Short-Term Facilitation-Then-Depression Enables Adaptive Processing of Sensory Inputs by Ion Channels in Biomolecular Synapses

Abstract

Providing AI platforms with perceptual capabilities at low energy cost is imperative to making them more human-like. This goal is contingent on developing stimuliresponsive materials that closely emulate diverse synaptic functions needed to enhance machine learning applications. Here a biomolecular device is reported, consisting of an insulating lipid membrane doped with voltage-activated, ion channel-forming monazomycin (Mz) species, that display diode-like current−voltage characteristics and emulate short-term facilitation-then-depression on time scales relevant to habituation in human sensory systems. Subjected to a slow train of voltage pulses, devices show only facilitation upon Mz channel formation. At higher pulse rates, faster facilitation creates later depression upon Mz inactivation. The device is integrated as a biomolecular synapse in an organic−inorganic hybrid afferent model to demonstrate its ability to process sensory inputs and mimic bidirectional retinal adaptation and sensitization. These findings highlight the value of instilling complex synaptic plasticity into engineered systems, which is useful for adaptive neuroprosthetics and biosignal processing.

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

“…Besides BMRs constructed from conventional tri-layer solid-state materials, devices made from a biomembrane sandwiched between two aqueous solutions also exhibited bio-voltage memristive and memcapacitive behaviors. 72–76 A bio-voltage signal can induce a sufficiently high electrical field across the biomembrane due to its ultrathin thickness ( e.g. , 3–5 nm), which is believed to change the structure ( e.g.…”

“…Besides BMRs constructed from conventional tri-layer solidstate materials, devices made from a biomembrane sandwiched between two aqueous solutions also exhibited biovoltage memristive and memcapacitive behaviors. [72][73][74][75][76] A biovoltage signal can induce a sufficiently high electrical field across the biomembrane due to its ultrathin thickness (e.g., 3-5 nm), which is believed to change the structure (e.g., through peptide insertion) or interfacing area (e.g., through electrowetting) in the biomembrane for resistive or capacitive modulation, respectively. The modulation was found to be reversible, thus yielding volatile switching behaviors.…”

Bioderived materials for stimuli-responsive, adaptive, and neuromorphic systems: A perspective

“…7,98,103,104 3-state model that describes the dynamic plasticity and built a hybrid afferent neuron-synapse circuit to demonstrate adaptive signal processing capability, specifically neural habituation to a strong light source. 86 The functionality of the nervous system relies on nonthreshold dependent transport of ions between neurons via electrical synapses as well as threshold-dependent communication. 96 In a similar approach to the sections above, Koner et al assembled lipid membranes containing gramicidin peptides that spontaneously form ion channels 84 (Figure 4(c)).…”

“…Whereas the work by Zocchi et al focused on emulating artificial axons, Sarles et al recently introduced voltageactivated biomembranes using the Droplet Interface Bilayer (DIB) platform for membrane assembly, which yielded volatile memory resistance (and capacitance) and short-term activity-dependent plasticity when voltage-activated ion channels were integrated. [81][82][83][84][85][86] Figure 4(a) shows ion current-voltage relationships for a lipid membrane with embedded Alamethicin pores (ALM was present on both sides of the membrane). These measurements (recorded at multiple scan rates) show that the threshold for ion channel formation is around 120 mV and exhibit a hysteretic path upon a subsequent drop in voltage.…”

“…The authors developed a kinetic 3-state model that describes the dynamic plasticity and built a hybrid afferent neuron-synapse circuit to demonstrate adaptive signal processing capability, specifically neural habituation to a strong light source. 86…”