well, humans have been asking this question for centuries. Philosophers, psychologists and even religions have their own theories.

Cognitive scientists study how the brain processes information, trying to figure out how this complex biological system demonstrates intelligent behaviour and gives rise to the mind. Surprisingly, its working principles are quite simple, and from a design perspective, quite delightful.

We don’t fully understand every single mechanism of the brain. What the latest research tells us are the basic principles of the mind. They answer questions like: How can we make such accurate and seemingly automatic predictions? How can we learn new things? What is behind the decisions we make?

We try to explain these principles in a very simple language. For everyone to have a better understanding of their mind.

The big network

The brain is constituted by loads of different types of cells, but the crucial ones are the neurons. These cells are connected to each other by synapses and this forms an extraordinarily complex network.

This network is what gives rise to the mind, everything you are and everything you know is encoded in that network. Other cells are just there to support and maintenance.

According to studies, there are about 86 billion neurons. That’s a big number, right? How big? Picture a tower, formed by stacking one billion sheets of paper. This tower will be as tall as the tallest building in the world, the Burj Khalifa. A recent estimate, tells us there are about 1,000 trillion connections (synapses) between neurons.

Obviously nobody knows the map of the network (i.e. how the neurons are connected to each other), although a recent project, called ‘connectome’, is trying to figure it out. Good luck! We do know that the map is different for everyone. Simply because everyone has different personalities, memories, etc…

The light

This network is awesome, but what does it do? Lets think of a neuron as a light bulb, it can be dark, or have be as bright as you want. The bulbs are connected by wires which let the electricity flow.

Imagine now that we light one bulb, if the bulb is bright enough, the bulbs that are connected to it will also light, with slightly less brightness. This will provoke a cascade effect, bulbs will make other bulbs light as well, with even less bright, and so on, until the bulbs have not enough brightness to light their neighbours.

Once we light a bulb, it doesn’t stay bright forever. Instead it starts to decrease it’s brightness until it is dark. A certain bulb lights when we think of the specific thing that the bulb represents.

We believe that the brightest light is our consciousness, and our current thought is what the brightest bulb represents. The unconscious is the rest of the bulbs, some are slightly bright, some are dark…

Keep in mind that this process happens incredibly fast. A neuron will ‘light’ for only a few nanoseconds.

Weights

The synapses that connect the neurons in our brain have a very peculiar characteristic. They can enhance or diminish the intensity of the electricity that goes through them. So, if bulb A is bright, and is connected to another bulb B by a wire that enhances intensity, bulb B will be brighter than bulb A.

Some wires enhance the intensity a bit, some enhance it a lot, some diminish it a bit, some diminish it a lot and there are even some that neither enhance or diminish it, they just transfer the same intensity from one bulb to another.

The change in intensity performed by a certain synapse is defined by what we call the synaptic weight. This weight is always changing according to what we learn and observe.

Now, take a moment to imagine how this light flows around the network when we light some bulbs. We can light whatever bulbs we want and even multiple bulbs at the same time. Since the connections change the intensity, by only lighting one bulb in the correct place we could keep the light flowing forever.

Patterns, patterns everywhere

Our universe is basically composed of subatomic particles. These particles are arranged in certain ways to form different atoms. These atoms are arranged in certain ways to form different materials. These materials are arranged in certain ways to form different objects. In a bigger scale, anything we build is made of smaller, simpler objects, arranged in a certain way.

But this is in space, what about time? Well, if I want to make a rice, I first put the pot in the stove, I put the oil in the pot, I add diced onions, I let the onion fry until it’s golden, I put the rice, I fry it just a little bit, I add boiling water, I wait until it’s cooked and voila, the perfect rice! I can do this over and over again and I will always have the perfect rice, of course I need to pay attention to the quantities and the timing. This arrangement of actions will give me a perfectly cooked rice.

Think about it for a while, and you will realize that everything can be thought of as a pattern. Everything is defined by the arrangement of it’s smaller parts. And that arrangement, along with the correct identification of the smaller parts, forms a pattern.

So, why are patterns useful? Well, they allow us to do what we do best, predictions…

Always making predictions

Have you ever had the need to look around a car, look inside, look at the stirring wheel, look at the benches, look inside the trunk, look at the engine, just to know that it’s really a car? Probably not, only if you are trying to figure out if it is in good condition…

The reality is that you can look at a car positioned behind a horse and still know it’s a car. Why? Because your mind automatically fills the missing elements in the patterns that it observes.

And even better, just by looking at a car from a certain perspective, you can imagine it’s stirring wheel and benches. You can imagine how they look like and where they are positioned inside the car. You can imagine the engine and how all it’s components are arranged (if you are a mechanic).

The amazing thing is that everything works kind of like this. You observe, study it, learn it and then you make ‘predictions’. You can imagine how things look and how they are arranged just by looking at part of that pattern. If you learned it, you can figure out how any system is working just by looking at some of it’s parts.

And even more amazing is that most of the time your mind is making automatic predictions. If it’s a complex thing, like an engine, you need some time to think about all it’s parts. But if it’s a word for example, you autom_tically fill in the g_ps.

This is the principle used by magicians to perform spectacular tricks. Illusions are only possible because your mind is constantly making predictions, filling in the gaps. When it has enough information, it will just assume that it has all the information.

The brightness of a light bulb in our network is the certainty that we are observing the pattern that the bulb represents. For example, if we see a wheel, the bulb that represents a car will be medium bright, if we also see a part of the door, the brightness will increase. If we see the full car, the bulb will be very bright. This happens because the bulb that represents a wheel is bright, the bulb that represents a car door is bright. These bulbs are both connected to the bulb that represents a car, so when they get bright, the bulb that represents a car will get bright as well.

Let’s think of a single bulb that represents some pattern. Every bulb that is connected to this bulb composes the pattern that this bulb represents. The bulb is really just there to light up. What really defines the pattern are all the bulbs connected to it. If some of the bulbs connected to it light up, then this bulb will light up and the rest of bulbs that were not bright will become bright. And this is how you make the prediction, you fill the missing pieces by lightening the light bulbs that weren’t bright. And you only do this when you are certain that you are observing the pattern.

For example, you see a wheel, a car door and a window arranged in a certain way. Then all the other elements of the car, like the other doors, windows, etc… will light up in your mind and form the full shape of the car.

You can see that if the network is properly organised you can observe anything from any perspective and the missing pieces will automatically come to mind. Like magic!

But how can this big network organize itself so well? With a key mechanism that reshaped everything, learning…

Hebb’s Rule

Yes, learning. Everything we know, we learned. Even memories were learned, just think of memories as very specific patterns that only happened once. But how does the network learn them? Donald Hebb, a very influential Canadian psychologist, had a theory:

“neurons that fire together, wire together”, The Organisation of Behaviour, 1949

Think of the firing of a neuron as lighting a bulb. When two bulbs, close to each other, get bright at the same time, a wire will be created that connect each other. If a wire is already there, it’s weight will increase, according to the brightness of the bulbs. But the weight of the wires will decrease in time so if the bulbs rarely light up together then the weight will decrease and the connection might even disappear. This is why some memories fade over time. If they are very strong and you think of them from time to time, the wires will keep its weight (when you think of them, you are lighting the bulbs). This is essential, otherwise you wouldn’t form reasonable patterns. The point of the pattern is that it repeats itself over time and space.

As you can realize, this makes associations between patterns. It’s why you have an associative memory. It’s why you can make such good predictions. Everything you learn are associations that let you make predictions. Although this are the basic mechanisms, our complexity came for millions of years of evolution…

The rest, they say, is history…

With Hebb’s rule and by observing the world everyday, the network gets very complex, very quickly. Think of how many things you have seen and how many patterns your mind was able to capture. Loads of patterns are identified in your everyday life, and used to classify, discriminate, decide and act. Most of the times you use multiple patterns to make decisions.

Some are very hard to identify, it can take years, even a life time to find patterns that explain certain systems in the world. This is where language and education comes in. Older generations, with more knowledge of the world, shared their knowledge with younger generations. These then had time to obtain more information and mixture it with the one they already had, creating new, better, more accurate knowledge. Thousands of years following this process, gave rise to our modern civilisation.