Engineers from the University of Massachusetts Amherst have developed an artificial neuron whose electrical activity closely mimics that of real brain cells. This innovation builds on the team's previous research, where they used protein nanowires derived from bacteria capable of generating electricity. The new approach paves the way for computers that operate with efficiency comparable to living systems, as well as for direct interaction between electronic devices and biological tissues.

Brain Efficiency and Technological Challenges

According to Shuai Fu, a graduate student in the Department of Electrical and Computer Engineering at the University of Massachusetts and lead author of the study published in Nature Communications, the human brain can process vast amounts of information while consuming extremely little energy. For comparison, modern large language models like ChatGPT require significantly more electricity to perform similar tasks. The human brain uses only about 20 watts of energy to write a story, whereas artificial intelligence may consume over a megawatt for the same task. This is because the human body and brain operate with remarkable electrical efficiency—over 100 times greater than traditional computer circuits.

The brain contains billions of neurons—specialized cells that transmit and receive electrical signals throughout the body. Engineers have long sought to create artificial neurons for more energy-efficient computing, but reducing their operating voltage to biological levels has remained a major challenge. As Associate Professor Jun Yao notes, previous versions of artificial neurons operated at voltages ten times higher and consumed 100 times more energy than the new development. This made them less efficient and prevented direct interaction with living neurons, which are sensitive to strong electrical signals. The new artificial neuron operates at just 0.1 volts—about the same as neurons in the human body.

Potential Applications and Advantages

The work of Fu and Yao opens up a wide range of possibilities—from creating biology-inspired computers with high efficiency to electronic devices capable of directly interfacing with the human body. Current wearable electronic systems for monitoring are often bulky and inefficient: every time they detect a signal from the body, it must be amplified for computer analysis, increasing energy consumption and complicating the circuitry. Sensors built on the new low-voltage neurons can bypass the amplification stage, making them simpler and more economical.

The Role of Protein Nanowires

A key element of the new energy-efficient neuron is a protein nanowire synthesized from unique Geobacter sulfurreducens bacteria, which can also generate electricity. Yao and his colleagues have already used such nanowires to create several efficient devices: a biofilm that runs on sweat and can power personal electronics; an "electronic nose" for disease detection; and a device capable of extracting electricity directly from the air.

Research Support

This work was supported by the Army Research Office, the U.S. National Science Foundation, the National Institutes of Health, and the Alfred P. Sloan Foundation.