Printed Neurons Communicate with Living Brain Cells

Why It Matters

By demonstrating real‑time communication between printed electronics and biological neurons, the research paves the way for scalable, low‑power neural interfaces that could transform AI hardware and medical implants. It also offers a greener alternative to power‑hungry data‑center processors.

Key Takeaways

• Printed neurons use MoS₂ and graphene inks via aerosol jet printing.
• Devices generate neuron-like spikes that trigger activity in mouse brain slices.
• Flexible, low‑cost fabrication could accelerate brain‑machine interface development.
• Signal timing matches biological neurons, enabling direct neuro‑electronic communication.
• Energy‑efficient hardware may reduce AI data‑center power and water use.

Pulse Analysis

The quest for hardware that mimics the brain’s efficiency has long driven neuromorphic research, yet most artificial neurons fall short of biological realism. Northwestern’s team sidesteps this gap by leveraging printable electronic inks—molybdenum disulfide as a semiconductor and graphene as a conductor—on a soft polymer substrate. By partially decomposing the polymer during operation, they create a conductive filament that produces rapid, localized spikes, replicating the diverse firing patterns of real neurons. This method contrasts with earlier rigid silicon or bulk‑oxide approaches that either lack signal fidelity or demand excessive power.

Aerosol‑jet printing, an additive manufacturing technique, enables precise placement of nanomaterial inks while minimizing waste. The resulting devices are not only flexible but also inexpensive to produce at scale, a crucial factor for widespread adoption in medical and consumer applications. In mouse cerebellum slices, the printed neurons elicited voltage responses identical in timing and shape to natural spikes, confirming true biocompatibility. Such fidelity suggests these printed interfaces could soon bridge the gap between electronic processors and living neural circuits, supporting next‑generation brain‑machine interfaces for hearing, vision, and motor restoration.

Beyond therapeutic uses, the technology promises a paradigm shift for AI hardware. Current data‑center models consume gigawatts of electricity and strain water resources for cooling. Neuromorphic systems built from energy‑sparing printed neurons could execute complex computations with orders‑of‑magnitude lower power, echoing the brain’s five‑order efficiency advantage. As AI workloads continue to expand, integrating such low‑energy, scalable components may become essential for sustainable growth, prompting both industry and academia to explore printable neuromorphic architectures further.