Oxford University Press's
Academic Insights for the Thinking World

Computational Theories and their Implementation in the Brain

How does the brain work?

The media is full of stories about how this or that area of the brain has been shown to be active when people are scanned while doing some task. The images are alluring and it is tempting to use them to support this or that just-so story. However, they are limited in that the majority of the studies simply tell us where in the brain things are happening.

But the aim of neuroscience is to discover how the brain works. Yet the spatial resolution of these images is only of the order of millimetres. This means that within any one area that is shown to be active there are many millions of nerve cells or neurons. It is an awesome challenge to find out how these are connected and how their combined activity carries out particular operations. And just think: there are an estimated 86 billion neurons in the human brain as a whole!

One way of trying to find answers is to appeal to the operations that are carried out by computers. In other words, we can appeal to ‘computational’ theories. One of the pioneers of what we now call ‘Computational Neuroscience’ was David Marr. In his seminal book on Vision (1982), he suggested that we address the problem at three levels: 

  1. The computational level – what is being computed?
  2. The algorithmic level – what transformations are being carried out?
  3. The level of implementation – how are these transformations implemented?

In a computer, the operations are carried out by silicon chips, but in the brain, by circuits of neurons. It is not so difficult to understand how silicon chips work because we have made them. The problem is that this is not the case for neuronal circuits.

…there are an estimated 86 billion neurons in the human brain as a whole.

So it is simply astonishing that when he was a graduate student in the late 1960s David Marr produced theories as to how three of the structures in the brain actually work. These were the cerebellum, involved in the automation of motor skills; the hippocampus, involved in retrieving memories from our life; and the neocortex, involved in categorising and classifying the sensory information that we receive from the world. At the time, the majority of scientists were addressing questions as to what each brain area did. Yet here was someone having the daring to ask what transformations were performed and how they were implemented in neuronal circuitry.

These theories turned out to have a very widespread influence in the scientific literature. And they were so far ahead of their time that it was not for ten years that the most basic prediction made in the cerebellar theory was tested and found to be correct. Since then, neuroscience has advanced rapidly, with the advent of many new methods. We can trace the anatomical connections of individual neurons or groups of neurons; we can record the activity of individual neurons or groups of neurons while animals or people are performing tasks; we can record the magnetic signals that are produced when neurons become active, and do so in people using detectors from outside the head; and we can even turn off the activity of particular neurons using light. So we are now in a position to ask how Marr’s early theories hold up in the light of the findings of today’s neuroscience.

Neuroscience is coming of age.

It is a tragedy that David Marr did not live to find out for himself. He died in his mid-thirties in 1980, though his wife, the neuroscientist Lucia Vaina, is still working. Had he lived he would have been 70 last year. Without question, there would have been a scientific gathering to congratulate him on his achievements.

As it is, there is no better way of celebrating one of the most influential scientists of the modern era than to use current knowledge to produce theories of how the cerebellum, hippocampus, and neocortex actually work. And there are scientists around the world who are now doing so. Neuroscience is coming of age.

Featured image credit: Computer by Lorenzo Cafaro. CC0 public domain via Unsplash.

Recent Comments

  1. Michael Taylor

    I try to sense where in my head my thoughts come from. When I read this article. Behind my 20/200 left eye, above my aching wisdom tooth; and as I type, there’s a shift to the center, still low in my head. As if I was looking out below my nose.
    Since my wife was diagnosed with Young Onset Alzheimer’s 7 years ago, any article about brain function attracts my attention, why can a stroke victims brain be retrained? But the Alzheimers forgetfulness, doesn’t seem to respond to a similar therapy.
    I often wonder where the level of concentration that scientists possess comes from. Why does my brain flit from this to What is wrong with US politics, the floods in Wales, the earthquakes in New Zealand, my 2 cats, or how different the coffee tastes today?
    Thank you to all of science for giving us insight to ourselves.

  2. Dick Passingham

    All good questions. The truth is that,whatever we say, we don’t really understand the brain. But there is a perfectly good reason for this. It is the most complicated thing that scientists have to study. Not only are there 86 billion neurons but on any one neuron there can be up to 26,000 connections. So when I said that neuroscience is coming of age all I meant is that we are trying. Have pity on us.

  3. kdn

    Studying the brain only gives us information on correlations (i.e., which function/mental phenomena is correlated with which part of the brain). This type of information has very limited applications. Also, before improved standards of antenatal care were available, scientists have studied hundreds of patients who displayed normal and above normal IQ’s in spite of having severely reduced brain tissue. [Reference: Lorber J. Is your brain really necessary? Nurs Mirror. 1981 Apr 30;152(18):29-30.]. We occasionally have such cases even today. Makes me wonder if a brain is even necessary?!

  4. Dick Passingham

    It is true that functional brain imaging studies correlations, but we have other methods that get at causes. We can interfere using transcranial magnetic brain stimulation, study the order of events using magnetoencephalography, stimulate using depth electrodes and so on. As for claims that people can have a thin cortex, yet normal IQ, it turns out that in such cases the number of neurons is not reduced; it is just that they are small. I promise you we really do need our brain!

Leave a Comment

Your email address will not be published. Required fields are marked *