Skip to main content
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2024 Jul 3;121(28):e2403143121. doi: 10.1073/pnas.2403143121

A nanofluidic spiking synapse

Yi-Tong Xu a, Si-Yuan Yu a, Zheng Li a, Bo-Han Kou a, Jian-Xiang Pang a, Wei-Wei Zhao a,1, Hong-Yuan Chen a, Jing-Juan Xu a,1
PMCID: PMC11252921  PMID: 38959041

Significance

Nanofluidics has recently revolutionized the traditional solid-state neuromorphic emulations, yet a realization of a robust nanofluidic spiking device remains a challenge. Here, we report a nanofluidic spiking synapse based on a poly (3,4-ethylenedioxythiophene) polystyrene sulfonate membrane confined within a nanopore, which mimics time-dependent plasticity and the biological transmembrane spiking with ionic and dopaminergic tunability. The shared ionic current-induced potential spiking in aqueous environment suggests that this device has the potential to communicate with biological synapses.

Keywords: nanofluidic, bionic synapse, transmembrane, potential spiking, chemical tunability

Abstract

Currently, the nanofluidic synapse can only perform basic neuromorphic pulse patterns. One immediate problem that needs to be addressed to further its capability of brain-like computing is the realization of a nanofluidic spiking device. Here, we report the use of a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate membrane to achieve bionic ionic current-induced spiking. In addition to the simulation of various electrical pulse patterns, our synapse could produce transmembrane ionic current-induced spiking, which is highly analogous to biological action potentials with similar phases and excitability. Moreover, the spiking properties could be modulated by ions and neurochemicals. We expect that this work could contribute to biomimetic spiking computing in solution.


Human brain processes information through discrete neuronal spikes with remarkable energy efficiency. Guided by the brain, spike-driven neuromorphic computing has been actively studied across both algorithmic and hardware domains. Training algorithms for spiking neural networks, very-large-scale integration neuromorphic chips, and algorithm-hardware codesign have made incredible progress (1). Nevertheless, the synaptic transmission in hardware implementation of algorithms takes place fundamentally different from the spike-based computations of biological synapses, which employ multiple carriers of ions and biomolecules in aqueous milieu. In this regard, organic electrochemical transistors have been explored to emulate neuronal spiking in electrolytes (2–4).

Nanofluidic technology has undergone a great boost in the past decade, from separation to biosensors and to memories (5–7), which is expected to produce the ionic memristors to revolutionize traditional solid-state neuromorphic engineering. Xiong et al. recently established polyelectrolyte-confined fluidic memristor, wherein the ion transport in spatial confinement could be affected by polyelectrolyte–ion interactions, making possible the neuroplastic functions as regulated by ions and adenosine triphosphate (8). Robin et al. simultaneously reported the ion-mediated neuroplasticity with nonvolatile ionic memory and Hebbian learning (9). Despite the progress, nanofluidic emulation of neuronal spiking remains a grand challenge as indicated by Noy and Darling (10).

Here, we report a nanofluidic spiking synapse via confinement of a polymeric membrane within the nanopipette tip, which could well mimic the biological spiking with ionic and dopaminergic tunability. In biological synapse (Fig. 1A), signaling transmission is mediated by action potentials (APs) that are induced by the transmembrane ionic current and modulated by neurochemicals. Inspired by this, poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS) membrane was confined within a nanopipette (Fig. 1B). PEDOT: PSS was chosen due to its cation permeability, operational stability, volumetric electronegativity, and easy formation of membrane morphology (11–13). Upon such a PEDOT: PSS nanofluidic synapse, various electrical pulse patterns, e.g., paired-pulse facilitation/depression (PPF/PPD), spike-rate-dependent-plasticity (SRDP), and spike-timing-dependent-plasticity (STDP), could be realized. Significantly, ionic current-induced spiking that is highly analogous to APs could be produced, with similar phases of resting potential, depolarization, and hyperpolarization. Especially, the bionic spiking could also be modulated by ions and neurochemicals.

Fig. 1.

Fig. 1.

PEDOT: PSS nanofluidic synapse. (A) Schematic of the biological APs transmission mediated by transmembrane ionic currents. (B) Schematic of the nanofluidic device with a PEODT: PSS membrane and the measurement setup. (C) I–V response, quasi-equivalent circuit, and derived conductance (Inset) of the device under periodic triangular wave of 100 mV/s scan rate in a 100 mM KCl solution. The arrows indicate the scan direction. (D) The ICP within the device was enhanced and inhibited during the scanning of external voltage. The larger/smaller size of solid arrows indicate the increasing/decreasing rate of enhancement or inhibition. (E) Evolution of the geometric hysteresis area against the scan rates. (F) Transient current and conductance changes under constant +0.3 (red curve) and −0.5 V (blue curve).

Results and Discussion

PEDOT: PSS Nanofluidic Synapse.

The confinement of PEDOT: PSS membrane at the nanotip was confirmed by the scanning electron microscope (SEM) and transmission electron microscope (TEM) images (SI Appendix, Fig. S1). The nanofluidic device was then tested upon periodic triangular wave between −1 V and +1 V at a scan rate of 100 mV/s in a 100 mM KCl solution, which manifested both ionic rectification and pinched hysteresis characteristics in the current-potential (I–V) curve (Fig. 1C) with good stability over 1,000 repeated cycles (SI Appendix, Fig. S2). The rectification and hysteresis behavior closely resembled the collected properties of a diode and a memristor, which stemmed from the asymmetrical charge distribution in the confined nanopore (14–17) and the capacitive effect of PEDOT: PSS membrane (18, 19), respectively. Note that the hysteresis loop intersected itself near 0 V and exhibited a clockwise/anticlockwise trend in the first/third quadrant, respectively, yielding the monocyclic conductance switching (Fig. 1 C, Inset). Incidentally, the pristine nanopipette without PEDOT: PSS membrane (SI Appendix, Fig. S3), the nanopipettes with overoxidized (SI Appendix, Fig. S4) or cross-linked PEDOT: PSS (SI Appendix, Fig. S5) could only produce linear curves without hysteresis, indicating the crucial role of PEDOT: PSS membrane in generating the ionic hysteresis.

Essentially, the negative rectification originated from negatively charged sulfonic acid moieties in the PEDOT: PSS membrane (20, 21) and geometrical asymmetry of the conical nanotip (22–24). These effects would lead to cations being the main carriers (SI Appendix, Fig. S6) and their accumulation at boundary layer of tip side and depletion at boundary layer of base side (25, 26), and thus a built-in ion concentration polarization (ICP) from the tip to base side (Fig. 1D). Based on the reversible redox process in PEDOT: PSS and its capacitive effect, the hysteresis could be observed following several processes (27): in the first quadrant, the external positive voltage would drive the cations from the lumen to the outside bulk solution, accompanied by an opposite flux of anion that is blocked by the negative charged membrane, leading to the salt depletion/accumulation at the base/tip side and the continuously enhanced ICP. With the scanning from 0 to +1 V and then back to 0 V, the polarization effect was time-accumulated, which would lead to a lower current value of the scanning from +1 to 0 V than that of the scanning from 0 to +1 V. In the third quadrant, the external negative voltage would drive the cations from the outside bulk solution into the lumen, accompanied by an opposite flux of anion that is blocked by the negative charged membrane, leading to the salt accumulation/depletion at the base/tip side and the continuously inhibited ICP. With the scanning from 0 to –1 V and then back to 0 V, the time-accumulated depolarization could lead to a larger current value in the scanning from –1 to 0 V than that in the scanning from 0 to – 1 V.

The hysteresis characteristics were further studied at different scan rates ranging from 20 to 5,000 mV/s (SI Appendix, Fig. S7). The geometric areas of the hysteresis loops were derived and plotted against the scan rate, revealing an upward trend before 100 mV/s and a downward trend thereafter (Fig. 1E). Such a phenomenon was due to sufficient ionic equilibrium at low scan rates and the lagged ionic changes at high scan rates (28), which was consistent with Hodgkin–Huxley ion-channel memristors (29–31). It implied a time-dependent process, which could be demonstrated by sampling transient current and calculating conductance changes under constant potential. Both of current and conductance became smaller/larger under positive/negative potential until reaching a steady state after 10 s (Fig. 1 F, Inset), confirming the time-dependent ionic polarization/depolarization processes and thus the memristive behaviors. This property could be also verified by the current responses under different switch conditions (SI Appendix, Fig. S8).

Time-Dependent Synaptic Plasticity.

Synaptic plasticity plays a significant role in forming learning and memory. As a form of short-term plasticity, PPD/PPF refers to the lower/higher response amplitude upon the second stimulus pulse (A2) than that of the first pulse (A1), respectively (32, 33). These phenomena could be independently mimicked using our nanofluidic synapse. Two presynaptic pulses with +0.4 V intensity, 10 ms duration, and 20 ms interval (Δt) were initially applied, yielding postsynaptic current peaks of 21.70 and 19.84 nA (Fig. 2A). In the scenario of −0.4 V stimuli, peaks of −28.42 and −29.04 nA could be produced (Fig. 2B). The signals originated from the preservation of polarization/depolarization states when the second pulse arrived, aligning well with the time-dependent characteristic. The preservation time in such conditions could be deduced by plotting A2/A1 against the Δt. As the Δt increased, the A2/A1 values increased for PPD and decreased for PPF (Fig. 2C). Time constants for PPD and PPF with different values could be derived, indicating the asymmetric polarization/depolarization dynamics (Fig. 2 C, Inset). Pair pulses with 100 ms duration were also applied (SI Appendix, Fig. S9), leading to larger time constants compared to that of 10 ms. These time constants spanned from several to hundreds of milliseconds, coinciding with that of the biological system (34).

Fig. 2.

Fig. 2.

Time-dependent synaptic plasticity. (A) PPD and (B) PPF behaviors of the synapse under Δt = 20 ms. (C) A2/A1 − Δt plots for PPD/PPF and the corresponding fitting curves. The inset shows the derived time constants. (D) SRDP signals under 50 successive pulses with +/−0.4 V intensity (+ for upper and – for lower), 10 ms duration, and different intervals, corresponding to the frequency ranging from 67 to 1 Hz (from left to right). (E) STDP signals and the calculated ΔConductance against different ΔT+/− values. The input signals at +/−0.1 V were used for calculation of ΔConductance.

The short-term plasticity contributes to the temporal filtering as the change of synaptic plasticity depends upon multiple spikes with varied rates. This evolution could be simulated by applying 50 presynaptic pulses with +/−0.4 V intensity, 10 ms duration, and different intervals (corresponding to the frequency of 1 to 67 Hz) upon our nanofluidic synapse. The postsynaptic current decreased/increased obviously at high frequency and showed little changes at the low frequency, which was attributable to the accumulation of polarization/depolarization degree at high frequency (Fig. 2D). STDP is another form of synaptic plasticity describing the connection change between the pre- and postsynaptic against their spikes order and interval (ΔT). This process could be imitated by treating a positive pulse (+1 V) as the presynaptic spike, a negative pulse (−1 V) as the postsynaptic spike, and small voltage (+/−0.1 V) of 10 ms after the latter pulse as the connection readout. When −1 V pulse first arrived (ΔT+/− < 0), the conductance variation (ΔConductance) increased as the ΔT+/− came near 0 s, which was due to a higher preservation of the predepolarization state that facilitated the ionic transportation (Fig. 2 E, Left Inset). When +1 V pulse first arrived (ΔT+/− > 0), the ΔConductance decreased as the ΔT+/− came near 0 s, owing to a higher preservation of the prepolarization state that hindered the ionic transportation (Fig. 2 E, Right Inset). These phenomena were consistent with the antisymmetric anti-Hebbian learning rule (35).

Bionic Spiking Behaviors.

The biological APs operate upon integration of incoming signals and firing output spikes based on predefined thresholds, which includes several main stages (29, 36–38): i) At rest, the equilibrium with the inherent ionic gradient (Fig. 3A, left column) maintains the transmembrane potential (Vmem) of ca. −70 mV (Fig. 3B stage 1). ii) Upon stimulus, the dominant sodium ion afflux (Fig. 3A, middle column) leads to the depolarization with Vmem positive than −70 mV (Fig. 3B, stage 2). iii) Then, the dominant potassium ion efflux (Fig. 3A, right column) results into repolarization with recovery of Vmem to ca. −70 mV, accompanied by a short hyperpolarization with Vmem negative than −70 mV before resting (Fig. 3B, stage 3). Such a transmembranous electrical dynamics could be well reiterated using our nanofluidic synapse. An initial resting current of −7 nA was applied to induce the high conductance (Fig. 3C, left column), which could easily ordinate the Vmem to ca. −70 mV (Fig. 3D, stage 1). The subsequent switch to the stimulus current of +7 nA allowed the cation accumulation at the boundary layer of tip side and thus the high resistance (Fig. 3C, middle column), which could prompt the Vmem positive than −70 mV to the predefined threshold of +40 mV that is identical to that of APs (Fig. 3D, stage 2). Then, the automatic switch to the resting current of −7 nA led to the fast release of accumulated cations and the recovery of high conductance (Fig. 3C, right column), which could decrease the Vmem to ca. −70 mV with a short overshoot (Fig. 3 D, Inset and stage 3). Importantly, the energy consumption of a single potential spiking was around 77 pJ in our nanofluidic device, which was obviously lower than the nJ level of the organic electrochemical transistors (2–4). The spiking durability of this synapse was further studied over 1,000 cycles (Fig. 3E). The little difference between the first and last 10 spikes indicated its good operational stability.

Fig. 3.

Fig. 3.

Bionic spiking behaviors. (A) Schematic of three main stages in biological generation of APs and (B) the corresponding Vmem evolution. (C) Schematic of three main stages in our bionic spiking synapes and (D) the corresponding highly analogous Vmem under input ionic currents of −/+7 nA. The Inset shows the magnified view of the Vmem signals in dotted box. (E) Stable spiking durability over 1,000 cycles.

Electrical and Chemical Tunability of the Spiking.

The biological generation of APs necessitates a certain intensity of stimulation, and the weak stimuli could not successfully initiate the APs (39). This phenomenon could be emulated in our nanofluidic device through adjusting stimulus current of −5, −1, +3, and +7 nA (Fig. 4A). As shown, different stimulus currents induced Vmem of varying magnitudes and shapes, and only +7 nA could generate the ideal spiking property. Besides, at fixed stimulus current of +7 nA, different threshold values, e.g., +30, +40, and +70 mV, could be realized (Fig. 4B) with more pronounced overshoots (Fig. 4 B, Inset), which could be attributed to the accumulation of more cations at the boundary layer of tip side. Moreover, at the predefined threshold of +40 mV, the level of resting potential could also be tuned (SI Appendix, Fig. S10). In addition, biological APs with different frequencies could further be simulated by controlling of the resting time. With the increased time from 0.5 to 10 s, spikes with frequencies ranging from 1.72 to 0.10 Hz could be generated (Fig. 4C). Spikes with larger profiles could be also achieved (SI Appendix, Fig. S11).

Fig. 4.

Fig. 4.

Spiking tunability by electrical and chemical inputs. (A) Varying spiking behaviors under different stimulus currents with resting current of −7 nA, threshold of +40 mV, and resting time of 2 s. (B) The tunable threshold values with stimulus current of +7 nA, resting current of −7 nA and resting time of 2 s. (C) Tunable spiking frequencies under varying resting time with stimulus current of +7 nA, resting current of −7 nA and threshold of +40 mV. (D) Spiking behaviors under different potassium concentrations. (E) The spiking modulation by neurotransmitter dopamine. For (D and E), the signals were acquired with stimulus current of +7 nA, resting current of −7 nA, threshold of +40 mV, and resting time of 2 s. Insets show the corresponding I–V signals and the right columns show magnified views of the second spikes.

In addition to the types of cations (SI Appendix, Figs. S6 and S12), the spike characteristics of our nanofluidic device could be further modulated by the cation concentrations. For example, upon increased potassium concentration at tip side (cK+,t), the resting potential became less negative (Fig. 4D), which was due to the increased conductance state that could be supported by the I–V curves showing larger current level under larger cK+,t (Fig. 4 D, Inset). A magnified view of the second stimulus reflected that more time was needed to complete a spike as the cK+,t decreased, corresponding to the decreased spike frequency. Such properties were consistent with the biological hypokalemia and hyperkalemia effects that result in hypoexcitability and hyperexcitability, respectively (40). Especially, the spiking could be reshaped by the neurotransmitter dopamine. Upon addition of 5 mM dopamine at the tip side (cDopa,t), the resting potential became more negative and the I–V curves exhibited a smaller current level (Fig. 4 E, Inset), which was due to the decreased conductance state and the electrostatic binding between the dopamine and the PEDOT: PSS (41) (SI Appendix, Fig. S13). Magnified view of the second stimulus indicated that more time was needed to complete a spike in the presence of dopamine, corresponding to the decreased spike frequency. Such a property resembled the binding of dopamine with the D2-like receptors on the cell membrane that could inhibit neuronal excitability (42, 43).

Conclusion

In summary, we have demonstrated a nanofluidic spiking synapse through confinement of a PEDOT: PSS membrane within the nanopipette. Stemming from geometrical asymmetry of the nanopore and capacitance of the PEDOT: PSS membrane, the bionic synapse manifested both ionic rectification and pinched hysteresis, the collected effects of which enabled time-dependent plasticity and especially the ionic current-induced spiking that highly analogous to the biological generation of APs. The corresponding mechanism could be explained by the ICP aside from the PEDOT: PSS membrane. Comparing with existing organic electrochemical transistor spiking devices in aqueous condition, this nanofluidic synapse could be operated with much simplified circuit and lower energy consumption. Moreover, the spiking characteristics could be chemically modulated by different ionic environments and real neurotransmitter dopamine. The shared ionic current-induced spiking in aqueous environment suggests its potential to communicate with biological synapses in a spiking manner. Nevertheless, we acknowledge the difference of this bionic synapse from the biological one, in which the complex coordination among different ion channels within synaptic membranes makes possible the occurrence of APs. Highly close imitation of such rich behaviors remains challenging for current neuromorphic engineering. Using the nanofluidic technique, this work realized the spiking-based signaling, the scalable fabrication of which might facilitate the brain-like neuromorphic computation based on temporally coded spikes using STDP-based learning rules.

Materials and Methods

Reagents and Materials.

Analytical grade chemicals were acquired from Sinopharm Chemical Reagent Co., Ltd. (China) and used as received unless stated otherwise. PEDOT: PSS (Clevios PH1000) was purchased from Heraeus. 3-glycidyloxypropyl trimethoxysilane (GOPS) was obtained from Sigma-Aldrich (St. Louis, MO). All aqueous solutions were prepared from deionized water (Millipore, 18 MΩ cm). Borosilicate capillaries with filament (outside diameter: 1.0 mm; inside diameter: 0.58 mm; length: 10 cm) by fire-polished process was purchased from Sutter Instrument.

Apparatus and Instruments.

The nanopipettes were fabricated with a laser micropuller (P-2000, Sutter Instruments). Current- and potential-related measurements were performed using a CHI 760E electrochemical workstation (Shanghai Chenhua, China). The SEM images were recorded using a Hitachi S4800 SEM (Hitachi Ltd., Japan). The TEM images were acquired using a Model JEOL 2000 TEM (JOEL Ltd., Japan).

Device Preparation and Setup.

The borosilicate capillaries were initially immersed in the piranha solution for 2 h, followed by the ultrasonic dilution and removal of piranha with water for several times. Then, the capillaries were dried at 80 °C for 8 h. After cooling, the capillaries were pulled by the following two steps: i) Heat 375, Fil 3, Vel 40, Del 200, Pul 0; ii) Heat 365, Fil 3, Vel 37, Del 170, Pul 100. The obtained nanotip was immersed into the mixture of PEDOT: PSS and water (v/v = 1: 1) at r. t. for 30 s, followed by a vacuum drying of 10 min to obtain the PEDOT: PSS membrane device. For the overoxidized PEDOT: PSS, the device was treated with 12% NaClO by weight for another 10 min. For the cross-linked PEDOT: PSS, the mixture was added 1% GOPS by volume and well stirred. After fabrication, the devices were filled with 100 mM KCl and inserted by an Ag/AgCl wire that was connected to a working electrode of the station. Then, the devices were immersed into a target solution that was connected with 100 mM KCl using agar-saturated KCl salt bridge. Another Ag/AgCl wire was also immersed into the 100 mM KCl and acting as the counter/reference electrodes. In most cases, the target solution was also 100 mM KCl. For the chronopotentiometry measurement, the input current switched automatically upon the predefined potential thresholds and the resting time. To derive the pair-pulses index, the A2/A1 − Δt plots were fitted by the biexponential decay model: A2/A1=A0+w1e-Δt/τ1+w2e-Δt/τ2, where A0, w, and τ are the final decay value, mixing ratio, and the time constant, respectively.

Supplementary Material

Appendix 01 (PDF)

Acknowledgments

We thank the National Natural Science Foundation of China (Grant Nos. 22034003, 22174063, 22374066, and 21327902), Excellent Research Program of Nanjing University (ZYJH004).

Author contributions

Y.-T.X., W.-W.Z., and J.-J.X. designed research; Y.-T.X. performed research; Y.-T.X., S.-Y.Y., Z.L., B.-H.K., J.-X.P., and H.-Y.C. contributed new reagents/analytic tools; Y.-T.X. and J.-J.X. analyzed data; S.-Y.Y. and Z.L. methodology, visualization, and discussion; B.-H.K. and J.-X.P. visualization, and discussion; W.-W.Z. conceptualization, funding acquisition, supervision, and writing—original draft, review, and editing; H.-Y.C. funding acquisition; J.-J.X. conceptualization, funding acquisition, supervision, and writing—review and editing; and Y.-T.X. and W.-W.Z. wrote the paper.

Competing interests

The authors declare no competing interest.

Footnotes

This article is a PNAS Direct Submission. A.N. is a guest editor invited by the Editorial Board.

Contributor Information

Wei-Wei Zhao, Email: zww@nju.edu.cn.

Jing-Juan Xu, Email: xujj@nju.edu.cn.

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.

Supporting Information

References

  • 1.Roy K., Jaiswal A., Panda P., Towards spike-based machine intelligence with neuromorphic computing. Nature 575, 607–617 (2019). [DOI] [PubMed] [Google Scholar]
  • 2.Harikesh P. C., et al. , Organic electrochemical neurons and synapses with ion mediated spiking. Nat. Commun. 13, 901 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Sarkar T., et al. , An organic artificial spiking neuron for in situ neuromorphic sensing and biointerfacing. Nat. Electron. 5, 774–783 (2022). [Google Scholar]
  • 4.Harikesh P. C., et al. , Ion-tunable antiambipolarity in mixed ion-electron conducting polymers enables biorealistic organic electrochemical neurons. Nat. Mater. 22, 242–248 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Esfandiar A., et al. , Size effect in ion transport through angstrom-scale slits. Science 358, 511–513 (2017). [DOI] [PubMed] [Google Scholar]
  • 6.Bocquet L., Nanofluidics coming of age. Nat. Mater. 19, 254–256 (2020). [DOI] [PubMed] [Google Scholar]
  • 7.Qi Y., Hou X., Bioinspired nanofluidic iontronics. Science 373, 628–629 (2021). [DOI] [PubMed] [Google Scholar]
  • 8.Xiong T., et al. , Neuromorphic functions with a polyelectrolyte-confined fluidic memristor. Science 379, 156–161 (2023). [DOI] [PubMed] [Google Scholar]
  • 9.Robin P., et al. , Long-term memory and synapse-like dynamics in two-dimensional nanofluidic channels. Science 379, 161–167 (2023). [DOI] [PubMed] [Google Scholar]
  • 10.Noy A., Darling S. B., Nanofluidic computing makes a splash. Science 379, 143–144 (2023). [DOI] [PubMed] [Google Scholar]
  • 11.Berggren M., Malliaras G. G., How conducting polymer electrodes operate. Science 364, 233–234 (2019). [DOI] [PubMed] [Google Scholar]
  • 12.Yu S.-Y., et al. , A light-driven nanopipette ion pump for probing subcellular oxygen levels. Device 1, 100001 (2023). [Google Scholar]
  • 13.Rivnay J., et al. , Structural control of mixed ionic and electronic transport in conducting polymers. Nat. Commun. 7, 11287 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Wei C., Bard A. J., Feldberg S. W., Current rectification at quartz nanopipet electrodes. Anal. Chem. 69, 22 (1997). [Google Scholar]
  • 15.Guo W., Tian Y., Jiang L., Asymmetric ion transport through ion-channel-mimetic solid-state nanopores. Acc. Chem. Res. 46, 2834–2846 (2013). [DOI] [PubMed] [Google Scholar]
  • 16.Lin C.-Y., Combs C., Su Y.-S., Yeh L.-H., Siwy Z. S., Rectification of concentration polarization in mesopores leads to high conductance ionic diodes and high performance osmotic power. J. Am. Chem. Soc. 141, 3691–3698 (2019). [DOI] [PubMed] [Google Scholar]
  • 17.Xu Y.-T., et al. , Ion current rectification-nanopipette technique for single-cell analysis. Trends Anal. Chem. 167, 117217 (2023). [Google Scholar]
  • 18.Rivnay J., et al. , High-performance transistors for bioelectronics through tuning of channel thickness. Sci. Adv. 1, e1400251 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Rivnay J., et al. , Organic electrochemical transistors. Nat. Rev. Mater. 3, 17086 (2018). [Google Scholar]
  • 20.Horii T., Hikawa H., Katsunuma M., Okuzaki H., Synthesis of highly conductive PEDOT: PSS and correlation with hierarchical structure. Polymer 140, 33–38 (2018). [Google Scholar]
  • 21.Jain K., Mehandzhiyski A. Y., Zozoulenko I., Wagberg L., PEDOT: PSS nano-particles in aqueous media: A comparative experimental and molecular dynamics study of particle size, morphology and z-potential. J. Colloid Interface Sci. 584, 57–66 (2021). [DOI] [PubMed] [Google Scholar]
  • 22.Wang H.-Y., et al. , An integrated electrochemical nanodevice for intracellular RNA collection and detection in single living cell. Angew. Chem. Int. Ed. 60, 13244–13250 (2021). [DOI] [PubMed] [Google Scholar]
  • 23.Shi X.-M., et al. , Electrochemical single-cell protein therapeutics using a double barrel nanopipette. Angew. Chem. Int. Ed. 62, e202215801 (2023). [DOI] [PubMed] [Google Scholar]
  • 24.Shi X.-M., et al. , An integrated dual-functional nanotool capable of studying single cell epigenetics and programmable gene regulation. Angew. Chem. Int. Ed. 62, e202302930 (2023). [DOI] [PubMed] [Google Scholar]
  • 25.Kim S. J., Wang Y.-C., Lee J. H., Jang H., Han J., Concentration polarization and nonlinear electrokinetic flow near a nanofluidic channel. Phys. Rev. Lett. 99, 044501 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Lin C.-Y., Combs C., Su Y.-S., Yeh L.-H., Siwy Z. S., Rectification of concentration polarization in mesopores leads to high conductance ionic diodes and high performance osmotic power. J. Am. Chem. Soc. 141, 3691–3698 (2019). [DOI] [PubMed] [Google Scholar]
  • 27.Wang D., et al. , Transmembrane potential across single conical nanopores and resulting memristive and memcapacitive ion transport. J. Am. Chem. Soc. 134, 3651–3654 (2012). [DOI] [PubMed] [Google Scholar]
  • 28.Pershin Y. V., Ventra M. D., Memory effects in complex materials and nanoscale systems. Adv. Phys. 60, 145–227 (2011). [Google Scholar]
  • 29.Hodgkin A. L., Huxley A. F., A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. 117, 500–544 (1952). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Chua L., Sbitnev V., Kim H., Hodgkin–Huxley axon is made of memristors. Int. J. Bifurcat. Chaos 22, 1230011 (2012). [Google Scholar]
  • 31.Sah M. P., Kim H., Chua L., Brains are made of memristors. IEEE Circuits Syst. Mag. 14, 12–36 (2014). [Google Scholar]
  • 32.Zucker R. S., Regehr W. G., Short-term synaptic plasticity. Annu. Rev. Physiol. 64, 355–405 (2002). [DOI] [PubMed] [Google Scholar]
  • 33.Blitz D. M., Foster K. A., Regehr W. G., Short-term synaptic plasticity: A comparison of two synapses. Nat. Rev. Neurosci. 5, 630–640 (2004). [DOI] [PubMed] [Google Scholar]
  • 34.Abbott L. F., Regehr W. G., Synaptic computation. Nature 431, 796–803 (2004). [DOI] [PubMed] [Google Scholar]
  • 35.Abbott L. F., Nelson S. B., Synaptic plasticity: Taming the beast. Nat. Neurosci. 3, 1178–1183 (2000). [DOI] [PubMed] [Google Scholar]
  • 36.Fukushima Y., Single channel potassium currents of the anomalous rectifier. Nature 294, 368–371 (1981). [DOI] [PubMed] [Google Scholar]
  • 37.Bader C. R., Bernheim L., Bertrand D., Sodium-activated potassium current in cultured avian neurons. Nature 317, 540–542 (1985). [DOI] [PubMed] [Google Scholar]
  • 38.Bean B. P., The action potential in mammalian central neurons. Nat. Rev. Neurosci. 8, 451–465 (2007). [DOI] [PubMed] [Google Scholar]
  • 39.Spruston N., Schiller Y., Stuart G., Sakmann B., Activity-dependent action potential invasion and calcium influx into hippocampal CA1 dendrites. Science 268, 297–300 (1995). [DOI] [PubMed] [Google Scholar]
  • 40.DiBartola S. P., Morais H. A. D., Disorders of Potassium: Hypokalemia and Hyperkalemia (American College of Veterinary Internal Medicine, 1995). [Google Scholar]
  • 41.Zeng M., et al. , Dopamine semiquinone radical doped PEDOT: PSS: Enhanced conductivity, work function and performance in organic solar cells. Adv. Energy Mater. 10, 2000743 (2020). [Google Scholar]
  • 42.Pisani A., Bonsi P., Centonze D., Calabresi P., Bernardi G., Activation of D2-like dopamine receptors reduces synaptic inputs to striatal cholinergic interneurons. J. Neurosci. 7, RC69 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Lin Y. S., Gu Q., Lee L.-Y., Activation of dopamine D2-like receptors attenuates pulmonary C-fiber hypersensitivity in rats. Am. J. Resp. Crit. Care 167, 1096–1101 (2003). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix 01 (PDF)

Data Availability Statement

All study data are included in the article and/or SI Appendix.


Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES