Due to physical limitations with conventional silicon-based technology, it is predicted that by 2040 our global energy supply will no longer be able to meet our computational demands. Hence there is a pressing need to develop new systems capable of processing data quicker and most importantly, in a more energy-efficient manner.

Scientists have long known that one way to achieve this would be to create a technology that mimics the human brain, the most powerful machine we know of. In silicon computing, different functions are performed by different physical entities which slows processing times and results in vast thermal waste. In contrast, neurons within our brain are able to send and receive information simultaneously throughout a huge network, at 10 times lower voltage than our most advanced computers.

The main advantage that our brain has over its silicon counterparts is the ability of parallel processing. Every one of our neurons is connected to thousands of others, and they can all act as both an input and an output of information. In silicon computing, transistor elements are only capable of three connections.

To create an artificial brain-like machine that could store and process information like we do, the key is finding physical materials that can seamlessly fluctuate between insulating and conducting an electrical current, as our neurons do.

A study published last month in the Cell Press journal Matter has discovered a material with such properties. Researchers at Texas A&M University created nanowires of β’-CuₓV₂O₅ which demonstrated the ability to oscillate between conductive states, in response to changes in temperature, voltage, and current.

On further inspection, they found that this ability stems from the movement of copper ions throughout β’-CuₓV₂O₅, which forces electrons to shift and transforms the conductive properties of the material. By manipulating this aspect of β’-CuₓV₂O₅, an electrical spike is generated on command, very similar to that produced when our neurons send signals between themselves.

A small electrical event such as this can be used to store information inside a circuit of the material. Our brain functions through firing specific neurons at a crucial time, in a unique sequence. One particular sequence of neuronal events results in the processing of information, whether it be recalling a memory or performing a physical action. A circuit of β’-CuₓV₂O₅ would act in the same way.

This is a huge breakthrough in the grand scheme of creating a circuitry network that could mimic how our brain works. This area of study is known as neuromorphic engineering and essentially aims to develop technologies that can replicate the power and architecture of our brain, matching its levels of processing efficiency and capabilities.

“The importance of this work is to show that chemists can rationally design and create electrically active materials with significantly improved neuromorphic properties. As we understand more, our materials will improve significantly, thus providing a new path to the continual technological advancement of our computing abilities.” R. Stanley Williams, Texas A&M electrical and computer engineer

The team is hopeful that this will be the first step towards creating such machines that could think and understand like us. With machine learning and neural network technology advancing each day, it will be fascinating to see how this technology could be combined with deep learning algorithms to store and process information more efficiently than we are currently able to with standard computers.

This article was originally published by Ellie harris on medium.