Artificial neurons are flexible, low-cost electronic devices engineered to replicate the signaling behavior of biological neurons by generating and transmitting electrical signals that can interface with living brain tissue. These biocompatible devices represent a convergence of neuroscience, materials science, and bioelectronics, designed to be soft enough to match brain tissue architecture while maintaining functional compatibility with neural systems. Such devices have potential applications in neural interfacing, brain-computer interfaces (BCIs), and energy-efficient alternative computing paradigms. Recent research at Northwestern University has demonstrated that these bio-inspired devices can trigger electrical responses in live mouse brain tissue, validating their potential as energy-efficient alternatives to power-hungry AI systems 1).
Artificial neurons are fundamentally inspired by the structure and function of biological neurons, which communicate through action potentials—rapid changes in electrical potential across cell membranes. Traditional computational approaches to modeling neurons, such as the artificial neural networks used in deep learning, abstract away the physical electrical signaling mechanism in favor of mathematical operations. Conversely, artificial neuron devices seek to more directly mimic the actual electrochemical signaling mechanisms of biological systems 2).
The key design challenge involves creating materials that are simultaneously electronically functional and mechanically soft. Biological neural tissue exhibits Young's moduli in the range of approximately 1-10 kPa, whereas traditional silicon-based electronics are orders of magnitude stiffer. This mechanical mismatch can cause inflammatory responses and signal degradation at the tissue-device interface. Modern artificial neurons employ soft materials such as conducting polymers, hydrogels, and organic semiconductors to achieve mechanical compliance while maintaining electrical responsiveness 3).
Artificial neuron devices typically incorporate several functional components: a stimulus reception mechanism, an integration zone that sums inputs, a threshold mechanism, and an output firing stage. These devices can be constructed using organic electronics, which offer superior biocompatibility compared to inorganic semiconductors. The electrical signals generated by artificial neurons must achieve sufficient amplitude and temporal precision to trigger responses in connected biological neurons through synaptic transmission or electrical coupling.
Energy efficiency represents a significant advantage of artificial neurons over traditional semiconductor-based computing. Biological neurons operate at approximately 10-20 nanowatts per spike, whereas conventional silicon transistors consume significantly more power. Artificial neurons constructed from organic materials can approach biological energy efficiency levels, making them potentially valuable for implantable long-term applications where power delivery through external sources is impractical 4).
The primary application domain for artificial neurons involves neural interface systems. Brain-computer interfaces (BCIs) that employ artificial neurons could enable bidirectional communication with neural tissue, supporting therapeutic applications for neurological disorders, sensory restoration, and motor rehabilitation. The mechanical compliance of artificial neurons minimizes chronic immune responses that plague conventional rigid electrode arrays, potentially extending functional lifespan from months to years 5).
Beyond medical applications, researchers are exploring artificial neurons as components of hybrid computing systems that combine biological neural tissue with electronic devices. Such biohybrid systems might leverage the pattern recognition capabilities of biological neural networks while maintaining electronic interfaces for data processing and external communication. Some research explores growing biological neurons on artificial neuron substrates or directly interfacing engineered electronic neurons with cultured neural networks.
Several significant challenges remain in the development of practical artificial neurons. Signal-to-noise ratio management is critical, as biological neural signals operate at millivolt scales, requiring sensitive amplification without introducing excessive electrical noise. The long-term stability of organic electronic materials in the aqueous, protein-rich environment of neural tissue remains incompletely resolved, with device degradation occurring over months to years in some applications.
Integration complexity increases substantially when attempting to create artificial neuron networks with comparable density and connectivity to biological brains. A single human brain contains approximately 86 billion neurons with trillions of synaptic connections; replicating this scale with current manufacturing capabilities would require revolutionary advances in microfabrication, materials synthesis, and system integration.
Biocompatibility remains a nuanced challenge—while soft materials reduce mechanical damage, they may still trigger immune responses or exhibit unfavorable protein absorption characteristics. Additionally, establishing reliable electrical coupling between artificial and biological neurons requires precise control over electrode surface properties and electrochemical conditions.
Emerging research directions include development of self-assembling artificial neuron networks, incorporation of ionic transport mechanisms to more closely mimic biological ion channels, and integration of on-device learning mechanisms using memristive elements. The field is gradually moving toward fully implantable, long-term stable systems that could support clinical translation for neurological applications.