Artificial neurons successfully communicate with living brain cells

Engineers at Northwestern University have achieved a groundbreaking feat: the creation of printed artificial neurons that transcend mere imitation and can directly interface with living brain cells. These novel, flexible, and cost-effective devices are engineered to generate electrical signals that mirror the complex patterns of biological neurons, enabling them to actively stimulate and interact with neural tissue. This breakthrough, detailed in a forthcoming publication in the prestigious journal Nature Nanotechnology on April 15, represents a significant leap toward seamless integration of artificial electronics with the intricate biological systems of the nervous system.

A New Era of Neural Interfacing

In a series of pivotal experiments, researchers demonstrated that these printed artificial neurons successfully elicited responses within slices of mouse brain tissue. This remarkable achievement signifies an unprecedented level of compatibility between synthetic electronic components and living neural networks, opening doors to a wide array of transformative applications.

The implications of this development are profound, particularly for the advancement of brain-machine interfaces (BMIs). These technologies, which enable direct communication between the brain and external devices, hold immense promise for restoring lost sensory and motor functions. Potential applications include the development of sophisticated neuroprosthetics capable of reintroducing hearing to the deaf, sight to the blind, or restoring voluntary movement to individuals with paralysis. Such devices could revolutionize rehabilitation and offer new avenues for treating neurological disorders.

Beyond medical applications, this innovation also paves the way for a new generation of computing systems inspired by the brain’s own architecture. The brain, renowned for its extraordinary energy efficiency, processes information through a complex web of interconnected neurons. By emulating these communication principles, future hardware could perform sophisticated computational tasks with a fraction of the energy consumed by conventional digital computers. Scientists have long marveled at the brain’s ability to achieve remarkable cognitive feats while consuming an estimated five orders of magnitude less energy than a typical digital computer. This stark contrast underscores the potential for bio-inspired computing to address the escalating energy demands of modern artificial intelligence.

The Brain’s Advantage: Heterogeneity and Dynamics Over Silicon’s Rigidity

Traditional silicon-based computing relies on packing billions of identical transistors onto rigid, two-dimensional chips. While this approach has driven immense progress, each component operates identically, and once manufactured, the system’s architecture remains largely fixed. This contrasts sharply with the brain’s fundamentally different operating principles.

The human brain is a marvel of biological engineering, comprised of diverse types of neurons, each specialized for distinct functions. These neurons are organized into intricate, three-dimensional networks that are not static but dynamically reconfigurable. As learning occurs, connections between neurons are constantly strengthened, weakened, or formed anew, allowing for remarkable adaptability and plasticity.

"Silicon achieves complexity by having billions of identical devices," explained Mark C. Hersam, the Walter P. Murphy Professor of Materials Science and Engineering at Northwestern’s McCormick School of Engineering, who led the study. "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 artificial neurons have been developed previously, many have produced signals that are too simplistic for nuanced interaction with biological systems. To achieve more complex behaviors, researchers have often resorted to building extensive networks of these simpler artificial neurons, which in turn leads to increased energy consumption. The Northwestern team’s breakthrough lies in their ability to imbue each individual artificial neuron with a greater degree of complexity, thereby reducing the overall number of components required for advanced functions.

Printable Materials Unlock Brain-Like Neural Signaling

To more accurately replicate the intricate signaling patterns of real neurons, Hersam’s team turned to soft, printable materials that better mimic the brain’s organic structure. Their innovative approach utilizes electronic inks composed of nanoscale flakes of molybdenum disulfide (MoS₂), a semiconductor material, and graphene, an excellent electrical conductor. These advanced materials were precisely deposited onto flexible polymer substrates using aerosol jet printing, a versatile and scalable manufacturing technique.

A key element of their success involved re-evaluating a component previously considered a limitation. In prior research, the polymer used in these electronic inks was often treated as an impediment to electrical performance and subsequently removed after the printing process. However, the Northwestern engineers ingeniously harnessed this polymer’s properties to enhance device functionality.

"Instead of fully removing the polymer, we partially decompose it," Hersam elaborated. "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 precisely controlled decomposition process creates a narrow conductive pathway within the artificial neuron. When electrical current flows through this pathway, it triggers a sudden and robust electrical response, closely mimicking the "firing" of a biological neuron. Crucially, this mechanism allows the resulting device to generate a rich repertoire of signals, including single electrical spikes, sustained firing patterns, and dynamic bursting activity, all of which bear a striking resemblance to natural neural communication. The ability of each artificial neuron to produce such complex signals means that fewer components are needed to achieve sophisticated computational tasks, a significant factor in improving overall computing efficiency.

Rigorous Testing: Artificial Neurons Engage with Living Brain Tissue

To validate the practical efficacy of their artificial neurons in interacting with living systems, the researchers collaborated with Indira M. Raman, the Bill and Gayle Cook Professor of Neurobiology at Northwestern’s Weinberg College of Arts and Sciences. Dr. Raman’s team conducted experiments applying the electrical signals generated by the artificial neurons to slices of mouse cerebellum, a brain region critical for motor control and coordination.

The results were highly encouraging. The electrical spikes produced by the artificial neurons closely matched key biological properties of real neuronal signals, including their precise timing and duration. These signals reliably activated the living neurons and successfully triggered neural circuits in a manner indistinguishable from natural brain activity.

"Other labs have tried to make artificial neurons with organic materials, and they spiked too slowly," Hersam noted, highlighting the temporal precision achieved. "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." This temporal and waveform fidelity is critical for bidirectional communication and integration with biological neural networks.

Sustainable Manufacturing and the Urgent Need for Energy-Efficient AI

Beyond their remarkable performance, the new approach offers substantial environmental and economic advantages. The manufacturing process is inherently simple and cost-effective. Furthermore, the additive nature of aerosol jet printing ensures that material is deposited only where it is needed, minimizing waste and contributing to a more sustainable production model.

The imperative for improved energy efficiency in artificial intelligence is becoming increasingly critical. As AI systems grow in complexity and capability, their power demands are escalating at an alarming rate. Large data centers, the backbone of modern AI infrastructure, already consume enormous amounts of electricity and require substantial water resources for cooling.

"To meet the energy demands of AI, tech companies are building gigawatt data centers powered by dedicated nuclear power plants," Hersam stated, underscoring the magnitude of the challenge. "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."

The research, titled "Multi-order complexity spiking neurons enabled by printed MoS₂ memristive nanosheet networks," was generously supported by the National Science Foundation, a testament to its perceived scientific significance and potential impact. This advancement by Northwestern engineers not only pushes the boundaries of what is possible in neural interfacing and computing but also offers a tangible path toward a more sustainable and energy-efficient technological future. The study’s publication in Nature Nanotechnology signifies its reception by the scientific community as a landmark contribution to the fields of materials science, electrical engineering, and neuroscience.