Memristive plasticity in artificial electrical synapses via geometrically reconfigurable, gramicidin-doped biomembranes

Artificial Funnel Nanochannel Device Emulates Synaptic Behavior

Creating artificial synapses that can interact with biological neural systems is critical for developing advanced intelligent systems. However, there are still many difficulties, including device morphology and fluid selection. Based on Micro-Electro-Mechanical System technologies, we utilized two immiscible electrolytes to form a liquid/liquid interface at the tip of a funnel nanochannel, effectively enabling a wafer-level fabrication, interactions between multiple information carriers, and electron-to-chemical signal transitions. The distinctive ionic transport properties successfully achieved a hysteresis in ionic transport, resulting in adjustable multistage conductance gradient and synaptic functions. Notably, the device is similar to biological systems in terms of structure and signal carriers, especially for the low operating voltage (200 mV), which matches the biological neural potential (∼110 mV). This work lays the foundation for realizing the function of iontronics neuromorphic computing at ultralow operating voltages and in-memory computing, which can break the limits of information barriers for brain-machine interfaces.

Modulation of Voltage-Gating and Hysteresis of Lysenin Channels by Cu2+ Ions

The intricate voltage regulation presented by lysenin channels reconstituted in artificial lipid membranes leads to a strong hysteresis in conductance, bistability, and memory. Prior investigations on lysenin channels indicate that the hysteresis is modulated by multivalent cations which are also capable of eliciting single-step conformational changes and transitions to stable closed or sub-conducting states. However, the influence on voltage regulation of Cu2+ ions, capable of completely closing the lysenin channels in a two-step process, was not sufficiently addressed. In this respect, we employed electrophysiology approaches to investigate the response of lysenin channels to variable voltage stimuli in the presence of small concentrations of Cu2+ ions. Our experimental results showed that the hysteretic behavior, recorded in response to variable voltage ramps, is accentuated in the presence of Cu2+ ions. Using simultaneous AC/DC stimulation, we were able to determine that Cu2+ prevents the reopening of channels previously closed by depolarizing potentials and the channels remain in the closed state even in the absence of a transmembrane voltage. In addition, we showed that Cu2+ addition reinstates the voltage gating and hysteretic behavior of lysenin channels reconstituted in neutral lipid membranes in which lysenin channels lose their voltage-regulating properties. In the presence of Cu2+ ions, lysenin not only regained the voltage gating but also behaved like a long-term molecular memory controlled by electrical potentials.

Recent progress in bio-voltage memristors working with ultralow voltage of biological amplitude

Neuromorphic systems built from memristors that emulate bioelectrical information processing in the brain may overcome the limitations of traditional computing architectures. However, functional emulation alone may still not attain all the merits of bio-computation, which uses action potentials of 50-120 mV at least 10 times lower than signal amplitude in conventional electronics to achieve extraordinary power efficiency and effective functional integration. Reducing the functional voltage in memristors to this biological amplitude can thus advance neuromorphic engineering and bio-emulated integration. This review aims to provide a timely update on the effort and progress in this burgeoning research direction, covering the aspects of device material composition, performance, working mechanism, and potential application.

Entrapment and Voltage-Driven Reorganization of Hydrophobic Nanoparticles in Planar Phospholipid Bilayers

Engineered nanoparticles (NPs) possess diverse physical and chemical properties, which make them attractive agents for targeted cellular interactions within the human body. Once affiliated with the plasma membrane, NPs can become embedded within its hydrophobic core, which can limit the intended therapeutic functionality and affect the associated toxicity. As such, understanding the physical effects of embedded NPs on a plasma membrane is critical to understanding their design and clinical use. Here, we demonstrate that functionalized, hydrophobic gold NPs dissolved in oil can be directly trapped within the hydrophobic interior of a phospholipid membrane assembled using the droplet interface bilayer technique. This approach to model membrane formation preserves lateral lipid diffusion found in cell membranes and permits simultaneous imaging and electrophysiology to study the effects of embedded NPs on the electromechanical properties of the bilayer. We show that trapped NPs enhance ion conductance and lateral membrane tension in 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DPhPC) bilayers while lowering the adhesive energy of the joined droplets. Embedded NPs also cause changes in bilayer capacitance and area in response to applied voltage, which are nonmonotonic for DOPC bilayers. This electrophysical characterization can reveal NP entrapment without relying on changes in membrane thickness. By evaluating the energetic components of membrane tension under an applied potential, we demonstrate that these nonmonotonic, voltage-dependent responses are caused by reversible clustering of NPs within the unsaturated DOPC membrane core; aggregates form spontaneously at low voltages and are dispersed by higher transmembrane potentials of magnitude similar to those found in the cellular environment. These findings allow for a better understanding of lipid-dependent NP interactions, while providing a platform to study relationships between other hydrophobic nanomaterials and organic membranes.

Homogeneous hybrid droplet interface bilayers assembled from binary mixtures of DPhPC phospholipids and PB-b-PEO diblock copolymers

Hybrid membranes built from phospholipids and amphiphilic block copolymers seek to capitalize on the benefits of both constituents for constructing biomimetic interfaces with improved performance. However, hybrid membranes have not been formed or studied using the droplet interface bilayer (DIB) method, an approach that offers advantages for revealing nanoscale changes in membrane structure and mechanics and offers a path toward assembling higher-order tissues. We report on hybrid droplet interface bilayers (hDIBs) formed in hexadecane from binary mixtures of synthetic diphytanoyl phosphatidylcholine (DPhPC) lipids and low molecular weight 1,2 polybutadiene-b-polyethylene oxide (PBPEO) amphiphilic block copolymers and use electrophysiology measurements and imaging to assess the effects of PBPEO in the membrane. This work reveals that hDIBs containing up to 15 mol% PBPEO plus DPhPC are homogeneously mixtures of lipids and polymers, remain highly resistive to ion transport, and are stable-including under applied voltage. Moreover, they exhibit hydrophobic thicknesses similar to DPhPC-only bilayers, but also have significantly lower values of membrane tension. These characteristics coincide with reduced energy of adhesion between droplets and the formation of alamethicin ion channels at significantly lower threshold voltages, demonstrating that even moderate amounts of amphiphilic block copolymers in a lipid bilayer provide a route for tuning the physical properties of a biomimetic membrane.

Enhancing membrane-based soft materials with magnetic reconfiguration events

Adaptive and bioinspired droplet-based materials are built using the droplet interface bilayer (DIB) technique, assembling networks of lipid membranes through adhered microdroplets. The properties of these lipid membranes are linked to the properties of the droplets forming the interface. Consequently, rearranging the relative positions of the droplets within the network will also alter the properties of the lipid membranes formed between them, modifying the transmembrane exchanges between neighboring compartments. In this work, we achieved this through the use of magnetic fluids or ferrofluids selectively dispersed within the droplet-phase of DIB structures. First, the ferrofluid DIB properties are optimized for reconfiguration using a coupled experimental-computational approach, exploring the ideal parameters for droplet manipulation through magnetic fields. Next, these findings are applied towards larger, magnetically-heterogeneous collections of DIBs to investigate magnetically-driven reconfiguration events. Activating electromagnets bordering the DIB networks generates rearrangement events by separating and reforming the interfacial membranes bordering the dispersed magnetic compartments. These findings enable the production of dynamic droplet networks capable of modifying their underlying membranous architecture through magnetic forces.

A skin-inspired soft material with directional mechanosensation

Lessons about artificial sensor design may be taken from evolutionarily perfected physiological systems. Mechanosensory cells in human skin are exquisitely sensitive to gentle touch and enable us to distinguish objects of different stiffnesses and textures. These cells are embedded in soft epidermal layers of gel-like consistency. Reproducing these mechanosensing capabilities in new soft materials may lead to the development of adaptive mechanosensors which will further enhance the abilities of engineered membrane-based structures with bioinspired sensing strategies. This strategy is explored here using droplet interface bilayers embedded within a thermoreversible organogel. The interface between two lipid-coated aqueous inclusions contained within a soft polymeric matrix forms a lipid bilayer resembling the lipid matrix of cell membranes. These interfaces are functionalized with bacterial mechanosensitive channels (V23T MscL) which convert membrane tension into changes in membrane conductance, mimicking mechanosensitive channel activation in mammalian mechanosensory cells. The distortion of encapsulated adhered droplets by cyclical external forces are first explored using a finite element composite model illustrating the directional propagation of mechanical disturbances imposed by a piston. The model predicts that the orientation of the droplet pair forming the membrane relative to the direction of the compression plays a role in the membrane response. The directional dependence of mechanosensitive channel activation in response to gel compression is confirmed experimentally and shows that purely compressive perturbations normal to the interface invoke different channel activities as compared to shearing displacement along a plane of the membrane. The developed system containing specially positioned pairs of droplets functionalized with bacterial mechanosensitive channels and embedded in a gel creates a skin-inspired soft material with a directional response to mechanical perturbation.

Droplet-Based Membranous Soft Materials

Inspired by the structure and functionality of natural cellular tissues, droplet interface bilayer (DIB)-based materials strategically combine model membrane assembly techniques and droplet microfluidics. These structures have shown promising results in applications ranging from biological computing to chemical microrobots. This Feature Article briefly explores recent advances in the areas of construction, manipulation, and functionalization of DIB networks; discusses their unique mechanics; and focuses on the contributions of our lab in the advancement of this platform. We also reflect on some of the limitations facing DIB-based materials and how they might be addressed, highlighting promising applications made possible through the refinement of the material concept.

Emulation of biphasic plasticity in retinal electrical synapses for light-adaptive pattern pre-processing

Electrical synapses provide rapid, bidirectional communication in nervous systems, accomplishing tasks distinct from and complementary to chemical synapses. Here, we demonstrate an artificial electrical synapse based on second-order conductance transition (SOCT) in an Ag-based memristor for the first time. High-resolution transmission electron microscopy indicates that SOCT is mediated by the virtual silver electrode. Besides the conventional chemical synaptic behaviors, the biphasic plasticity of electrical synapses is well emulated by integrating the device with a photosensitive element to form an optical pre-processing unit (OPU), which contributes to the retinal neural circuitry and is adaptive to ambient illumination. By synergizing the OPU and spiking neural network (SNN), adaptive pattern recognition tasks are accomplished under different light and noise settings. This work not only contributes to the further completion of synaptic behaviour for hardware-level neuromorphic computing, but also potentially enables image pre-processing with light adaptation and noise suppression for adaptive visual recognition.

Memristive synapses with high reproducibility for flexible neuromorphic networks based on biological nanocomposites

Memristive synapses from biomaterials are promising for building flexible and implantable artificial neuromorphic systems due to their remarkable mechanical and biological properties. However, these biological devices have relatively poor memristive switching characteristics, and thus fail to meet the requirement of neuromorphic networks for high learning accuracy. Here, memristive synapses based on carrageenan nanocomposites that possess desirable characteristics are demonstrated. These devices show highly reproducible analog resistive switching behaviors with 250 conductance states, low write noise, good write linearity, high retention of more than 104 s and endurance for at least 106 pulses. The enhanced switching properties are attributed to controllable and confined conductive filament growth, owing to the synergistic effect of self-assembled silver nanocluster doping and nanocone-shaped electrode contact. Moreover, the devices exhibit excellent reliability after 1000 bending cycles. Simulations including the non-ideal factors prove that the synaptic device array can operate with an online learning accuracy of 94.3%. These findings enable broader applications of biomaterials in flexible memristive devices and neuromorphic systems.