Neuromorphic Circuits Research Articles

Analysis of the spike responses in the neuromorphic implementation of the two-compartmental model of hippocampal pyramidal neuron

Designing a low-power neuromorphic circuit for the main units of the nervous system, neurons, plays a crucial role in neural engineering and neuroprosthesis. Due to the CA1 pyramidal neurons’ specific characteristics, these neurons have been a focus of research for several years. The Pinsky-Rinzel (P-R) model is a nonlinear two-compartmental model of the Hippocampal neuron that reveals various modes of firing patterns. In this paper, a low-power neuromorphic circuit for the Hippocampal P-R model is designed at 180 nm TSMC technology. Then, we analyze the recorded spiking responses from the neuroscience point of view. We also quantitatively compare the P-R model simulated in MATLAB with the proposed circuit in the Cadence simulator to show that the obtained responses are in good agreement. For the proposed circuit, the maximum energy consumption per spike is 97pJ/spike in the worst case. The effect of the device mismatch is also studied by performing a 200-point Monte Carlo simulation for different values of input currents. It is demonstrated that the proposed circuit is robust under process variation. Next, considering temporal coding, the ISI (inter-spike interval) and VP-d (Victor-Purpura distance) are extracted from the spike responses collected from the designed circuit. Finally, utilizing machine-learning algorithms, the applied input current to the circuit was accurately identified from the recorded spike trains. In this way, the spike patterns produced by the designed P-R circuit are diverse and consistent to be able to decode the input stimulus using typical machine learning classifiers. This supports the application of the proposed circuit in the hybrid neurocomputing systems and thus it can be integrated easily into large neuromorphic arrays in future works.

Role of ambient temperature in modulation of behavior of vanadium dioxide volatile memristors and oscillators for neuromorphic applications

Volatile memristors are versatile devices whose operating mechanism is based on an abrupt and volatile change of resistivity. This switching between high and low resistance states is at the base of cutting edge technological implementations such as neural/synaptic devices or random number generators. A detailed understanding of this operating mechanisms is essential prerequisite to exploit the full potentiality of volatile memristors. In this respect, multi-physics device simulations provide a powerful tool to single out material properties and device features that are the keys to achieve desired behaviors. In this paper, we perform 3D electrothermal simulations of volatile memristors based on vanadium dioxide (VO_{2}) to accurately investigate the interplay among Joule effect, heat dissipation and the external temperature T_{0} over their resistive switching mechanism. In particular, we extract from our simulations a simplified model for the effect of T_{0} over the negative differential resistance (NDR) region of such devices. The NDR of VO_{2} devices is pivotal for building VO_{2} oscillators, which have been recently shown to be essential elements of oscillatory neural networks (ONNs). ONNs are innovative neuromorphic circuits that harness oscillators’ phases to compute. Our simulations quantify the impact of T_{0} over figures of merit of VO_{2} oscillator, such as frequency, voltage amplitude and average power per cycle. Our findings shed light over the interlinked thermal and electrical behavior of VO_{2} volatile memristors and oscillators, and provide a roadmap for the development of ONN technology.

A large scale photonic matrix processor enabled by charge accumulation

Abstract Integrated neuromorphic photonic circuits aim to power complex artificial neural networks (ANNs) in an energy and time efficient way by exploiting the large bandwidth and the low loss of photonic structures. However, scaling photonic circuits to match the requirements of modern ANNs still remains challenging. In this perspective, we give an overview over the usual sizes of matrices processed in ANNs and compare them with the capability of existing photonic matrix processors. To address shortcomings of existing architectures, we propose a time multiplexed matrix processing scheme which virtually increases the size of a physical photonic crossbar array without requiring any additional electrical post-processing. We investigate the underlying process of time multiplexed incoherent optical accumulation and achieve accumulation accuracy of 98.9% with 1 ns pulses. Assuming state of the art active components and a reasonable crossbar array size, this processor architecture would enable matrix vector multiplications with 16,000 × 64 matrices all optically on an estimated area of 51.2 mm2, while performing more than 110 trillion multiply and accumulate operations per second.

Si-based self-programming neuromorphic integrated circuits for intelligent morphing wings

Unlike artificial intelligent systems based on computers, which must be programmed for specific tasks, the human brain can learn in real-time to create new tactics and adapt to complex, unpredictable environments. Computers embedded in artificial intelligent systems can execute arbitrary inference algorithms capable of outperforming humans at specific tasks. However, without real-time self-programming functionality, they must be preprogrammed by humans and will likely to fail in unpredictable environments beyond their preprogrammed domains. In this work, a Si-based synaptic resistor (synstor) was developed by integrating Al2Ox/TaOy materials to emulate biological synapses. The synstors were characterized, and their operation mechanism based on the charge stored in the oxygen vacancies in the Al2Ox material was simulated and analyzed, to understand the inference, learning, and memory functions of the synstors. A self-programming neuromorphic integrated circuit (SNIC) based on synstors was fabricated to execute inference and learning algorithms concurrently in real-time with an energy efficiency more than six-orders of magnitudes higher than those of standard digital computers. The SNIC dynamically modified its algorithm in a real-time learning process to control a morphing wing, thus successfully improving its lift-to-drag force ratio and recovering the wing from stall in complex aerodynamic environments. The synaptic resistor circuits can potentially circumvent the fundamental limitations of computers, thus providing a platform analogous to neurobiological network with real-time self-programming functionality for artificial intelligent systems.

Analog Tunnel Memory Based on Programmable Metallization for Passive Neuromorphic Circuits.

Although experimental implementations of memristive crossbar arrays have indicated the potential of these networks for in-memory computing, their performance is generally limited by an intrinsic variability on the device level as a result of the stochastic formation of conducting filaments. A tunnel-type memristive device typically exhibits small switching variations, owing to the relatively uniform interface effect. However, the low mobility of oxygen ions and large depolarization field result in slow operation speed and poor retention. Here, we demonstrate a quantum-tunneling memory with Ag-doped percolating systems, which possesses desired characteristics for large-scale artificial neural networks. The percolating layer suppresses the random formation of conductive filaments, and the nonvolatile modulation of the Fowler-Nordheim tunneling current is enabled by the collective movement of active Ag nanocrystals with high mobility and a minimal depolarization field. Such devices simultaneously possess electroforming-free characteristics, record low switching variabilities (temporal and spatial variation down to 1.6 and 2.1%, respectively), nanosecond operation speed, and long data retention (>104 s at 85 °C). Simulations prove that passive arrays with our analog memory of large current-voltage nonlinearity achieve a high write and recognition accuracy. Thus, our discovery of the unique tunnel memory contributes to an important step toward realizing neuromorphic circuits.

Intelligent Control of Seizure-Like Activity in a Memristive Neuromorphic Circuit Based on the Hodgkin–Huxley Model

Memristive neuromorphic systems represent one of the most promising technologies to overcome the current challenges faced by conventional computer systems. They have recently been proposed for a wide variety of applications, such as nonvolatile computer memory, neuroprosthetics, and brain–machine interfaces. However, due to their intrinsically nonlinear characteristics, they present a very complex dynamic behavior, including self-sustained oscillations, seizure-like events, and chaos, which may compromise their use in closed-loop systems. In this work, a novel intelligent controller is proposed to suppress seizure-like events in a memristive circuit based on the Hodgkin–Huxley equations. For this purpose, an adaptive neural network is adopted within a Lyapunov-based nonlinear control scheme to attenuate bursting dynamics in the circuit, while compensating for modeling uncertainties and external disturbances. The boundedness and convergence properties of the proposed control scheme are rigorously proved by means of a Lyapunov-like stability analysis. The obtained results confirm the effectiveness of the proposed intelligent controller, presenting a much improved performance when compared with a conventional nonlinear control scheme.

Electrochemically Tunable Mixed Valence Conduction in Ruthenium Hexacyanoruthenate

Reversible electrochemical doping of ruthenium hexacyanoruthenate, a type of Prussian blue analogue (PBA), enables the on-demand tuning of electronic conductivity by more than four orders of magnitude. Inkjet-printed electrochemical random access memory (ECRAM) devices based on Ru-PBA and lithium- or proton-conducting ionogel electrolytes exhibit excellent switching efficiency and long-term memory retention, important characteristics for analog artificial synapses in neuromorphic circuits. We also demonstrate excellent biocompatibility with live neurons and the use of Ru-PBA ECRAM devices to detect dopamine, promising first steps toward connecting artificial and biological neural networks. In-situ probing of metal-metal charge transfer by UV/Vis/NIR absorption spectroscopy reveals a switching mechanism whereby electrochemically tunable valence mixing between N-coordinated Ru sites controls the carrier concentration and mobility, as independently supported by both Marcus-Hush electron transfer theory and more conventional band structure predictions from DFT. The experimental agreement achieved by both theoretical approaches supports a general mechanistic picture that intramolecular charge transfer reactions, more commonly studied in polynuclear mixed valence small molecules, are central to electronic conductivity in extended coordination frameworks.

A Fully Analytical Current Model for Gate–Source Overlap Tunneling FETs as the Ternary Devices

Gate–source overlap tunneling FETs (GSO-TFETs) as a novel ternary device are very promising in low-power neuromorphic circuits. In this article, an accurate potential model of the ternary GSO-TFETs is presented for the face-tunnel region considering the quasi-mobile charges (QMCs) based on the analysis of the face-tunnel mechanism. Then a potential-based analytical current model is developed for the first time to predict the face-tunnel process and line-tunnel current simultaneously with both the gate and the drain modulations. The modeling results are validated with TCAD simulations and good agreement within a wide biasing range is achieved, which indicates the great potential of this SPICE-friendly model for the tunneling-based ternary device in the commercial IC design fields.

Disentangling ionic and electronic contributions to the switching dynamics of memristive Pr0.7Ca0.3MnO3/Al devices by employing a two-resistor model

Area-dependent memristive devices such as $\mathrm{Al}/{\mathrm{Pr}}_{0.7}{\mathrm{Ca}}_{0.3}{\mathrm{MnO}}_{3}$ (PCMO) stacks are highly interesting candidates for synapses in neuromorphic circuits due to their gradual switching properties, their reduced variability and the possibility to tune the resistance with the device area. However, due to the complexity of the different processes taking place, the electronic and ionic transport in theses devices is so far only poorly understood and physical compact models to simulate their behavior are missing so far. We developed a mathematical description of the dynamics of theses devices based on a simple two-resistor model that reproduces the device behavior very well. Based on x-ray photoelectron spectroscopy and impedance spectroscopy we assign the two resistors to the ${\mathrm{AlO}}_{x}$ layer and a depletion zone at the ${\mathrm{Pr}}_{0.7}{\mathrm{Ca}}_{0.3}{\mathrm{MnO}}_{3}$ layer, respectively. We assign the parameters used within the mathematical model to physical parameters and make use of them in order to explain the dynamics of the switching processes during the SET and RESET process in different voltage regimes. For both poly- and single crystalline PCMO thin film devices, oxygen migration between the ${\mathrm{AlO}}_{x}$ and the PCMO depletion zone is responsible for the resistance change. However, the dynamics differ significantly due to the increased mobility of oxygen vacancies with increasing defect density in the case of the polycrystalline samples. Moreover, we observe volatile subloops in our current-voltage curves, which vanish within millisecond time scale. Based on our two-resistor model and the band diagram derived from spectroscopic measurements, we assign these subloops to the injection of electrons into traps within the ${\mathrm{AlO}}_{x}$ barrier.

Neuromorphic circuit based on the un-supervised learning of biologically inspired spiking neural network for pattern recognition

One of the most sophisticated platforms for hosting intelligent systems is bio-inspired. This study proposes pattern recognition hardware using a biologically inspired Spiking Neural Network (SNN) and the new dimensionality reduction approach. The SNN model is based on real neural networks consisting of spiking neurons linked by excitatory and inhibitory synapses activated by excitatory and inhibitory neurotransmitters. Also, a semi-supervised (un-supervised STDP based learning with supervised weight initialization), spike-based learning strategy based on the learning procedure of the nervous system is used to teach the spiking output layer neurons, albeit that the hardware implementation benefits from a semi-supervised approach. The goal of this research is to accurately categorize patterns in the MNIST and CIFAR10 datasets using an SNN-based hardware platform. Due to the limitations of the latter’s resources, a dimensionality reduction based on principal component analysis (PCA) is proposed to speed up the processing procedure and reduce the hardware implementation cost. The presented pattern recognition platform is implemented using the Xilinx® VIVADO high-level synthesis platform (HLS). Finally, optimization approaches are used to improve the used space, reduce hardware implementation delay, and speed up the design process.

Physical Compact Model for Three‐Terminal SONOS Synaptic Circuit Element

Neuromorphic Circuit Design In article number 2200070, R. Stanley Williams and co-workers present SONOS three-terminal synaptic circuit element, known as Flash memory, which is compact-modeled for neuromorphic applications, mimicking a brain. Similar to most two-terminal non-volatile memories such as RRAMs presenting memristors, the compact model of three-terminal non-volatile memory reveals the memristor on the third terminal. The physical compact model opens up opportunities at the intersection of the manufacturable non-volatile memory and the artificial brain.

Heterosynaptic plasticity in biomembrane memristors controlled by pH

In biology, heterosynaptic plasticity maintains homeostasis in synaptic inputs during associative learning and memory, and initiates long-term changes in synaptic strengths that nonspecifically modulate different synapse types. In bioinspired neuromorphic circuits, heterosynaptic plasticity may be used to extend the functionality of two-terminal, biomimetic memristors. In this article, we explore how changes in the pH of droplet interface bilayer aqueous solutions modulate the memristive responses of a lipid bilayer membrane in the pH range 4.97–7.40. Surprisingly, we did not find conclusive evidence for pH-dependent shifts in the voltage thresholds (V*) needed for alamethicin ion channel formation in the membrane. However, we did observe a clear modulation in the dynamics of pore formation with pH in time-dependent, pulsed voltage experiments. Moreover, at the same voltage, lowering the pH resulted in higher steady-state currents because of increased numbers of conductive peptide ion channels in the membrane. This was due to increased partitioning of alamethicin monomers into the membrane at pH 4.97, which is below the pKa (~5.3–5.7) of carboxylate groups on the glutamate residues of the peptide, making the monomers more hydrophobic. Neutralization of the negative charges on these residues, under acidic conditions, increased the concentration of peptide monomers in the membrane, shifting the equilibrium concentrations of peptide aggregate assemblies in the membrane to favor greater numbers of larger, increasingly more conductive pores. It also increased the relaxation time constants for pore formation and decay, and enhanced short-term facilitation and depression of the switching characteristics of the device. Modulating these thresholds globally and independently of alamethicin concentration and applied voltage will enable the assembly of neuromorphic computational circuitry with enhanced functionality.Impact statementWe describe how to use pH as a modulatory “interneuron” that changes the voltage-dependent memristance of alamethicin ion channels in lipid bilayers by changing the structure and dynamical properties of the bilayer. Having the ability to independently control the threshold levels for pore conduction from voltage or ion channel concentration enables additional levels of programmability in a neuromorphic system. In this article, we note that barriers to conduction from membrane-bound ion channels can be lowered by reducing solution pH, resulting in higher currents, and enhanced short-term learning behavior in the form of paired-pulse facilitation. Tuning threshold values with environmental variables, such as pH, provide additional training and learning algorithms that can be used to elicit complex functionality within spiking neural networks.Graphical abstract

An Electronically Tunable Meminductor Emulator and Its Application in Chaotic Oscillator and Adaptive Learning Circuit

In this paper, a new design of meminductor emulator has been suggested using voltage differencing buffered amplifier (VDBA), current differencing buffered amplifier (CDBA), a resistor and two grounded capacitors. The key advantages of the proposed meminductor emulator circuits are their design simplicity, implementation without memristors and easily adjustable design for grounded/ floating decremental to incremental configurations. The proposed circuits have been designed and simulated using 180-nm process technologies in the Mentor Graphics Eldo simulation tool. The nonvolatility and Monte Carlo analysis of the proposed meminductor emulators have been given. A comparison table has also been presented to discuss the benefits of the suggested meminductor emulator designs with existing designs. The operation of the designed emulator has been corroborated by using it in the chaotic oscillator and neuromorphic circuit applications.

Enabling efficient rate and temporal coding using reliability‐aware design of a neuromorphic circuit

AbstractReliability aspects such as bias temperature instability (BTI) and hot carrier injection (HCI) affecting devices in advanced CMOS‐based technology have been the subject of active research in recent decades. Due to these reliability issues, various digital and analog circuits were investigated for degradation. However, circuit blocks like the neuron circuits of neuromorphic systems are not fully explored. This work is inclined toward examining the collective degradation impact of BTI and HCI due to aging in an adaptive exponential “integrate and fire” (I&F) model‐based, neuromorphic neuron circuit. Detailed degradation analysis of the stimulated neuron circuit aided in identifying possible mismatches/faults associated with neuron spikes. These factors could reduce the efficiency of the neuronal circuit by potentially affecting the transmission of information in a neuromorphic system. Various performance parameters were then derived to quantify the extent of circuit deterioration. The proposed reliability‐aware design aims to improve the circuit degradation through its effectiveness in alleviating the overall reliability impact. It demonstrates enhanced circuit operation in spike generation even after aging. The circuit performance is validated through simulations at “Time0” (pre‐degradation) and “Aged” (post‐degradation) neuron netlists, which is then compared with the proposed reliability‐aware circuit.

Polymer-based electronics that can learn to drive: That’s smart

In a recent paper published in <i>Science Advances</i> by Krauhausen and co-authors,¹ the Gkoupidenis' research group from the Max Planck Institute for Polymer Research demonstrated the use of organic neuromorphic electronics to drive a small robot through a maze. Using organic electrochemical transistors (OECTs), the authors designed a neuromorphic circuit and demonstrated energy efficient and reliable learning capabilities enabled by the mixed ionic-electronic conduction of the organic semiconductor. The work demonstrates unprecedented capabilities in organic neuromorphics toward effective local and decentralized learning, which is attractive for smart robotics.

Polydopamine‐Based All Solid‐State Flexible Organic Neuromorphic Devices for Access Device‐Free Artificial Neural Networks

AbstractRecent developments in organic neuromorphic devices and biohybrid interfaces are promising examples that show potential to improve implantable devices toward organic adaptive brain‐machine interfaces. However, fully integrated neuromorphic arrays still require relatively complex circuitry that includes multiple access devices to ensure synaptic weight stability and prevent sneak paths. Here, it is shown that polydopamine (PDA), the byproduct of dopamine autoxidation, promotes proton conductivity and can serve as a solid‐state electrolyte. Slow kinetics and high energy barriers of the PDA solid electrolyte prevent loss of conductance state for the device with a three‐terminal configuration without an access device, while partial dedoping of the conductive polymer channel by PDA simultaneously increases its stability in ambient environments. Fabricating the neuromorphic device on a flexible poly(styrene‐block‐isobutylene‐block‐styrene) substrate and the inherent biocompatibility of PDA demonstrates its potential toward more sophisticated implantable neuromorphic circuits for advanced neuroprosthetics.

Real-Time Generation of Hyperbolic Neuronal Spiking Patterns.

Neuronal spikes are referred to as the electric activity of neurons (in terms of voltage) in response to various biological events such as the sodium and calcium ionic current channels in the brain. Currently, both biological models as well as mathematical models of neuronal spiking patterns have been introduced in the literature. However, very little attempt has been made to run these models in real-time. With applications ranging from hardware neuromorphic circuit designs, artificial intelligence (AI) architectures, to deep brain stimulation, real-time generation of these models is of particular interest in the brain-inspired computing/architecture and neuro-modulation/stimulation research communities. This paper proposes the development of a framework for generating the hyperbolic based single neuronal spiking patterns in real-time. Simulation results confirm that the generated spikes resemble the existing models of neuronal spiking patterns, with additional real-time run capability as well as the ability to change the parameters on the fly. Clinical relevance-Real-time models of neuronal spiking patterns have significant clinical relevance with respect to applications of neuromorphic/AI chips for medical image processing/computer vision, as well as clinical neuroscience, neuromodulation and neurostimulation such as deep brain stimulation for modulating the abnormal effects of neurological diseases.

Neuromorphic solutions: digital implementation of bio-inspired spiking neural network for electrocardiogram classification

Conventional techniques of off-chip processing for wearable devices cause high hardware resource usage which leads to heat generation and increased power consumption. &lt;span&gt;Hence, edge computing methods such as neuromorphic computing are considered the most promising modern technology to replace conventional processing. It is beneficial to employ neuromorphic processing in electrocardiogram (ECG) classification, enabling engineers to overcome the constraints of heat generation caused by hardware utilization. Thus, this work aims to investigate common building blocks in a spiking neural network (SNN), analyze the spike-based plasticity mechanism and implement ECG classification on a neuromorphic circuit. The MIT-BIH Arrhythmia database (MITDB) is preprocessed in MATLAB, then used to train and test an SNN designed for field programmable gate arrays (FPGA), employing spike-based plasticity and Izhikevich neurons. The behaviour of spike timing dependent plasticity (STDP) in a neuromorphic circuit is also visualized in this work. The state-of the-art performance of this work lies in providing a generic mechanism to adapt ECG classification into a neuromorphic solution, a non-Von Neumann architecture. The proposed digital design utilizes 1.058% of hardware resources on a Zedboard. Application-wise, this work provides a foundation for development of neuromorphic computing in wearable medical devices that perform continuous monitoring of ECG&lt;/span&gt;.

IEEE Journal on Flexible Electronics Call for Papers: Special Issue on Neuromorphic Devices and Circuits for Next Generation Flexible Electronics

IEEE Journal on Flexible Electronics Call for Papers: Special Issue on Neuromorphic Devices and Circuits for Next Generation Flexible Electronics

Neuro-inspired electronic skin for robots.

Touch is a complex sensing modality owing to large number of receptors (mechano, thermal, pain) nonuniformly embedded in the soft skin all over the body. These receptors can gather and encode the large tactile data, allowing us to feel and perceive the real world. This efficient somatosensation far outperforms the touch-sensing capability of most of the state-of-the-art robots today and suggests the need for neural-like hardware for electronic skin (e-skin). This could be attained through either innovative schemes for developing distributed electronics or repurposing the neuromorphic circuits developed for other sensory modalities such as vision and audio. This Review highlights the hardware implementations of various computational building blocks for e-skin and the ways they can be integrated to potentially realize human skin-like or peripheral nervous system-like functionalities. The neural-like sensing and data processing are discussed along with various algorithms and hardware architectures. The integration of ultrathin neuromorphic chips for local computation and the printed electronics on soft substrate used for the development of e-skin over large areas are expected to advance robotic interaction as well as open new avenues for research in medical instrumentation, wearables, electronics, and neuroprosthetics.