Unique Artificial Neurons Trigger Neural Activity in Living Cells

Why It Matters

The ability to generate biologically realistic spikes with low‑power, flexible printed hardware opens a viable path toward energy‑efficient AI and clinically viable neuroprosthetics, accelerating the shift from simulated to functional brain‑computer interfaces.

Key Takeaways

• Printable artificial neurons trigger real spikes in mouse brain tissue
• Molybdenum disulfide and graphene inks enable flexible, low‑cost fabrication
• Partial polymer residue creates conductive filament mimicking biological spike patterns
• Device operates within neuronal timing window, overcoming speed mismatches
• Additive printing reduces waste, supporting neuroprosthetic and BCI applications

Pulse Analysis

Neuromorphic computing promises to close the energy chasm between today’s data‑center AI models and the brain’s five‑order‑of‑magnitude efficiency advantage. Conventional silicon‑based accelerators consume gigawatts, prompting researchers to look for hardware that mimics neuronal mechanisms rather than merely emulating them. Northwestern’s printable neurons represent a tangible step toward that goal, delivering biologically faithful electrical spikes without the power overhead of traditional digital processors. By leveraging materials that naturally operate at millivolt scales, the approach aligns hardware consumption with the brain’s low‑energy signaling paradigm.

The core innovation lies in the material chemistry and the intentional retention of a polymer residue during aerosol‑jet 3D printing. Nanoscale flakes of molybdenum disulfide and graphene are deposited onto a flexible polymer substrate, where the residual polymer forms a narrow conductive filament. This filament produces abrupt, varied spikes that replicate the amplitude and temporal dynamics of natural neurons, including single spikes, sustained firing, and burst patterns. Unlike earlier artificial neurons that emit uniform pulses, these devices encode richer information per unit, reducing the component count needed for complex computations.

Beyond academic intrigue, the technology has clear translational potential. Its additive, waste‑minimizing manufacturing process lowers production costs and enables conformal, soft interfaces ideal for neuroprosthetic implants and brain‑machine interfaces. The demonstrated compatibility with mouse cerebellum slices suggests a pathway to clinical trials, where precise timing and biocompatibility are paramount. As other institutions, such as Carnegie Mellon, adopt the same aerosol‑jet printing for high‑density microelectrode arrays, the ecosystem for scalable, low‑power neural interfaces is rapidly coalescing, positioning printable artificial neurons as a cornerstone of next‑generation neurotechnology.