Neuron Model Research Articles

Towards NeuroAI: introducing neuronal diversity into artificial neural networks

Throughout history, the development of artificial intelligence, especially artificial neural networks, has been continuously influenced by a deeper understanding of the brain. This influence includes the development of the neocognitron, considered a precursor to convolutional neural networks. The emerging field of NeuroAI posits that leveraging neuroscience knowledge could significantly advance AI by imbuing networks with enhanced capabilities. Unlike the human brain, which features a variety of morphologically and functionally distinct neurons, artificial neural networks typically rely on a homogeneous neuron model. In the human brain, the diversity of neurons facilitates a wide range of intelligent behaviors. Given that artificial networks aim to mimic the human brain, incorporating a diversity of neuron models could address key challenges in artificial intelligence, such as efficiency, interpretability, and memory capacity. This perspective begins by examining the basics of biological neuronal diversity and how biological neurons transmit and process information. We then explore research efforts to design novel neuron models for artificial networks and discuss the potential benefits of neuronal diversity, including applications in several critical areas. Finally, we address the challenges and future directions for integrating neuronal diversity into artificial networks, highlighting its potential to enrich NeuroAI.

Chondroitin Sulfate and Proteinoids in Neuron Models.

This study examines the relationship between chondroitin sulfate, proteinoids, and computational neuron models, with a specific emphasis on the Izhikevich neuron model. We investigate the effect of chondroitin sulfate-proteinoid complexes on the behavior and dynamics of simulated neurons. Through the use of computational simulations, we provide evidence that these biomolecular components have the power to regulate the responsiveness of neurons, the patterns of their firing, and the ability of their synapses to change within the Izhikevich architecture. The findings suggest that the interactions between chondroitin sulfate and proteinoid cause notable alterations in the dynamics of membrane potential and the timing of spikes. We detect adjustments in the features of neuronal responses, such as shifts in the thresholds for firing, alterations in spike frequency adaptation, and changes to bursting patterns. The findings indicate that chondroitin sulfate and proteinoids may have a role in precisely adjusting neuronal information processing and network behavior. This study offers valuable information about the complex connection between the many components of the extracellular matrix, protein-based structures, and the functioning of neurons. In addition, our analysis of the proteinoid-chondroitine system using game theory uncovers a significant Prisoner's Dilemma scenario. The system's inclination toward defection, due to the appeal of cheating and the significant penalty for cooperation, with a mean voltage of -9.19 mV, indicates that defective behaviors may prevail in the long term dynamics of these neuronal interactions.

Nonlinearities and timescales in neural models of temporal interference stimulation.

In temporal interference (TI) stimulation, neuronal cells react to two interfering sinusoidal electric fields with a slightly different frequency ( , in the range of about 1-4 kHz, in the range of about 1-100 Hz). It has been previously observed that for the same input intensity, the neurons do not react to a purely sinusoidal signal at or . This study seeks a better understanding of the largely unknown mechanisms underlying TI neuromodulation. To this end, single-compartment models are used to simulate computationally the response of neurons to the sinusoidal and TI waveform. This study compares five different neuron models: Hodgkin-Huxley (HH), Frankenhaeuser-Huxley (FH), along with leaky, exponential, and adaptive-exponential integrate-and-fire (IF). It was found that IF models do not entirely reflect the experimental behavior while the HH and FH model did qualitatively replicate the observed neural responses. Changing the time constants and steady state values of the ion gates in the FH model alters the response to both the sinusoidal and TI signal, possibly reducing the firing threshold of the sinusoidal input below that of the TI input. The results show that in the modified (simplified) model, TI stimulation is not qualitatively impacted by nonlinearities in the current-voltage relation. In contrast, ion channels have a significant impact on the neuronal response. This paper offers insights into neuronal biophysics and computational models of TI stimulation.

Neuronal TRPV1-CGRP axis regulates peripheral nerve regeneration through ERK/HIF-1 signaling pathway.

Severe trauma frequently leads to nerve damage. Peripheral nerves possess a degree of regenerative ability, and actively promoting their recovery can help restore the sensory and functional capacities of tissues. The neuropeptide calcitonin gene-related peptide (CGRP) is believed to regulate the repair of injured peripheral nerves, with neuronal transient receptor potential vanilloid type 1 (TRPV1) potentially serving as a crucial upstream factor. In this study, we established a mouse model of sciatic nerve (SN) crush injury and found that intrathecal injection of capsaicin (Cap) activated the neuronal TRPV1-CGRP axis, thereby promoting SN repair. Conversely, the application of capsazepine (Cpz), which inhibits the neuronal TRPV1-CGRP axis, delayed SN repair. Local restoration of CGRP expression at the injury site enhanced the repair process. Invitro experiments, we employed the rat Schwann cell (SC) line RSC96 to establish an indirect co-culture model of neurons and SCs. We observed that the proliferation, migration, expression of myelination-associated proteins, and neurotrophic secretion functions of RSC96 cells are positively correlated with the degree of activation of neuronal TRPV1. Inhibition of neuronal TRPV1, followed by the restoration of CGRP levels, improved these functions in RSC96 cells. Furthermore, activation of the neuronal TRPV1-CGRP axis resulted in an upregulation of extracellular signal-regulated kinases 1/2 (ERK1/2) phosphorylation levels and an increase in hypoxia-inducible factor 1α (HIF-1α) accumulation in RSC96 cells, thereby promoting their proliferation and migration. In summary, this study demonstrates that neuronal TRPV1-CGRP axis can regulate biological behavior of SCs and axon regeneration by activating the ERK/HIF-1 signaling pathway following peripheral nerve injury. This finding clarifies the role of CGRP in neuroregulatory networks and provides a novel reference point for the development of drugs and biomaterials for treating nerve damage.

Power spectrum based early warning signal of neuronal firing

Brain diseases often occur simultaneously with critical changes in neural system and abnormal neuronal firing. Studying the early warning signals (EWSs) of critical changes can provide a promising approach for predicting neuronal firing behaviors, which is conducible to the early diagnosis and prevention of brain diseases. Traditional EWSs, such as autocorrelation and variance, have been widely used to detect the critical transitions in various dynamical systems. However, these methods have limitations in distinguishing different types of bifurcations. In contrast, the EWSs with power spectrum have shown a significant advantage in not only predicting bifurcation points but also distinguishing the types of bifurcations involved. Previous studies have demonstrated its predictive capability in climate and ecological models. Based on this, this study applies the EWS with power spectrum to neuronal systems in order to predict the neuronal firing behaviors and distinguish different classes of neuronal excitability. Specifically, we compute the EWSs before the occurrence of saddle-node bifurcation on the invariant circle and subcritical Hopf bifurcation in the Morris-Lecar neuron model. Additionally, we extend the analysis to the Hindmarsh-Rose model, calculating the EWSs before both saddle-node bifurcation and supercritical Hopf bifurcation. This study contains the four types of codimension-1 bifurcations corresponding to the neuronal firing. For comparison, we also calculate two types of conventional EWSs: lag-1 autocorrelation and variance. In numerical simulations, the stochastic differential equations are simulated by the Euler-Maruyama method. Then, the simulated responses are detrended by the Lowess filter. Finally, the EWSs are calculated by using the rolling window method to ensure the detection of EWS before bifurcation points. Our results show that the EWS with power spectrum can effectively predict the bifurcation points, which means that it can predict neuronal firing activities. Compared with the lag-1 autocorrelation and the variance, the EWSs with power spectrum not only accurately predict the neuronal firing, but also distinguish the classes of excitability in neurons. That is, according to the different characteristics of the power spectrum frequencies, the EWS with power spectrum can effectively distinguish between saddle-node bifurcations and Hopf bifurcations during neuronal firing. This work provides a novel approach for predicting the critical transitions in neural system, with potential applications in diagnosing and treating brain diseases.

A robust Parkinson’s disease detection model based on time-varying synaptic efficacy function in spiking neural network

Parkinson’s disease (PD) is a neurodegenerative disease affecting millions of people around the world. Conventional PD detection algorithms are generally based on first and second-generation artificial neural network (ANN) models which consume high energy and have complex architecture. Considering these limitations, a time-varying synaptic efficacy function based leaky-integrate and fire neuron model, called SEFRON is used for the detection of PD. SEFRON explores the advantages of Spiking Neural Network (SNN) which is suitable for neuromorphic devices. To evaluate the performance of SEFRON, 2 publicly available standard datasets, namely (1) UCI: Oxford Parkinson’s Disease Detection Dataset and (2) UCI: Parkinson Dataset with replicated acoustic features are used. The performance is compared with other well-known neural network models: Multilayer Perceptron Neural Network (MLP-NN), Radial Basis Function Neural Network (RBF-NN), Recurrent Neural Network (RNN) and Long short-term memory (LSTM). The experimental results demonstrate that the SEFRON classifier achieves a maximum accuracy of 100% and an average accuracy of 99.49% on dataset 1. For dataset 2, it attains a peak accuracy of 94% and an average accuracy of 91.94%, outperforming the other classifiers in both cases. From the performance, it is proved that the presented model can help to develop a robust automated PD detection device that can assist the physicians to diagnose the disease at its early stage.

Simulation Study of Envelope Wave Electrical Nerve Stimulation Based on a Real Head Model.

In recent years, the modulation of brain neural activity by applied electromagnetic fields has become a hot spot in neuroscience research. Transcranial direct current stimulation (tDCS) and transcranial alternating current stimulation (tACS) are two common non-invasive neuromodulation techniques. However, conventional tACS has limited stimulation effects in the deeper parts of the brain. In this study, a method of low and medium frequency envelope wave neurostimulation is proposed, and its effectiveness and safety are evaluated by simulation and human experiment. First, we built a real head model from head MRI image data and used the finite element method to calculate the current distribution of the envelope wave in the brain. Then, a single-compartment neuron model was constructed in NEURON software to simulate the action potential generation of neurons under different frequencies of electrical stimulation. Finally, a human experiment was conducted to investigate the threshold of human perception of envelope wave electrical stimulation. The results show that envelope wave can both increase the depth of stimulation and induce neurons to generate effective action potentials. In envelope wave electrical stimulation, the optimal modulating wave frequency was 50 Hz, and the carrier frequency was 2 kHz-3 kHz. This method is expected to play an important role in the non-invasive treatment of neurological and psychiatric disorders.

Neurobiological transition of magnetized and demagnetized dynamism for fractional Hindmarsh–Rose neuron model via fractal numerical simulations

Neurobiological transition of magnetized and demagnetized dynamism for fractional Hindmarsh–Rose neuron model via fractal numerical simulations

Nonautonomous Single Inertial Neuron Model: Coexisting Patterns, Hamilton Energy, and Analog Circuit Implementation

The main objective of this study is to investigate the chaotic dynamics and analog circuit implementation of a nonautonomous single inertial neuron. First, the dimensionless mathematical model of such neuron is established. The equilibria with three kinds of stabilities are then depicted, and system symmetry and Hamilton energy function are analyzed, respectively. In numerical simulations, by using the two-dimensional/one-dimensional bifurcation diagrams, Lyapunov exponent spectra, phase plots, and basins of attraction, bifurcation of coexisting bubbles as well as bi-stable patterns are revealed, where the generated coexisting attractors are symmetric about the origin. In addition, we confirm that the complexity of the basins of attraction deteriorates with the increase of the frequency parameter. Finally, analog circuit experiments with few analog circuit elements are deployed, via which the bi-stable patterns of the model are well verified.

SLC35A2 loss of function variants affect glycomic signatures, neuronal fate, and network dynamics.

SLC35A2 encodes a UDP-galactose transporter essential for glycosylation of proteins and galactosylation of lipids and glycosaminoglycans. Germline genetic SLC35A2 variants have been identified in congenital disorders of glycosylation and somatic SLC35A2 variants have been linked to intractable epilepsy associated with malformations of cortical development. However, the functional consequences of these pathogenic variants on brain development and network integrity remain elusive. In this study, we use an isogenic human induced pluripotent stem cell-derived neuron model to comprehensively interrogate the functional impact of loss of function variants in SLC35A2 through the integration of cellular and molecular biology, protein glycosylation analysis, neural network dynamics, and single cell electrophysiology. We show that loss of function variants in SLC35A2 result in disrupted glycomic signatures and precocious neurodevelopment, yielding hypoactive, asynchronous neural networks. This aberrant network activity is attributed to an inhibitory/excitatory imbalance as characterization of neural composition revealed preferential differentiation of SLC35A2 loss of function variants towards the GABAergic fate. Additionally, electrophysiological recordings of synaptic activity reveal a shift in excitatory/inhibitory balance towards increased inhibitory drive, indicating changes occurring specifically at the pre-synaptic terminal. Our study is the first to provide mechanistic insight regarding the early development and functional connectivity of SLC35A2 loss of function variant harboring human neurons, providing important groundwork for future exploration of potential therapeutic interventions.

Analog Spiking Neuron Model for Unsupervised STDP-based learning in Neuromorphic Circuits

Analog Spiking Neuron Model for Unsupervised STDP-based learning in Neuromorphic Circuits

CHMP2B promotes CHMP7 mediated nuclear pore complex injury in sporadic ALS

Alterations to the composition and function of neuronal nuclear pore complexes (NPCs) have been documented in multiple neurodegenerative diseases including Amyotrophic Lateral Sclerosis (ALS). Moreover, recent work has suggested that injury to the NPC can at least in part contribute to TDP-43 loss of function and mislocalization, a pathological hallmark of ALS and related neurodegenerative diseases. Collectively, these studies highlight a role for disruptions in NPC homeostasis and surveillance as a significant pathophysiologic event in neurodegeneration. The ESCRT-III nuclear surveillance pathway plays a critical role in the surveillance and maintenance of NPCs and the surrounding nuclear environment. Importantly, pathologic alterations to this pathway and its protein constituents have been implicated in neurodegenerative diseases such as ALS. However, the mechanism by which this pathway contributes to disease associated alterations in the NPC remains unknown. Here we use an induced pluripotent stem cell (iPSC) derived neuron (iPSN) model of sALS to demonstrate that CHMP7/ESCRT-III nuclear maintenance/surveillance is overactivated in sALS neurons. This overactivation is dependent upon the ESCRT-III protein CHMP2B and sustained CHMP2B dependent “activation” is sufficient to contribute to pathologic CHMP7 nuclear accumulation and POM121 reduction. Importantly, partial knockdown of CHMP2B was sufficient to alleviate NPC injury and downstream TDP-43 dysfunction in sALS neurons thereby highlighting CHMP2B as a potential therapeutic target in disease.

Non-conventional deep brain stimulation in a network model of movement disorders

Conventional deep brain stimulation (DBS) for movement disorders is a well-established clinical treatment. Over the last few decades, over 200,000 people have been treated by DBS worldwide for several neurological conditions, including Parkinson's disease and Essential Tremor. DBS involves implanting electrodes into disorder-specific targets in the brain and applying an electric current. Although the hardware has developed in recent years, the clinically used stimulation pattern has remained as a regular frequency square pulse. Recent studies have suggested that phase-locking, coordinated reset or irregular patterns may be as or more effective at desynchronising the pathological neural activity. Such studies have shown efficacy using detailed neuron models or highly simplified networks and considered one frequency band. We previously described a population level model which generates oscillatory activity in both the beta band (20 Hz) and the tremor band (4 Hz). Here we use this model to look at the impact of applying regular, irregular and phase dependent bursts of stimulation, and show how this influences both tremor- and beta-band activity. We found that bursts are as or more effective at suppressing the pathological oscillations compared to continuous DBS. Importantly however, at higher amplitudes we found that the stimulus drove the network activity, as seen previously. Strikingly, this suppression was most apparent for the tremor band oscillations, with beta band pathological activity being more resistant to the burst stimulation compared to continuous, conventional DBS. Furthermore, our simulations showed that phase-locked bursts of stimulation did not convey much improvement on regular bursts of oscillation. Using a genetic algorithm optimisation approach to find the best stimulation parameters for regular, irregular and phase-locked bursts, we confirmed that tremor band oscillations could be more readily suppressed. Our results allow exploration of stimulation mechanisms at the network level to formulate testable predictions regarding parameter settings in DBS.

Estradiol elicits distinct firing patterns in arcuate nucleus kisspeptin neurons of females through altering ion channel conductances.

Hypothalamic kisspeptin (Kiss1) neurons are vital for pubertal development and reproduction. Arcuate nucleus Kiss1 (Kiss1ARH) neurons are responsible for the pulsatile release of gonadotropin-releasing hormone (GnRH). In females, the behavior of Kiss1ARH neurons, expressing Kiss1, neurokinin B (NKB), and dynorphin (Dyn), varies throughout the ovarian cycle. Studies indicate that 17β-estradiol (E2) reduces peptide expression but increases Slc17a6 (Vglut2) mRNA and glutamate neurotransmission in these neurons, suggesting a shift from peptidergic to glutamatergic signaling. To investigate this shift, we combined transcriptomics, electrophysiology, and mathematical modeling. Our results demonstrate that E2 treatment upregulates the mRNA expression of voltage-activated calcium channels, elevating the whole-cell calcium current that contributes to high-frequency burst firing. Additionally, E2 treatment decreased the mRNA levels of canonical transient receptor potential (TPRC) 5 and G protein-coupled K+ (GIRK) channels. When Trpc5 channels in Kiss1ARH neurons were deleted using CRISPR/SaCas9, the slow excitatory postsynaptic potential was eliminated. Our data enabled us to formulate a biophysically realistic mathematical model of Kiss1ARH neurons, suggesting that E2 modifies ionic conductances in these neurons, enabling the transition from high-frequency synchronous firing through NKB-driven activation of TRPC5 channels to a short bursting mode facilitating glutamate release. In a low E2 milieu, synchronous firing of Kiss1ARH neurons drives pulsatile release of GnRH, while the transition to burst firing with high, preovulatory levels of E2 would facilitate the GnRH surge through its glutamatergic synaptic connection to preoptic Kiss1 neurons.

Implementation of Self-Organized Operational Neural Networks for R Peak Detection in Holter ECG

While a number of R-peak detectors have been created, their performance may be significantly impacted when handling noisy, low-quality data from mobile ECG sensors, such as Holter monitors. Even though deep 1-D convolutional neural networks (CNNs) have recently produced state-of-the-art results, their high complexity and need for specialized parallel hardware for real-time processing can limit performance, particularly with compact network configurations. Because CNNs only use one linear neuron model, their learning capacity is limited, leading to this constraint. To tackle this problem, operational neural networks (ONNs) integrate neurons with several types of nonlinear operators in a network architecture that is heterogeneous. The goal of this work is to improve R-peak detection performance in 1-D Self-Organized ONNs (Self-ONNs) while maintaining computing efficiency through the use of generative neurons. Because each generating neuron in a 1-D Self-ONN learns its ideal configuration through adaptation, the Self-Organizing feature eliminates the need for human operator set selection. Our experimental results, utilizing the MIT-BIH Arrhythmia dataset, which contains over a million ECG beats, reveal that 1-D Self-ONNs outperform state-of-the-art deep CNNs in terms of both performance and computational economy.

Mitochondrial oxidant stress promotes α-synuclein aggregation and spreading in mice with mutated glucocerebrosidase

In this study, heterozygous expression of a common Parkinson-associated GBA1 variant, the L444P mutation, was found to exacerbate α-synuclein aggregation and spreading in a mouse model of Parkinson-like pathology targeting neurons of the medullary vagal system. These neurons were also shown to become more vulnerable to oxidative and nitrative stress after L444P expression. The latter paralleled neuronal formation of reactive oxygen species and led to a pronounced accumulation of nitrated α-synuclein. A causal relationship linked mutation-induced oxidative/nitrative stress to enhanced α-synuclein aggregation and spreading that could indeed be rescued by neuronal overexpression of mitochondrial superoxide dismutase 2. Further evidence supported a key involvement of mitochondria as sources of reactive oxygen species as well as targets of oxidative and nitrative damage within L444P-expressing neurons. These findings support the conclusion that enhanced vulnerability to mitochondrial oxidative stress should be considered an important mechanism predisposing to pathology conversion in carriers of GBA1 mutations.

Pulse Modulated Neural Nets with Adaptive Temporal Windows for Enhanced Video Perception

A key challenge in modern artificial intelligence is the limited awareness of temporal dynamics. This is especially evident in AI's struggles with video animation and interpreting sequential frames or time-dependent data. This paper presents a new method for modulating neuron spikes to synchronize with incoming frames, addressing the temporal limitations of voltage AI models. This approach involves a specialized class of neuron, similar to Wolfgang Maass's liquid state machine concept. However, our model differs significantly in its synchronization strategy. Neurons are synchronized with frame import, incorporating a flexible time window that enables spike modulation ranging from 1 to 1024 spikes per frame. This adaptability is crucial for accurately processing and interpreting time-based data. The main focus of this paper is to explore the operational principles of these specialized neurons. While the wiring and architectural design of these neurons will be discussed in a future article, this paper aims to establish the foundational understanding of how these neurons function individually and in response to temporal data inputs. By synchronizing neuron spikes with frame inputs and allowing for variable spike modulation, this method aims to enhance AI's ability to handle time-based data. The proposed neuron model represents a step towards more advanced and temporally aware AI systems, potentially unlocking new capabilities in video animation and other time-sensitive applications.

Single inertial neuron with forced bipolar pulse: chaotic dynamics, circuit implementation, and color image encryption

Abstract The bipolar pulse current can effectively mimic the external time-varying stimulus of neurons, and its effect of neuronal dynamics has rarely been reported. To this end, this paper reports the effects of bipolar pulses on a two-dimensional single inertial neuron model, showcasing the chaotic dynamics of hidden attractors and coexisting symmetric attractors, which is of significant importance for understanding the complex behaviors of neuron dynamics under time-varying external stimuli and its application. Firstly, the mathematical model of the single intertial neuron model with forced bipolar pulse is presented, and then the equilibrium states behaving as unstable saddle point (USP), stable node-focus (SNF), and stable node point (SNP) are analyzed. Additionally, by using multiple dynamical methods including bifurcation plots, basins of attraction, and phase plots, complex dynamics of interesting bifurcation behaviors and coexisting attractors are revealed, which are induced by the forced bipolar pulse current as well as initial values, both. In addition, such effets are well valideted via a simple multiplerless electronic neuron circuit. The implementation circuit of presented model is constructed on the analog level and executed using PSIM circuit platform. The measurement results verified the double-scroll chaotic attractors and the coexisting period/chaos behaviors. Finally, the chaotic sequences of the model are applied to color image encryption for the benefit of requirements on modern security field. The encryption effectiveness is demonstrated through various evaluation indexes, including histogram analysis, information entropy, correlation coefficient, plaintext sensitivity, and resistance to noise attacks.

Dynamical analysis of the Rulkov model with quasiperiodic forcing

This paper investigates the dynamical behavior of the Rulkov neuron model under quasiperiodic forcing, focusing on the emergence and mechanisms of strange nonchaotic attractors (SNAs). SNAs are characterized by fractal geometry without sensitive dependence on initial conditions, and their formation mechanisms, such as torus-doubling, fractal, bubble, Heagy–Hammel, and multi-intermittency routes, are explored in detail. The study employs Lyapunov exponents, singular continuous spectrum, and phase sensitivity to detect and characterize the strange and nonchaotic properties of attractors. Numerical simulations reveal the transition dynamics between strange nonchaotic, quasiperiodic, and chaotic states, emphasizing the role of transient dynamics in understanding the complex behavior of neural systems under multifrequency stimuli. The findings provide deeper insights into the intricate dynamical processes in neuron models, potentially aiding the development of advanced models for neural response prediction in varying environments.

TLR3 Knockdown Attenuates Pressure-Induced Neuronal Damage InVitro.

The disruption of nerve parenchyma and axonal networks triggered by spinal cord injury (SCI) can initiate a cascade of events associated with secondary injury. Toll-like receptors play a critical role in initiating and regulating immune-inflammatory responses following SCI; however, the precise involvement of Toll-like receptor-3 (TLR3) in secondary neuronal injury remains incompletely understood. To investigate the potential contribution of TLR3 in mediating neuronal pressure-induced damage, we established a stress-induced neuronal damage model using rat anterior horn motor neuron line (VSC4.1), which was subjected to varying levels and durations of sustained pressure. Our findings suggest that pressure induces neuronal damage and apoptosis, and reduced proliferation rates in VSC4.1 cells. Furthermore, this pressure-induced neuronal injury is accompanied by upregulation of TLR3 expression and activation of downstream TLR3 signalling molecules. Knockdown experiments targeting TLR3 significantly alleviate pressure-induced motor neuron injury and apoptosis within the anterior horn region while promoting mitochondria-related autophagy and reducing mitochondrial dysfunction via the TLR3/IRF3 and TLR3/NF-κB pathways.