A team of scientists at École Polytechnique Fédérale de Lausanne (EPFL) has redesigned bacterial nanopores to behave like adaptive neural elements, a development that could reshape molecular sensing and inspire new ion-based computing hardware. Led by Matteo Dal Peraro and Aleksandra Radenovic, the researchers reported their results in Nature Nanotechnology on November 11, 2025, describing how careful control of charge and mechanics inside a pore produces behavior analogous to synaptic plasticity.
Solving long-standing nanopore puzzles
Biological nanopores such as aerolysin are already staples of single-molecule detection and DNA sequencing, but they sometimes act erratically, limiting performance. Two persistent problems, rectification, where ionic current depends on the direction of applied voltage, and gating, where conduction stops abruptly, have proven difficult to predict or control. The EPFL group systematically probed these behaviors and revealed the molecular origins that link pore chemistry and structure to electrical response.
A systematic engineering approach
The researchers constructed 26 distinct versions of the aerolysin pore, varying where and which electrical charges line the channel. By applying alternating voltage waveforms, running simulations and fitting biophysical models to experimental data, they were able to separate fast rectification effects from slower gating events. Their experiments showed that rectification arises when asymmetric charge distributions steer ions preferentially in one direction, effectively making the pore act like a nanoscale diode. Gating, by contrast, appears when intense ion flow perturbs the internal charge balance and the pore’s conformation, momentarily collapsing ion passage.
Controlling gating through polarity and rigidity
Two striking controls emerged. Reversing the sign of internal charges tuned the threshold and timing of gating, enabling the team to program when a pore would shut down under electrical stimulation. Increasing the mechanical stiffness of the pore, making the protein scaffold less flexible, eliminated gating entirely, confirming that structural dynamics are essential for the stop-start behavior. These knobs give designers the means either to suppress unwanted interruptions for more reliable sensing or to exploit gating deliberately as a computational mechanism.
A nanopore that “learns” from voltage pulses
The most eye-catching demonstration was a pore engineered to display an activity-dependent change in conductance reminiscent of synaptic plasticity. When subjected to trains of voltage pulses, the modified aerolysin altered its ionic throughput in a persistent way, akin to how a synapse strengthens or weakens with neural activity. That functional analogy points to a new class of molecular devices that encode and adapt information in response to electrical signals, opening the door to ion-based processors that compute by changing transport properties rather than switching transistors.
Applications: better sensors and bio-inspired processors
For the sensing community, the ability to minimize rectification and gating means more stable, high-fidelity readouts for DNA and molecular analytes. For the nascent field of bio-inspired computing, engineered gating and plasticity suggest compact, low-energy primitives that mimic biological information processing at the nanoscale. Combining precision charge placement with controlled mechanical design could enable arrays of learning nanopores that perform pattern recognition or memory storage using ionic rather than electronic currents.
Team and collaboration
The study was carried out at EPFL with contributions from groups at the Institute of Science and Technology Austria, the University of Washington, and ENS de Lyon. By bridging nanoscale transport mechanisms, computational models, and device-level performance, the collaboration establishes charge distribution and structural flexibility as central design parameters for future nanopore technologies.
What’s next
Future work will likely expand the palette of pore scaffolds and explore how networks of adaptive channels can be integrated into practical devices. Whether the immediate payoff is more reliable molecular diagnostics or an emergent class of ion-based neuromorphic components, the EPFL advance provides a clear roadmap: control the chemistry and mechanics inside a pore, and you can tune not only what it senses but how it “learns.”
