How can three pounds of jelly inside our skulls enable us to do everything that makes us human? For centuries, scientists have been fascinated and puzzled by the mysterious workings of the brain. Now, for the first time, they can re-create in the computer the shapes of every one of the billions of nerve cells that make up our brains, the component parts of the intricate neural circuits that allow us to move, see and hear, to feel and to think. Armed with this new tool, scientists are beginning to decipher the secrets of the brain’s architecture, which may one day enable us to build smart technologies that surpass the capabilities of anything we have today.
This video is based on One Rule to Grow Them All: A General Theory of Neuronal Branching and Its Practical Application, a paper published by neuroscientist Hermann Cuntz and colleagues in the online journal PLoS Computational Biology.
Following is a statement by Dr. Hermann Cuntz:
The brain is composed of billions of cells called neurons. One neuron receives inputs from thousands of other neurons and sends out its signals to thousands more. We believe that if we understood the precise pattern with which neurons connect to each other, i.e. which neuron is connected with which other, we would understand how the brain works and how thoughts come about within the brain’s circuitry.
More than a hundred years ago, Ramón y Cajal’s big discovery was that you could read out the connection patterns between neurons by simply looking at their shapes: the tree-like input and output structures of neurons are not only beautiful but they are also optimized to connect to other neurons in an efficient manner.
Inspired by Cajal’s observations, we have designed a computer program that generates artificial shapes of neurons once the connection pattern is known. If we therefore are able to generate shapes that are indistinguishable from their real counterparts, we can conclude that we entered the correct connection patterns in order to generate the cells. We have then understood the connection pattern that led to these particular neuron shapes in the real brain.
Following is a statement by Dr. Eugene Izhikevich:
At Brain Corporation we’re working on the next generation of smart consumer products that will have what we call an artificial nervous system. The idea is that by creating mathematical models of the way the brain works, we can help build products that behave more like animals and less like robots.
These days we’re concentrating on creating large-scale models of the visual system that faithfully reproduce the biology of vision in animals. Some of these models depend on highly sophisticated algorithms that require enormous amounts of computer processing to generate millions of artificial neurons that have the same structures as real neurons.
To implement such large-scale models in a low-power mobile device, we’ve been collaborating with Qualcomm Incorporated (the largest mobile semiconductor company on the planet) to design new chips with massive computing power that will be the brains of our artificial nervous systems.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Use the Form Below to comment on this post.