A team at Northwestern University has developed printable artificial neurons capable of triggering real neural activity in living tissue, according to a study published in Nature Nanotechnology.
The finding matters because lab-built hardware isn’t just simulating brain signaling but producing responses indistinguishable enough from biological signals that actual neurons react to them.
Led by professor Mark C. Hersam and research associate professor Vinod K. Sangwan at Northwestern’s McCormick School of Engineering, the research used electronic inks made from nanoscale flakes of molybdenum disulfide and graphene, deposited onto flexible polymer substrates through a method called aerosol jet 3D printing.
What set this apart from previous artificial neuron designs was a manufacturing quirk the team chose to exploit rather than eliminate. Earlier researchers had treated a residual polymer in the ink as contamination and burned it off. Hersam’s team only partially removed it, which caused current passing through the device to form a narrow conductive filament. That filament produced the kind of sudden, varied electrical spikes that real neurons generate.
Most existing artificial neurons produce simplified, uniform pulses, which forces engineers to compensate by building larger, more power-hungry systems. This device generates single spikes, sustained firing, and burst patterns, covering more of the signaling range that biological neurons actually use, and because each neuron can encode more information on its own, the total number of components needed to handle the same computational load drops significantly.
Proving Compatibility in Living Tissue
To test biological compatibility, the team worked with neuroscientist Indira Raman, whose lab applied signals from the artificial neurons to slices of mouse cerebellum. The artificial spikes matched the timing and shape of natural neuron signals closely enough to reliably activate neural circuits in the tissue.
The timing matters because it has been the persistent failure point of earlier attempts. Organic material-based artificial neurons have spiked too slowly; metal oxide versions too fast. This device operates within the temporal window that living neurons actually use.
The broader context for this work is a growing power problem in computing. Training large AI models now requires data centers drawing gigawatts of electricity, with some facilities planning dedicated nuclear power plants. The brain operates at a level of energy efficiency that conventional digital computers fall short of by five orders of magnitude.
Neuromorphic hardware, which takes structural and functional cues from biological neural architecture, has long been proposed as one path toward closing that gap. Most of it, however, has stayed closer to metaphor than mechanism. Getting artificial signals to actually trigger responses in biological tissue is a harder and more specific claim than most prior work in the field has supported.
Since the manufacturing process is additive, and deposits material only where needed, it reduces both waste and cost relative to conventional semiconductor fabrication. The team says this makes the approach more viable for medical applications including neuroprosthetics and brain-machine interfaces, where device flexibility and biological tolerance are requirements, not optional features.
The fact that the same printing method has drawn independent institutional interest points to something beyond this study. Carnegie Mellon University (CMU) researchers there used aerosol jet 3D printing to fabricate high-density microelectrode arrays for brain-machine interfaces, a project backed by a $1.95 million National Institutes of Health (NIH) grant under the BRAIN Initiative. That work was focused on recording neural signals with greater precision and less tissue damage.
Northwestern’s device works in the opposite direction, generating signals that biological tissue responds to. This shared reliance on aerosol jet 3D printing across separate, independently funded projects demonstrates that the method is scaling past isolated lab experiments into a reproducible manufacturing standard for neural interfaces.
3D Printing Industry is inviting speakers for its 2026 Additive Manufacturing Applications (AMA) series, covering Energy, Healthcare, Automotive and Mobility, Aerospace, Space and Defense, and Software. Each online event focuses on real production deployments, qualification, and supply chain integration. Practitioners interested in contributing can complete the call for speakers form here.
To stay up to date with the latest 3D printing news, don’t forget to subscribe to the 3D Printing Industry newsletter or follow us on LinkedIn.
Explore the full Future of 3D Printing and Executive Survey series from 3D Printing Industry, featuring perspectives from CEOs, engineers, and industry leaders on the industrialization of additive manufacturing, 3D printing industry trends 2026, qualification, supply chains, and additive manufacturing industry analysis.
Featured image shows to move closer to a biological model, Mark 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. Photo via Mark Hersam | Northwestern Univeristy.
