Penn State researchers have developed soft networks of water-based droplets that can process information in ways similar to the human brain. The work lays groundwork for a new kind of soft computer designed for wet environments, one that uses ions to carry signals at far lower energy cost than conventional electronics.
How the droplet networks work
The researchers immerse tiny droplets containing lipids in oil. Lipid molecules naturally arrange around each droplet, and when neighboring droplets touch, they form a thin two-layer membrane between them. These soft surfaces contain voltage-sensitive channels that allow electrical signals to pass between droplets through the movement of ions. The membranes store charge, while the channels respond to present and recent signals. Together, these properties let the droplets form an interconnected network the researchers call "neuromorphic tissue."
The term neuromorphic refers to brain-like computing. Reservoir computing is a technique where a changing input produces responses that retain traces of earlier inputs - in this network, membrane charging and ion-channel activity generate those responses.
Individual droplets can connect to as many as six neighbors in honeycomb-like arrangements. When coupled this way, the network develops rich electrical dynamics capable of processing complex signals that change over time. The connections allow droplets to influence one another, accumulating charge and passing it along - behavior similar to neurons that allows rapid signal transfer without losing clarity.
The membranes give the network short-term, fading memory, so its response depends partly on recent inputs. In testing, the team found that connected networks reduced prediction error by nearly 50% compared to independent nodes when modeling chaotic behavior.
This matters because traditional electronics and biology process information differently. Electronics use electrons moving through metal wires, while biological signaling relies on ions moving through fluids and across cell membranes. Creating interfaces between these systems requires translating between electronic and ionic signals, which can delay and distort signals while consuming significant energy. This approach suggests a way to compute directly with ions, potentially enabling better integration of sensors and computers with biological systems.
This could result in commercially scalable ionic computing, with potential applications in adaptive sensors, soft robotics, and computing at biological interfaces. Their ultimate goal is creating "jelly-like computer chips" that could complement and optimize electronics in the body.