Engineers at Northwestern University reported on April 15, 2026, in the journal Nature Nanotechnology that they have printed artificial neurons capable of generating electrical signals realistic enough to activate real, living brain cells, a step the team says moves flexible electronics closer to genuine communication with the nervous system. The work was led by Mark C. Hersam, the Walter P. Murphy Professor of Materials Science and Engineering at Northwestern’s McCormick School of Engineering, alongside research associate professor Vinod K. Sangwan and neurobiology professor Indira M. Raman.
Printed, Not Etched: A Different Way to Build a Chip
Unlike conventional computer chips, which are built through rigid, expensive silicon fabrication processes in cleanrooms, the Northwestern team created its artificial neurons using electronic inks deposited through aerosol jet printing, a technique that sprays a fine mist of material onto flexible polymer substrates in precise patterns. The inks were formulated from nanoscale flakes of molybdenum disulfide, which behaves as a semiconductor, and graphene, which conducts electricity, with partial decomposition of the surrounding polymer creating the conductive filaments needed to mimic a working neuron’s electrical behavior.
Testing the Signals on Real Brain Tissue
To find out whether the printed neurons could do more than mimic brain signals on a bench-top oscilloscope, the researchers applied their artificial neurons’ output directly to slices of mouse cerebellum tissue kept alive in the lab. The printed neurons successfully triggered responses in the living cells, with the timing and duration of the artificial voltage spikes closely matching what real neurons produce, according to the study, effectively proving the printed devices could speak the brain’s own electrical language rather than just approximating it.
Why Efficiency Is the Real Prize
Hersam has framed the motivation behind the project around energy efficiency rather than raw computing speed, noting that the brain is roughly five orders of magnitude, or on the order of 100,000 times, more energy efficient than a digital computer at processing information, which is why his team looked to biological neurons for design inspiration. Today’s most advanced AI systems require enormous data centers consuming vast amounts of electricity, and brain-inspired, or neuromorphic, computing hardware is one of the more promising paths researchers have identified for cutting that energy cost, since biological neurons process information using far less power than transistor-based circuits performing equivalent tasks.
From Lab Bench to Possible Medical Devices
Beyond computing, the Northwestern team points to brain-machine interfaces and neuroprosthetics as a direct application of the technology, including devices that could one day help restore hearing, vision, or movement in patients with neurological damage or limb loss by directly and efficiently interfacing with the nervous system. Because the printing process is inexpensive and additive, meaning it deposits material only where needed rather than carving it away from a larger block as in traditional chip fabrication, the researchers argue it could make such devices cheaper to manufacture and more adaptable to the flexible, curved surfaces found throughout the human body, unlike rigid silicon components.
How Far This Is From a Real Implant
The experiment took place on isolated slices of tissue from a mouse brain, not in a living animal, let alone a human, and researchers in the broader neurotechnology field caution that stimulating cultured tissue in a dish is a far simpler test than achieving safe, stable, long-term integration inside a living, moving organism, where issues like immune response, signal drift, and mechanical wear on implanted materials become significant obstacles. Bringing this kind of printed neuromorphic material into an actual clinical neuroprosthetic would likely require years of additional engineering, biocompatibility testing, and regulatory review before any human trials could begin, and academic experts note that many promising brain-interface materials fail during that longer translation process.
What It Means and What’s Next
For now, the Northwestern results serve as an early proof of concept that a cheap, printable, flexible material can produce brain-like electrical signals precise enough to communicate with actual living neurons, rather than a finished product ready for patients. The research team’s stated next steps include refining the printing process for greater reliability and exploring how arrays of these artificial neurons might be scaled up into more complex circuits capable of processing, not just generating, brain-like signals. If that work succeeds, it could feed into two very different long-term goals: energy-efficient neuromorphic computer chips that reduce the power appetite of future AI systems, and next-generation neural implants designed to restore lost sensory or motor function by speaking the brain’s native electrical language.
