Making computer chips act more like brain cells

The human brain is an amazing computing machine. Weighing only three pounds, it can process information a thousand times faster than the fastest supercomputer, store a thousand times more information than a powerful laptop, and do it all without using more energy than a 20-watt light bulb.

Researchers are trying to replicate this success using soft, flexible organic materials that can act like biological neurons, and may one day even connect to them. Eventually, soft “neuromorphic” computer chips could be implanted directly into the brain, allowing people to control an artificial arm or computer monitor simply by thinking about it.

Schematic of biological neural network and organic nanowire synaptics
a transistor that emulates a biological synapse.

Like real neurons – but unlike conventional computer chips – these new devices can send and receive both chemical and electrical signals. “Your brain works with chemicals, with neurotransmitters like dopamine and serotonin. Our materials are able to interact electrochemically with them,” says Alberto Salleo, a materials scientist at Stanford University who wrote about the potential for organic neuromorphic devices in 2021 Annual Review of Materials Research.

Salleo and other researchers have created electronic devices using these soft organic materials that can act as transistors (which amplify and switch electrical signals) and memory cells (which store information) and other basic electronic components.

The work grows out of a growing interest in neuromorphic computer circuits that mimic how human neural connections, or synapses, work. These circuits, whether made of silicon, metal or organic materials, work less like those in digital computers and more like the networks of neurons in the human brain.

Conventional digital computers work step by step, and their architecture creates a fundamental division between computation and memory. This division means that ones and zeros must be switched back and forth between locations on the computer’s processor, creating bottleneck for speed and energy consumption.

The brain does things differently. A single neuron receives signals from many other neurons, and all these signals together affect the electrical state of the receiving neuron. In effect, each neuron serves as both a computing device—integrating the value of all the signals it receives—and a memory device: it stores the value of all those combined signals as an infinitely variable analog value, rather than the zero-or-one of digital computers.

Researchers have developed a number of different “memristive” devices that mimic this ability. When you pass electric currents through them, you change the electrical resistance. Like biological neurons, these devices calculate by summing the values ​​of all the currents they have been exposed to. And they remember through the resulting value their resistance.

A simple organic memristor, for example, may have two layers of electrically conductive materials. When a voltage is applied, an electric current drives positively charged ions from one layer to another, changing how easily the other layer will conduct electricity the next time it is exposed to an electric current. (See diagram.) “It’s a way to let physics do the math,” he says Matthew Marinellacomputer engineer at Arizona State University in Tempe who researches neuromorphic computing.

The technique also frees the computer from strictly binary values. “When you have a classic computer memory, it’s either zero or one. We create a memory that can be any value between zero and one. So you can set it up in an analog way,” says Salleo.

Currently, most memristors and related devices are not based on organic materials, but use standard silicon chip technology. Some are even used commercially as a way to speed up artificial intelligence programs. But organic components have the potential to get the job done faster while using less energy, Salleo says. Even better, they could be designed to integrate with your brain. The materials are soft and flexible, and have electrochemical properties that allow them to interact with biological neurons.

For example, Francesca Santoro, an electrical engineer now at RWTH Aachen University in Germany, is developing polymer device that receives data from real cells and “learn” from it. In her device, the cells are separated from the artificial neuron by a small space, similar to the synapses that separate real neurons from each other. As cells produce dopamine, a chemical that signals nerves, dopamine changes the electrical state of the artificial half of the device. The more dopamine the cells produce, the more the electrical state of the artificial neuron changes, just as you can see with two biological neurons. (See diagram.) “Our ultimate goal is to really design electronics that look like neurons and behave like neurons,” says Santoro.

The approach could offer a better way to use brain activity for driving prosthetics or computer monitors. Today’s systems use standard electronics, including electrodes that can only capture broad patterns of electrical activity. And the equipment is bulky and requires external computers to operate.

Flexible, neuromorphic circuits could improve this in at least two ways. They would be able to translate neural signals in a much more detailed way, responding to the signals of individual neurons. And the devices could also handle some of the necessary calculations themselves, Salleo says, which could save energy and increase processing speed.

Decentralized low-level systems of this kind — with small, neuromorphic computers that process information as it is received by local sensors — are a promising avenue for neuromorphic computing, Salleo and Santoro say. “The fact that they so beautifully resemble the electrical workings of neurons makes them ideal for physical and electrical coupling with neural tissue,” says Santoro, “and ultimately the brain.”

This article originally appeared in Knowable Magazinean independent journalistic venture from Annual Reviews.