Summary: Engineers have achieved a major breakthrough in bio-electronics: printing artificial neurons that can communicate directly with the biological brain. The study details the development of flexible, low-cost devices that generate electrical signals so realistic they successfully triggered responses in mouse brain tissue.
Unlike traditional silicon chips, these “printable” neurons mimic the brain’s energy efficiency and signaling complexity, paving the way for advanced neuroprosthetics and AI hardware that consumes a fraction of the power used by modern data centers.
Key Findings
- Biocompatibility in Action: In tests with the lab of Indira M. Raman, the artificial neurons were hooked up to slices of a mouse cerebellum. The living neurons responded to the artificial spikes as if they were coming from a biological peer.
- Energy Efficiency: The brain is five orders of magnitude more energy-efficient than a digital computer. By mimicking its structure, these devices could handle “Big Data” without the massive power and water cooling requirements of modern AI data centers.
- Structural Mimicry: Unlike silicon chips with billions of identical, fixed transistors, these artificial neurons are heterogeneous and dynamic, mimicking the soft, 3D changing networks of the human brain.
- Additive Manufacturing: The printing process is low-cost and environmentally friendly, only placing material where it is needed, which significantly reduces waste.
Source: Northwestern University
Northwestern University engineers printed artificial neurons that don’t just imitate the brain, they talk to it.
In a new study, the Northwestern team developed flexible, low-cost devices that generate electrical signals realistic enough to activate living brain cells. When tested on slices of tissue from mouse brains, the artificial neurons successfully triggered responses from real neurons, demonstrating a new level of biocompatibility.
The work marks a step toward electronics that can communicate directly with the nervous system, with potential applications in brain-machine interfaces and neuroprosthetics, including implants for hearing, vision and movement.
It also lays the groundwork for more efficient, brain-like computing systems. By mimicking how neurons signal, a key feature of the brain, which is the most energy-efficient computer known, futuristic systems could perform complex operations using far less power than today’s data-hungry technologies.
The study will be published on Wednesday (April 15) in the journal Nature Nanotechnology.
“The world we live in today is dominated by artificial intelligence (AI),” said Northwestern’s Mark C. Hersam, who led the study.
“The way you make AI smarter is by training it on more and more data. This data-intensive training leads to a massive power-consumption problem. Therefore, we have to come up with more efficient hardware to handle big data and AI. Because the brain is five orders of magnitude more energy efficient than a digital computer, it makes sense to look to the brain for inspiration for next-generation computing.”
An expert in brain-like computing, Hersam is the Walter P. Murphy Professor of Materials Science and Engineering at Northwestern’s McCormick School of Engineering, professor of medicine at Northwestern University Feinberg School of Medicine and professor of chemistry at Northwestern’s Weinberg College of Arts and Sciences. He also is the chair of the department of materials science and engineering, director of the Materials Research Science and Engineering Center and member of the International Institute for Nanotechnology. Hersam co-led the study with Vinod K. Sangwan, a research associate professor at McCormick.
From rigid silicon to dynamic brains
As computing tasks become more complex and data-intensive, computers meet these demands by adding more identical components — billions of transistors packed onto rigid, two-dimensional silicon chips. Each transistor behaves the same way. And, once fabricated, those systems remain fixed.
The brain operates in a strikingly different way. Rather than comprising uniform building blocks, the brain relies on diverse types of neurons — each performing specialized roles — organized across regions. These soft, three-dimensional networks constantly change, forming and reshaping connections over time as people learn and adapt.
“Silicon achieves complexity by having billions of identical devices,” Hersam said. “Everything is the same, rigid and fixed once it’s fabricated. The brain is the opposite. It’s heterogeneous, dynamic and three-dimensional. To move in that direction, we need new materials and new ways to build electronics.”
While other artificial neurons do exist, they fall short of biological realism. Most produce simplified signals, forcing engineers to rely on large, energy-intensive networks of devices to achieve complex behavior.
Turning an imperfection into a feature
To move closer to a biological model, Hersam’s team developed artificial neurons using soft, printable materials that better mimic the brain’s structure and behavior. The backbone of that advance is a series of electronic inks, formulated from nanoscale flakes of molybdenum disulfide (MoS2), which acts as a semiconductor, and graphene, which serves an electrical conductor. Using a specialized printing technique called aerosol jet printing, the researchers deposited these inks onto flexible polymer substrates.
In the past, other researchers viewed the stabilizing polymer in the inks as a problem that interfered with electrical current flow, so they burned the polymer away after printing the electronic circuit. But Hersam leveraged this minor imperfection to add brain-like functionality to his device.
“Instead of fully removing the polymer, we partially decompose it,” he said. “Then, when we pass current through the device, we drive further decomposition of the polymer. This decomposition occurs in a spatially inhomogeneous manner, leading to formation of a conductive filament, such that all the current is constricted into a narrow region in space.”
This narrow region becomes a localized pathway that produces a sudden, neuron-like electrical response. The result is a new type of artificial neuron that can generate a rich range of electrical signals.
Instead of generating simple, one-off pulses, the new device produces more complex signaling patterns — including single spikes, continuous firing and bursting patterns — that resemble how real neurons communicate.
By capturing this signaling diversity, each neuron can encode more information and perform more sophisticated functions. And that can reduce the number of components needed in a computing system, drastically improving overall efficiency.
Putting artificial neurons to the test
To test whether its artificial neurons truly could interface with biology, Hersam’s team collaborated with Indira M. Raman, the Bill and Gayle Cook Professor of Neurobiology at Weinberg. Raman’s team applied electrical signals from the artificial neurons to slices of mouse cerebellum.
They found the artificial voltage spikes matched key biological features, including timing and duration of living neuron voltage spikes. This reliably triggered activity in real neurons, activating neural circuits in a way similar to natural signals.
“Other labs have tried to make artificial neurons with organic materials, and they spiked too slowly,” Hersam said.
“Or they used metal oxides, which are too fast. We are within a temporal range that was not previously demonstrated for artificial neurons. You can see the living neurons respond to our artificial neuron. So, we’ve demonstrated signals that are not only the right timescale but also the right spike shape to interact directly with living neurons.”
The approach comes with several environmentally friendly advantages. In addition to improving energy efficiency, the neuron’s manufacturing process is simple and low-cost. Because the printing process is additive — placing material only where it’s needed — it also reduces waste.
“To meet the energy demands of AI, tech companies are building gigawatt data centers powered by dedicated nuclear power plants,” Hersam said.
“It is evident that this massive power consumption will limit further scaling of computing since it’s hard to imagine a next-generation data center requiring 100 nuclear power plants.
“The other issue is that when you’re dissipating gigawatts of power, there’s a lot of heat. Because data centers are cooled with water, AI is putting severe stress on the water supply. However you look at it, we need to come up with more energy-efficient hardware for AI.”
Funding: The study, “Multi-order complexity spiking neurons enabled by printed MoS2 memristive nanosheet networks,” was supported by the National Science Foundation.
Key Questions Answered:
A: Not quite, but we can print the “translators.” These artificial neurons act as a bridge. Because they speak the same electrical language as your body, they could be used to create better cochlear implants for hearing, visual prosthetics for the blind, or interfaces that allow paralyzed patients to move robotic limbs with their thoughts.
A: Modern AI is a “power hog.” Tech companies are currently looking into building dedicated nuclear power plants just to keep up with the data centers needed for tools like ChatGPT. By making hardware that works like a brain (which runs on about the same power as a dim lightbulb), we can make AI smarter without draining the world’s energy and water.
A: Your computer’s “neural networks” are just software, math code running on rigid metal. These are physical neural networks. They are soft, flexible, and have “memory” built into their physical structure, making them much closer to the “wetware” inside your skull.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by our staff.
About this neurotech research news
Author: Amanda Morris
Source: Northwestern University
Contact: Amanda Morris – Northwestern University
Image: The image is credited to Neuroscience News
Original Research: Closed access.
“Printed MoS2 memristive nanosheet networks for spiking neurons with multi-order complexity” by Shreyash S. Hadke, Carol N. Klingler, Spencer T. Brown, Meghana Holla, Xudong Zhuang, Linda Li, M. Iqbal Bakti Utama, Santiago Diaz-Arauzo, Anurag Chapagain, Siyang Li, Jung Hun Lee, Indira M. Raman, Vinod K. Sangwan & Mark C. Hersam. Nature Nanotechnology
DOI:10.1038/s41565-026-02149-6
Abstract
Printed MoS2 memristive nanosheet networks for spiking neurons with multi-order complexity
Artificial neurons that reproduce the rich dynamical behaviour of biological spiking are essential for neuromorphic hardware and biohybrid interfaces, yet scalable solution-processed devices with physiologically relevant spiking characteristics remain elusive.
Here we demonstrate aerosol-jet-printed memristive networks of MoS2 nanosheets that exhibit thermally activated filamentary switching and snap-back negative differential resistance, enabling volatile threshold switching in fully printed graphene/MoS2/graphene devices on flexible substrates.
In situ thermal imaging and circuit modelling reveal that current-constricted filaments formed through Joule heating govern the nonlinear switching dynamics. These printed memristors enable oscillatory and spiking neuron circuits with tunable frequencies up to 20 kHz and stable operation over more than 106 cycles.
Simple neuristor circuits realize first-, second- and third-order spiking complexity, including integrate-and-fire behaviour, spike latency, tonic firing, class 1 excitability, tonic bursting and phasic dynamics.
The generated spike waveforms match physiological timescales and stimulate Purkinje neurons in mouse cerebellar slices.
Our results establish printed nanosheet memristive networks as a scalable platform for bio-realistic neuromorphic hardware and flexible brain–machine interfaces.
