Elsevier

Nano Energy

Volume 86, August 2021, 106038
Nano Energy

Multiplexed neurotransmission emulated for emotion control

https://doi.org/10.1016/j.nanoen.2021.106038Get rights and content

Highlights

  • An artificial synapse showed strong n-type/weak p-type bipolar characteristics for the synergy of two neurotransmitters.

  • This device presented a switchable plasticity between short-term of several seconds and long-term of hours to days.

  • The device achieved so far the longest memory duration of ion-regulated neuromorphic transistor up to 105 s.

  • By introducing an additional monitoring terminal, our device implemented main functions of automatic emotion regulation (AER).

Abstract

Emotions are closely related to multiplexed neurotransmission of glutamate and acetylcholine in the hippocampus, which are competitors and partners respectively. This paper demonstrates the first hetero-junction two-dimensional dual-excitatory artificial synapse (AS) for mimicking co-release of the two different neurotransmitters. This AS for the first time presents an immediate switchable plasticity between short-term of < 10 s and long-term of > 105 s, as both avoidance tendency of negative stimuli and lasting memory of positive emotion. 105 s represents so far the best long term memory of ion-regulated ASs. Inclusion of a fourth terminal as a monitor enables the emulation of automatic emotion regulation (AER). The AS array-based neural network achieves up to ~97% recognition accuracy for emotion-correlated electroencephalogram patterns; a new concept of early warning of negative emotions is proposed for monitoring mood changes, especially useful for babies and patients with mental disease, paving the way to future bio-inspired intelligent systems.

Introduction

Emotion is an attitude experience of human beings towards objective things, and the corresponding behavior response [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]. Emotion has a binary nature. People generally feel anxious in an uncertain circumstance, but pleasant in a comfortable environment [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]. Automatic emotion regulation (AER) helps individuals to process their emotions [1]. Under different external stimuli, multiple unconscious regulation forms begin with activation of sensory receptors in the dorsomedial prefrontal cortex, anterior cingulate cortex and orbitofrontal cortex [2]. Forms of AER mainly include automatic cognitive change [3], [4], [5], [6], automatic attentional control [7], [8], and automatic behavior control [9], [10].

The ability of our brain to select an appropriate AER strategy and subsequent response to a complex external environment requires multiplexed neurotransmission of different neurotransmitters, i.e. co-release and co-transmission of glutamate and acetylcholine from distinct microdomain of the same axon, has recently been found to underlie deserve functions of ventral tegmental area (VTA) region in striatal interneurons or medial habenula neurons [11], [12], [13], [14], [15], [16]. Different combinations of neurotransmitters could be released from various subgroups of VTA neurons to convey different information in a neural circuit [15], [16]. Multiplexed neurotransmission underlies many important synaptic functions [17]. both neurotransmitters are passed through biological synapses in the same type of nerve cell, and show a mutual restriction relationship: glutamate promotes the activation of emotion, whereas acetylcholine calms it [11], [12], [13], [14].

These emotional signals, often in the form of memory of external stimuli, can empower us to react appropriately to adapt to changing conditions in the real world [18]. Endowing human-inspired or human-integrated robots with such automatic emotion-regulated ability could greatly extend their adaptability and cognitive ability [19]. As emotional partners, they can realize emotion monitoring, early warning and psychotherapy when people are in a complex environment or in a bad psychological state. Achieving artificial synapses that emulate the complex functions of biological ones are the key to construct a comparable artificial system capable of automatic regulation of emotions.

Due to the compatibility with complementary logic circuits and multiple input terminals for obtaining signals from a diversity of sensors or actuators, transistor-structured ASs are widely applied to analog biomimetic systems. Some previously-reported transistor-structured artificial synapses (ASs) realized important synaptic functions in a single device; examples include short-term plasticity and long-term plasticity [20], [21], [22], [23]. Despite of these achievements, a major deficiency is that these synaptic transistors exhibit a limited retention time (seconds to minutes) and ambiguous short-term/long-term signals due to the non-switchable conductivity change based on single type of charge carrier that cannot be reset and switched immediately as needed. Especially for long-term plasticity, stimuli with either high frequency or large voltage amplitude are required, and this requirement impedes information consolidation and memory in applications of neuromorphic electronics. Another concern is that many ASs can only simulate the release of only one type of neurotransmitter from a pre-synaptic membrane [20], [21], [22], [23]. However, synergistic responses to a complex external environment need co-release of more than one type of neurotransmitter [24], [25]. Inspired by such properties, we proposed a dual-excitatory AS that can simulate emotion-regulated function by altering the polarity of charge carriers.

In this work, a dual-excitatory AS based on two-dimensional (2D) graphene/hexagonal boron nitride (h-BN) heterojunction is demonstrated by using a four-terminal device structure. Due to the strong n-type and weak p-type bipolar characteristics in the two-dimensional heterojunction structure, the effects of glutamate and acetylcholine in a multiplexed transmission process are simulated. With a fourth terminal, our AS realizes real-time monitoring of external stimuli, and simulates emotion-regulation process. Plasticity modes of the AS can be immediately switched between short-term of several seconds and long-term of hours to days, as both of avoidance tendency and lasting memory required. The device demonstrates so far the longest memory duration of ASs up to 105 s, and excellent endurance of > 103 cycles. This graphene/h-BN artificial synapse (GHAS) may improve cognitive neuromorphic computing, and could facilitate creation of human-computer interface. With the conventional stochastic gradient ascent/descent learning algorithm in our GHAS array constructed neural network, recognition accuracies are up to 97% for emotion-correlated electroencephalogram (EEG) patterns. Furthermore, the feature extraction of warning features in different moods is firstly introduced into the pattern recognition to realize a early warning process of negative emotions, which is important for real-time monitoring babies and patients diagnosed with depression, anxiety and bipolar disorder. Therefore, this work may contribute to the development of next-generation artificial intelligence and robotics.

Section snippets

Device structure and material characteristics

The proposed GHAS could mimic the competition of two kinds of excitatory neurotransmitters (glutamate and acetylcholine) in a biological synapse (Fig. 1). This GHAS device is composed of an ion gel, two metal contact pads, and a graphene/h-BN heterojunction layer (GHHL) on a p + Si/SiO2 substrate. The ion gel mimics the synaptic cleft; when pre-synaptic spikes are applied to the ion gel top gate, the single-layered graphene channel capped by the multi-layered h-BN can induce carriers (electrons

Conclusion

In summary, a dual-excitatory AS based on a 2D heterojunction was fabricated to emulate the cooperation and competition between two kinds of excitatory neurotransmitters in the same synaptic gap. Due to the unique strong n-type and weak p-type bipolar characteristics, our GHAS achieved switchable synaptic plasticity, with three typical behavioral regulation modes: potentiation-depression regulation, potentiation-erasure regulation, and STP-LTP immediate switching between short-term of several

Fabrication of the devices

A few-layer h-BN film that had been peeled off and transferred to P+ Si/300 nm SiO2/single-layer graphene was purchased from Sixcarbon Technology Shenzhen. The source and drain Au electrodes (60 nm) were thermally deposited through a rectangular shadow mask (width 1000 µm; length 100 µm) onto the h-BN films. Then the ion-gel top-gate dielectric (1:3 mass ratio between polymer Poly(vinylidenefluoride-co-hexafluoropropylene) [PVDF-HFP] and ionic liquid (1-ethyl-3-methylimidazolium

CRediT authorship contribution statement

Yao Ni: Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Writing - original draft. Mingxue Ma: Conceptualization, Data curation, Investigation, Supervision, Visualization, Writing - original draft. Huanhuan Wei: Data curation, Investigation. Jiangdong Gong: Methodology, Validation. Hong Han: Supervision, Validation. Lu Liu: Validation. Zhipeng Xu: Validation. Wentao Xu: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision,

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by Guangdong Key Research and Development Project No. 2018B030338001, the Tianjin Science Foundation for Distinguished Young Scholars (19JCJQJC61000), the Fundamental Research Funds for the Central Universities (075–63191740, 075–63191745, 075–92022027), Hundred Young Academic Leaders Program of Nankai University (2122018218), Natural Science Foundation of Tianjin (18JCYBJC16000), the 111 Project (B16027), the International Cooperation Base (2016D01025), and Tianjin

Yao Ni obtained his Master’s degree at College of Microelectronics and Communication Engineering, Chongqing University in 2019. He is currently a Ph.D. candidate at the Institute of Photoelectronic Thin Film Devices and Technology, Nankai University. His research project is on transistor structured memories, artificial synapse devices, and flexible electronics.

References (57)

  • Y. Chen et al.

    Graphene-ferroelectric transistors as complementary synapses for supervised learning in spiking neural network

    npj 2D Mater. Appl.

    (2019)
  • Y. Liu et al.

    Self-powered artificial synapses actuated by triboelectric nanogenerator

    Nano Energy

    (2019)
  • G.M. Fitzsimons et al.

    Automatic self-regulation

    Handb. Self-Regul.

    (2004)
  • M.L. Phillips et al.

    Neural systems underlying voluntary and automatic emotion regulation: toward a neural model of bipolar disorder

    Mol. Psychiatry

    (2008)
  • J.A. Gray et al.

    The Neuropsychology of Anxiety: An Enquiry into Functions of the Septo-Hippocampal System

    (2000)
  • U. Schroeder et al.

    Functional neuroanatomy of perceiving surprised faces

    Hum. Brain Map.

    (2004)
  • D.C. Jackson et al.

    Now you feel it, now you don’t: frontal brain electrical asymmetry and individual differences in emotion regulation

    Psychol. Sci.

    (2003)
  • C. Gras et al.

    The vesicular glutamate transporter VGLUT3 synergizes striatal acetylcholine tone

    Nat. Neurosci.

    (2008)
  • M.J. Higley et al.

    Cholinergic interneurons mediate fast VGluT3-dependent glutamatergic transmission in the striatum

    PLoS One

    (2011)
  • G. Feng et al.

    A Sub‐10 nm vertical organic/inorganic hybrid transistor for pain‐perceptual and sensitization‐regulated nociceptor emulation

    Adv. Mater.

    (2020)
  • C. Wan et al.

    An artificial sensory neuron with tactile perceptual learning

    Adv. Mater.

    (2018)
  • P. Gkoupidenis et al.

    Neuromorphic functions in PEDOT: PSS organic electrochemical transistors

    Adv. Mater.

    (2015)
  • T. Ohno et al.

    Short-term plasticity and long-term potentiation mimicked in single inorganic synapses

    Nat. Mater.

    (2011)
  • W. Xu et al.

    Organometal halide perovskite artificial synapses

    Adv. Mater.

    (2016)
  • Y. v d Burgt et al.

    A non-volatile organic electrochemical device as a low-voltage artificial synapse for neuromorphic computing

    Nat. Mater.

    (2017)
  • D. Sulzer et al.

    Dopamine neurons make glutamatergic synapses in vitro

    J. Neurosci.

    (1998)
  • K.Y. Tseng et al.

    Dopamine–glutamate interactions controlling prefrontal cortical pyramidal cell excitability involve multiple signaling mechanisms

    J. Neurosci.

    (2004)
  • X. Li et al.

    Transfer of large-area graphene films for high-performance transparent conductive electrodes

    Nano Lett.

    (2009)
  • Cited by (19)

    • A fibrous neuromorphic device for multi-level nerve pathways implementing knee jerk reflex and cognitive activities

      2022, Nano Energy
      Citation Excerpt :

      Resent biomimetic studies have attempted to develop an organically operating system based on artificial synaptic devices that can emulate the functions of the biological neural system [13,14]. However, most artificial synapses emulate the transmission of a single type of neurotransmitter, multiplexed neurotransmission in fibrous devices has not yet been reported [15–17]. The development of fibrous artificial synapse that emulates the release and coordinative effect of different neurotransmitters is highly desired to further extend their application in future bio-inspired electronics [18–22].

    • Tactile tribotronic reconfigurable p-n junctions for artificial synapses

      2022, Science Bulletin
      Citation Excerpt :

      Three-terminal transistors [11–13], by contrast, are capable of effectively simulating the parallel signal processing because of separated read (drain) and write (gate) terminals, which offers more feasibility and flexibility for the sophisticated requirements. By applying a fourth terminal as a real-time monitor, synaptic transistor can even achieve avoidance tendency of negative stimuli and lasting memory of positive emotion, thereby emulating automatic emotion regulation [14]. It is also necessary to exploit the synaptic transistor to more complex neuromorphic hardware systems with high recognition accuracy and computing efficiency [15–17].

    • Mimicking ion-balance-dependent synaptic plasticity in body fluid for adaptive environment-responsive artificial neuromuscular reflexes

      2022, Materials Today Nano
      Citation Excerpt :

      However, the realization of the above functions requires an artificial body-fluid-dependent neuristor (BFDN) in which learning and neuromorphic recognition abilities vary along with the change in inner ion balance. State-of-the-art artificial synapses successfully mimic important forms of short-term and long-term plasticity (LTP) [15–17]. However, a non-adjustable internal environment with a single kind of cation impedes the ability to emulate electrical synaptic functions under complex body-fluid ionic balancing conditions that simultaneously respond to stimuli from the external environment and represent the internal environment of a body.

    • A neuromorphic device mimicking synaptic plasticity under different body fluid K<sup>+</sup> homeostasis for artificial reflex path construction and pattern recognition

      2022, Fundamental Research
      Citation Excerpt :

      In biology, different neural units and network structures afford complex and diverse brain functions. Inspired by this phenomenon, neuromorphic electronic devices have been extensively developed for mimicking real biological nerves [13–23]. In artificial neural devices, emphasis has always been placed on the principle simulation [13–15], structural design [16,17], storage calculation [18,19], and applications in the artificial neural system [20–23].

    View all citing articles on Scopus

    Yao Ni obtained his Master’s degree at College of Microelectronics and Communication Engineering, Chongqing University in 2019. He is currently a Ph.D. candidate at the Institute of Photoelectronic Thin Film Devices and Technology, Nankai University. His research project is on transistor structured memories, artificial synapse devices, and flexible electronics.

    Mingxue Ma received her B.S. degree from the College of Chemistry at Nankai University (2018). She is an M.S. candidate at the College of Electronic Information and Optical Engineering, Nankai University. Her major research focus is on three-terminal artificial synaptic devices

    Huanhuan Wei received his Master’s Degree from the College of Environmental and Chemical Engineering at Shanghai University of Electric Power (2018). He is currently a Ph.D. candidate at the Institute of Photoelectronic Thin Film Devices and Technology, Nankai University. His research activity is focused on the study of artificial synaptic devices and flexible electronics.

    Jiangdong Gong received his B.S. degree from the College of Physics and Electronics at Henan University in 2017. He is an Ph.D. candidate at the College of Electronic Information and Optical Engineering, Nankai University. His major research focuses on halide-perovskite-based artificial bionic devices.

    Hong Han received her B.S. degree from the College of Chemistry at Nankai University (2018). She is an M.S. candidate at the College of Electronic Information and Optical Engineering, Nankai University. Her major research focuses on three-terminal artificial synaptic devices.

    Lu Liu obtained his M.S. Materials Science and Engineering, University of Jinan (2018). She is currently a Ph.D. candidate at the Institute of Optoelectronic Thin Film Devices and Technology, Nankai University. Her research interests are the fabrication and developments of neuromorphic electronics, flexible electronics and electrohydrodynamic nanowire (e-NW) printing.

    Zhipeng Xu received his B.S. degree from the East China University of Technology (2019). He is currently an M.S. candidate at the College of Electronic Information and Optical Engineering, Nankai University. His major research focuses on three-terminal artificial synaptic devices.

    Wentao Xu is a professor in the Institute of Photoelectronic Thin Film Devices and Technology of Nankai University. He received his B. S. at Beijing Normal University and his Ph.D. at the Pohang University of Science and Technology (POSTECH). He had been a research associate professor at Seoul National University (SNU), and visiting scholar at Stanford University and the University of Illinois at Urbana-Champaign. His research interests include neuromorphic electronic devices, flexible electronics, electrohydrodynamic nanowire printing, memory devices, and thin-film transistors.

    View full text