When Songrui Zhao calls his team's latest creation a "crazy idea," he's not being modest. The McGill University engineer and his colleagues have built a nanoscale device from inorganic materials that behaves exactly like a single biological neuron—and they're doing it in a lab in Montreal.
The device, a light-detecting structure engineered from layers of atoms, can receive signals, briefly store information, and fire a response once a threshold is reached. In other words, it processes data the same way a living brain cell does. The findings, published in the journal Nanoscale, mark the first time researchers have closely mimicked single neuron dynamics using purely physical materials.
"It's a bit of a crazy idea—to create something like a biological system using an inorganic material," Zhao said.
The breakthrough sidesteps a fundamental limitation of current technology: most devices capture information first, then ship it elsewhere for processing. Zhao's team built theirs to do both in the same place, much like the human eye handles visual data before it ever reaches the brain. By engineering atom-thin layers using a technique called molecular beam epitaxy and exposing the device to light of varying colors, intensities, and timing patterns, the researchers demonstrated that sophisticated neural behavior can emerge directly from the physics of a material—without software or complex circuitry.
"By carefully engineering the layers, we created a device with a tunable response to light, which forms the basis for emulating how a single neuron behaves," Zhao said. "We were able to design the flow of electrical current to produce the behavior we wanted."
That control opens up practical possibilities. Because artificial neural networks—the systems powering machine learning and modern AI—are built from interconnected neurons, Zhao envisions the device serving as a modular building block. "A single artificial neuron is like a cell you can use as a building block, allowing us to construct networks from the bottom up," he said.
The energy savings could be substantial. Devices that rely on separate processing stages typically demand more power; this approach cuts out the middle step, reducing what the researchers call the "energy demand associated with similar devices that rely on circuits or software."
For patients with degenerative eye conditions, the technology could eventually inform more responsive artificial retinas. For cybersecurity experts, Zhao's team is already exploring how processing information directly at the sensor could improve data encryption. Smart optical sensors, advanced computing, and more efficient forms of information processing all sit within reach, the researchers say.
Next, the team plans to expand the device's light-response range and refine its performance. It is early work, but the foundation is solid: a single cell, behaving as nature does, built atom by atom in a Montreal laboratory.