you're reading...


How does the brain work?

In the movie Avatar, Dr. Grace Augustine says to Selfridge that the roots of the trees on Pandora have “more connections than the human brain”. Implied is that having connections leads to intelligence. Indeed, you often hear claims that childhood experience is important because it leads to new connections in the human brain. Commentators will sometimes even argue that more connections provide more pathways for information to proceed through the brain, leading to greater intelligence. But does this really make sense?

In my book Intelligence and the Brain, I point out the flaws with thinking about the brain in this way. In Avatar, Augustine goes on to say that the tree network functions like a data network, similar to our own internet. In data networks, each connection represents a switch that may be changed by the program or data currently running. In this way, data may be arbitrarily routed through the network. Typing a website address into your browser can result in information from that website’s server being transmitted to your computer. Similarly, in your own computer, data may be stored at any location in computer memory. Software then consists of instructions that tell switches or transistors to change, resulting in the data being retrieved. Further instructions then identify which switches to change to implement standard computing functions. The results are then stored in new memory locations, such as a screen buffer that displays the result on the computer’s monitor. All computing may be reduced down to this use of switches.

However, recent advances in the brain sciences over the last few decades have revealed that the brain is not like a data network. Neurons are not like transistors. They do not function like switches. They cannot take an input and then channel it to one of their outputs and not others based on some instruction. Instead, brain science has found that a neuron functions as a gating mechanism. If its input exceeds a certain threshold, a signal is produced that is carried to all of a neuron’s output connections. This tells us that simply increasing the number of connections in the brain would not lead to increased intelligence. It would just eventually lead to the situation where any input to the nervous system would have the same outcome–all of the neurons firing simultaneously.

The Role of Abstraction

So if the brain does not work like a data network, how does it work? Psychologists have known for a long time that abstraction is central to human intelligence. Humans are intelligent because we are good at abstracting out information from specific concrete experiences. Indeed, IQ tests get more difficult by employing problems that rely on more and more difficult abstractions to solve them. So if we want to understand how the brain works, we really need to look at abstraction.

We find that the ability to understand abstractions increases over childhood. Unfortunately, some people do not appreciate this because the concept of IQ can be confusing to those who are not familiar with how IQ test scores are calculated. They will hear that IQ stays constant over childhood. This is then taken to imply that intelligence does not change over this time. However, this is not true.

IQ is a measure of performance relative to people of the same age–so it says nothing about how performance changes with age. Indeed, originally IQ was determined by the formula of Mental Age / Chronological Age * 100. Using this formula, since Chronological Age increases, IQ can only stay constant if Mental Age increases as well. Supporting this, it is readily observed that a 16-year-old with an IQ of 130 is able to solve much more difficult IQ problems than a younger child with the same IQ.

Pruning of the Neural Connections

How do the connections of the brain then develop? Brain science has found that the child’s brain will initially develop an abundance of neural connections. It then goes through a pruning process, gradually reducing the number of connections over childhood until adulthood is reached. This pruning of the connections over childhood corresponds with the increasing ability to understand abstractions.

This role of pruning should not be surprising. The challenge in perceiving an abstraction is to filter out all of the information that is present in a concrete instance of an abstraction that is irrelevant to the commonality across situations. By pruning the connections, only relevant information is retained, enabling an abstraction to be perceived.

How Memory Works

This role of pruning becomes more clear when we look at how memory works in the brain. Memory typically refers to our ability to memorize facts and figures. For instance, learning that the capital of France is Paris. Unlike abstractions that take months or even years to build up, we can memorize facts after only a single experience.

Brain science has found that memory is due to an auto-associative network in the brain. Auto-associative networks have the characteristic that they can be presented with a pattern, and they learn to respond with the identical pattern. This does not initially seem to be that useful. However, once an auto-associative has been presented with a pattern, it can then recall that same pattern even if it is only presented with a fragment of the original pattern later on.

This may still not sound that interesting. However, think of how your own memory works. If you are presented with the information capital-France-Paris, your auto-associative network can store this as a pattern or configuration. If you are then presented with capital and France, your auto-associative network will automatically complete the pattern and return the answer–Paris.

Many other examples of memory fit this process of pattern completion. On your birthday, you ate pizza and cake. If you then think of your birthday, your auto-associative network will recall the pizza and cake. If you are shown the steps to send an email, your auto-associative network will store this as a pattern. When you next think of sending an email, your auto-associative network will recall the exact steps involved. In short, we use pattern completion every day to recall previously learnt information.

Limitations of Memory

Given that pattern completion is so fundamental to behavior, this leads to the question of whether human intelligence consists of only memory. It does not, for the reason implied above. While pattern completion can be very useful, it is, by itself, also very limited.

For instance, imagine if your brain had an auto-associative network that was plugged in to only your sensory receptors. You might then be told that the capital of France is Paris. If someone else later asks you what the capital of France is, your auto-associative network would make no response. The reason is that the verbal question would activate completely different sensory inputs to what you read, so there would be no way that the auto-associative network could recognize the pattern and complete it.

This is where abstraction comes in. By abstracting out information, the same representation can be activated across different concrete instances. In this way, information can generalize across situations, leading to successful performance in what are otherwise novel situations.

Moving On

In this brief article, I have tried to briefly summarize what is now known about how the brain works. More detailed explanations and additional information is available in my book Intelligence and the Brain, including the environmental and genetic factors that affect IQ, further differences between the human brain and computers, what characterizes savant syndrome, and factors that influence the development of genius.

About admin

has written 135 post in this blog.