Characterizing the neural representations underlying somatic perception is crucial for neural decoding of external stimulations. Acupuncture is an important therapeutic method of traditional Chinese medicine and can effectively modulate brain activity for the treatment of neural diseases. In this work, we investigate the neural representations based on the power spectral density (PSD) estimated from electroencephalogram (EEG) across the whole brain with deep learning. Frequency and spatial characteristics of PSD can reliably represent the dynamical brain responses to acupuncture with different manipulations, manifesting enhanced alpha power in parietal lobe. By removing aperiodic components, periodic spatial spectrum shows a higher representation ability of different brain states during acupuncture stimulations, and twiring-rotating (TR) manipulation have a more pronounced modulatory effect than lifting-thrusting (LT) manipulation. Moreover, we further infer the low-dimensional feature-disentangled representations with generative adversarial network (GAN), i.e., w-latents of StyleGAN, which can capture the latent features of periodic spatial spectrum and strike a balance between separability and generalizability. The effectiveness of feature-disentangled representations is evaluated by decoding the acupuncture states, which can achieve a highest accuracy of 95.71% with Transformer classifier. Compared with high-dimensional spatial spectrum, low-dimensional latent features can best characterize different brain states, indicating a precise representation of somatic perception of acupuncture stimulations. Taken together, our results highlight the significant role of spatial spectral representation underlying somatic perception and serve as an important benchmark for the evaluation of acupuncture effect on human brain.

Acupuncture stimulations in somatosensory system can modulate spatiotemporal brain activity and improve cognitive functions of patients with neurological disorders. The correlation between these somatosensory stimulations and dynamical brain responses is still unclear. We proposed a deep learning framework using electroencephalographic activity of stimulated subjects to decode the needling processes of various acupuncture manipulations performed on Zusanli acupoint. Contrastive representation learning integrated with domain adaptation strategy was applied to estimate 3D hand postures and hand joint motion trajectories of acupuncturist with video recordings, by which finite dimensional representations of behavior manifolds for needling operations were inferred. Distinct transition dynamics of behavior manifold were observed for acupuncture with lifting-thrusting and twisting-rotating manipulations. Moreover, latent neural manifolds of acupuncture evoked EEG signals were estimated in low dimensional state space of brain activities with unsupervised manifold learning, which can reliably represent acupuncture stimulations. Furthermore, a nonlinear decoder based on neural networks was designed to transform neural manifolds to behavior manifolds and further predict acupuncture manipulation as well as needling process. Experimental results demonstrated a high performance of the proposed decoding framework for four types of acupuncture manipulations with a precision of 92.42%. The EEG decoder provides an acupuncture-brain interface linking somatosensory stimulations with neural representations, an effective scheme for revealing clinical efficacy of acupuncture treatment.

Though the wireless in-flight charging is an ideal way for the energy supply of drones, it also faces the practical challenges, namely dual-disturbance of continuous fluctuation of mutual inductance and battery load variation, changed expected charging power, and the lightweight and no communication design demands for pickup, which have been nearly unexplored in previous article on wireless power transfer (WPT) system. To address the issue, this article proposes a novel communication-free output power control algorithm based on the load identification scheme with automatic frequency adjust. Only the primary current needs to be measured in the proposed algorithm, which can reduce the system complexity with the enhanced real-time performance for WPT system. Simulated and experimental results validate the feasibility of the proposed control algorithm with the identification accuracy of more than 92%, control accuracy of 95%, and average response time within 100 ms. Furthermore, the follow-up of desired output power and the implementation of lightweight pickup demonstrate the flexibility of the power control algorithm, which ensures the rapid and safe energy supply for drones in the drone wireless in-flight charging system.

Distributed edge computing platforms are of great significance for the implementation of brain-like computing research. Due to the limited power consumption and real-time requirements, hardware acceleration of computing units is a challenging task. Taking advantage of both scalable hardware framework and lower cost, this article designs a multicore distributed computing platform for spiking neural networks. Particularly, a shared memory partition structure is utilized to participate in hardware acceleration. Through the spike-queue-based synaptic mapping mechanism, each parallel computing unit deals with efficient point-to-point connections. In addition, this platform provides a basic community unit (BCU) that encapsulates a standard neuron model library and rich peripheral interfaces. With the support of GUI, users can quickly build large-scale systems. The experimental results show that a single BCU can accommodate more than 10 000 neurons updated in real-time at a power consumption of 273.6 mW. The extended BCU group is able to perform network dynamic simulations in the basal ganglia-thalamus plausible biological network composed of Hodgkin–Huxley neurons as well as MNIST dataset classification in the leaky-integrate-fire network. The outstanding flexibility and real-time performance of the proposed hardware architecture provide great potential for embedded applications of neural computation.

Variations in brain activity patterns reveal impairments of motor and cognitive functions in the human brain. Electroencephalogram (EEG) microstates embody brain activity patterns at a microscopic time scale. However, current microstate analysis method can only recognize less than 90% of EEG signals per subject, which severely limits the characterization of dynamic brain activity. As an application to early Parkinson's disease (PD), we propose an enhanced EEG microstate recognition framework based on deep neural networks, which yields recognition rates from 90% to 99%, as accompanied by a strong anti-artifact property. Additionally, gradient-weighted class activation mapping, as a visualization technique, is employed to locate the activated functional brain regions of each microstate class. We find that each microstate class corresponds to a particular activated brain region. Finally, based on the improved identification of microstate sequences, we explore the EEG microstate characteristics and their clinical associations. We show that the decreased occurrences of a particular microstate class reflect the degree of cognitive decline in early PD, and reduced transitions between certain microstates suggest injury in motor-related brain regions. The novel EEG microstate recognition framework paves the way to revealing more effective biomarkers for early PD.

This work explores the effectiveness of the intensity-varied closed-loop noise stimulation on the oscillation suppression in the Parkinsonian state. Deep brain stimulation (DBS) is the standard therapy for Parkinson's disease (PD), but its effects need to be improved. The noise stimulation has compelling results in alleviating the PD state. However, in the open-loop control scheme, the noise stimulation parameters cannot be self-adjusted to adapt to the amplitude of the synchronized neuronal activities in real time. Thus, based on the delayed-feedback control algorithm, an intensity-varied closed-loop noise stimulation strategy is proposed. Based on a computational model of the basal ganglia (BG) that can present the intrinsic properties of the BG neurons and their interactions with the thalamic neurons, the proposed stimulation strategy is tested. Simulation results show that the noise stimulation suppresses the pathological beta (12-35 Hz) oscillations without any new rhythms in other bands compared with traditional high-frequency DBS. The intensity-varied closed-loop noise stimulation has a more profound role in removing the pathological beta oscillations and improving the thalamic reliability than open-loop noise stimulation, especially for different PD states. And the closed-loop noise stimulation enlarges the parameter space of the delayed-feedback control algorithm due to the randomness of noise signals. We also provide a theoretical analysis of the effective parameter domain of the delayed-feedback control algorithm by simplifying the BG model to an oscillator model. This exploration may guide a new approach to treating PD by optimizing the noise-induced improvement of the BG dysfunction.

Acupuncture can regulate the functions of human body and improve the cognition of brain. However, the mechanism of acupuncture manipulations remains unclear. Here, we hypothesis that the frontal cortex plays a gating role in information routing of brain network under acupuncture. To that end, the gating effect of frontal cortex under acupuncture is analyzed in combination with EEG data of acupuncture at Zusanli acupoints. In addition, recurrent neural network (RNN) is used to reproduce the dynamics of frontal cortex under normal state and acupuncture state. From low-dimensional view, it is shown that the brain networks under acupuncture state can show stable attractor cycle dynamics, which may explain the regulation effect of acupuncture. Comparing with different manipulations, we find that the attractor of low-dimensional trajectory varies under different frequencies of acupuncture. Besides, a strip gated band of neural dynamics is found by changing the frequency of stimulation and excitatory-inhibitory balance of network. This reverse engineering of brain network indicates that the differences among acupuncture manipulations are caused by interaction and separation in the neural activity space between attractors that encode acupuncture function. Consequently, our results may provide help for quantitative analysis of acupuncture, and benefit for the clinical guidance of acupuncture clinicians.

The further exploration of the neural mechanisms underlying the biological activities of the human brain depends on the development of large-scale spiking neural networks (SNNs) with different categories at different levels, as well as the corresponding computing platforms. Neuromorphic engineering provides approaches to high-performance biologically plausible computational paradigms inspired by neural systems. In this article, we present a biological-inspired cognitive supercomputing system (BiCoSS) that integrates multiple granules (GRs) of SNNs to realize a hybrid compatible neuromorphic platform. A scalable hierarchical heterogeneous multicore architecture is presented, and a synergistic routing scheme for hybrid neural information is proposed. The BiCoSS system can accommodate different levels of GRs and biological plausibility of SNN models in an efficient and scalable manner. Over four million neurons can be realized on BiCoSS with a power efficiency of 2.8k larger than the GPU platform, and the average latency of BiCoSS is 3.62 and 2.49 times higher than conventional architectures of digital neuromorphic systems. For the verification, BiCoSS is used to replicate various biological cognitive activities, including motor learning, action selection, context-dependent learning, and movement disorders. Comprehensively considering the programmability, biological plausibility, learning capability, computational power, and scalability, BiCoSS is shown to outperform the alternative state-of-the-art works for large-scale SNN, while its real-time computational capability enables a wide range of potential applications.

Neuromorphic computing is a promising technology that realizes computation based on event-based spiking neural networks (SNNs). However, fault-tolerant on-chip learning remains a challenge in neuromorphic systems. This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework. We show how this system can learn associations between stimulation and response in two context-dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes. Furthermore, we demonstrate how our novel fault-tolerant neuromorphic spike routing scheme can avoid multiple fault nodes successfully and can enhance the maximum throughput of the neuromorphic network by 0.9%-16.1% in comparison with previous studies. By utilizing the real-time computational capabilities and multiple-fault-tolerant property of the proposed system, the neuronal mechanisms underlying the spiking activities of neuromorphic networks can be readily explored. In addition, the proposed system can be applied in real-time learning and decision-making applications, brain-machine integration, and the investigation of brain cognition during learning.

The cerebellum plays a vital role in motor learning and control with supervised learning capability, while neuromorphic engineering devises diverse approaches to high-performance computation inspired by biological neural systems. This article presents a large-scale cerebellar network model for supervised learning, as well as a cerebellum-inspired neuromorphic architecture to map the cerebellar anatomical structure into the large-scale model. Our multinucleus model and its underpinning architecture contain approximately 3.5 million neurons, upscaling state-of-the-art neuromorphic designs by over 34 times. Besides, the proposed model and architecture incorporate 3411k granule cells, introducing a 284 times increase compared to a previous study including only 12k cells. This large scaling induces more biologically plausible cerebellar divergence/convergence ratios, which results in better mimicking biology. In order to verify the functionality of our proposed model and demonstrate its strong biomimicry, a reconfigurable neuromorphic system is used, on which our developed architecture is realized to replicate cerebellar dynamics during the optokinetic response. In addition, our neuromorphic architecture is used to analyze the dynamical synchronization within the Purkinje cells, revealing the effects of firing rates of mossy fibers on the resonance dynamics of Purkinje cells. Our experiments show that real-time operation can be realized, with a system throughput of up to 4.70 times larger than previous works with high synaptic event rate. These results suggest that the proposed work provides both a theoretical basis and a neuromorphic engineering perspective for brain-inspired computing and the further exploration of cerebellar learning.

An extracellular electric field (EF) induces transmembrane polarizations on extremely inhomogeneous spaces. Evidence shows that EF-induced somatic polarization in pyramidal cells can modulate the neuronal input-output (I/O) function. However, it remains unclear whether and how dendritic polarization participates in the dendritic integration and contributes to the neuronal I/O function. To this end, we built a computational model of a simplified pyramidal cell with multi-dendritic tufts, one dendritic trunk, and one soma to describe the interactions among EF, dendritic integration, and somatic output, in which the EFs were modeled by inserting inhomogeneous extracellular potentials. We aimed to establish the underlying relationship between dendritic polarization and dendritic integration by analyzing the dynamics of subthreshold membrane potentials in response to AMPA synapses in the presence of constant EFs. The model-based singular perturbation analysis showed that the equilibrium mapping of a fast subsystem can serve as the asymptotic subthreshold I/O relationship for sublinear dendritic integration. This allows us to predict the tendency of EF-mediated dendritic integration by showing how EF changes modify equilibrium mapping. EF-induced hyperpolarization of distal dendrites receiving synapses inputs was found to play a key role in facilitating the AMPA receptor-evoked excitatory postsynaptic potential (EPSP) by enhancing the driving force of synaptic inputs. A significantly higher efficacy of EF modulation effect on global AMPA-type dendritic integration was found compared with local AMPA-type dendritic integration. During the generation of an action potential (AP), the relative contribution of EF-modulated dendritic integration and EF-induced somatic polarization was determined to show their collaboration in promoting or inhibiting the somatic excitability, depending on the EF polarity. These findings are crucial for understanding the EF modulation effect on neuronal computation, which provides insight into the modulation mechanism of noninvasive brain modulation.

Suppression of excessively synchronous beta frequency (12-35 Hz) oscillatory activity in the basal ganglia is believed to correlate with the alleviation of hypokinetic motor symptoms of the Parkinson's disease. Delayed feedback is an effective strategy to interrupt the synchronization and has been used in the design of closed-loop neuromodulation methods computationally. Although tremendous efforts in this are being made by optimizing delayed feedback algorithm and stimulation waveforms, there are still remaining problems in the selection of effective parameters in the delayed feedback control schemes. In most delayed feedback neuromodulation strategies, the stimulation signal is obtained from the local field potential (LFP) of the excitatory subthalamic nucleus (STN) neurons and then is administered back to STN itself only. The inhibitory external globus pallidus (GPe) nucleus in the excitatory-inhibitory STN-GPe reciprocal network has not been involved in the design of the delayed feedback control strategies. Thus, considering the role of GPe, this paper proposes three schemes involving GPe in the design of the delayed feedback strategies and compared their effectiveness to the traditional paradigm using STN only. Based on a neural mass model of STN-GPe network having capability of simulating the LFP directly, the proposed stimulation strategies are tested and compared. Our simulation results show that the four types of delayed feedback control schemes are all effective, even if with a simple linear delayed feedback algorithm. But the three new control strategies we propose here further improve the control performance by enlarging the oscillatory suppression space and reducing the energy expenditure, suggesting that they may be more effective in applications. This paper may guide a new approach to optimize the closed-loop deep brain stimulation treatment to alleviate the Parkinsonian state by retargeting the measurement and stimulation nucleus.

Thalamic deep brain stimulation (DBS) generates excitatory postsynaptic currents and action potentials (APs) by triggering large numbers of synaptic inputs to local cells, which also activates axonal spikes to antidromically invade the soma and dendrites. To maintain signaling, the evoked dendritic responses require metabolic energy to restore ion gradients in each dendrite. The objective of this study is to estimate the energy demand associated with dendritic responses to thalamic DBS. We use a morphologically realistic computational model to simulate dendritic activity in thalamocortical (TC) relay neurons with axonal intracellular stimulation or DBS-like extracellular stimulation. We determine the metabolic cost by calculating the number of adenosine triphosphate (ATP) expended to pump Na+and Ca2+ions out of each dendrite. The ATP demand of dendritic activity exhibits frequency dependence, which is determined by the number of spikes in the dendrites. Each backpropagating AP from the soma activates a spike in the dendrites, and the dendritic firing is dominated by antidromic activation of the soma. High stimulus frequencies decrease dendritic ATP cost by reducing the fidelity of antidromic activation. Synaptic inputs and stimulus-induced polarization govern the ATP cost of dendritic responses by facilitating/suppressing antidromic activation, which also influences the ATP cost by depolarizing/hyperpolarizing each dendrite. These findings are important for understanding the synaptic signaling energy in TC relay neurons and metabolism-dependent functional imaging data of thalamic DBS.

We employed high-density microelectrode arrays to investigate spontaneous firing patterns of neurons in brain circuits of the primary somatosensory cortex (S1) in mice. We recorded from over 150 neurons for 10[Formula: see text]min in each of eight different experiments, identified their location in S1, sorted their action potentials (spikes), and computed their power spectra and inter-spike interval (ISI) statistics. Of all persistently active neurons, 92% fired with a single dominant frequency — regularly firing neurons (RNs) — from 1 to 8[Formula: see text]Hz while 8% fired in burst with two dominant frequencies — bursting neurons (BNs) — corresponding to the inter-burst (2–6[Formula: see text]Hz) and intra-burst intervals (20–160[Formula: see text]Hz). RNs were predominantly located in layers 2/3 and 5/6 while BNs localized to layers 4 and 5. Across neurons, the standard deviation of ISI was a power law of its mean, a property known as fluctuation scaling, with a power law exponent of 1 for RNs and 1.25 for BNs. The power law implies that firing and bursting patterns are scale invariant: the firing pattern of a given RN or BN resembles that of another RN or BN, respectively, after a time contraction or dilation. An explanation for this scale invariance is discussed in the context of previous computational studies as well as its potential role in information processing.

Stochasticity and oscillation play a vital role in neural signal processing. Time delay, which is inevitable in biological neural systems, has significant effect on the dynamics of neuronal networks. This paper provides an analysis of how time delay affects stochastic resonance and firing rate oscillation of cortical neuronal networks. A cortical network is established and mean-field theory is applied to analytically compute the dynamical response of networks. When the frequency of external stimulation is close to intrinsic frequency of neuronal networks, firing rate exhibits coherent oscillation and the phenomenon of stochastic resonance occurs in inhibitory neurons. Time delay can induce multiple stochastic resonances, which appear intermittently at integer multiples of the period of input signal, due to the transition of network dynamics induced by time delays. The fluctuation of membrane potential and instantaneous firing rate of cortical networks achieve maximal periodically with the variation of time delay. Furthermore, time delay and electrical coupling play complementary roles in determining network responses. Network oscillation can transit from unstable to stable when coupling strength exceeds a critical value. Transition threshold is lower for time delays close to integer multiples of input period where resonant response of cortical network enhances the formation of stable oscillation.

The stochastic dynamics of neural network are studied and the phenomenon of stochastic resonance is found in inhibitory neurons, whose firing rate is close to the frequency of external stimulation. Time delay in the neural coupling process can induce multiple stochastic resonances, which appear intermittently at the integer multiples of the oscillation period of the input signal. It is found that the time delay can induce the periodic oscillation of neural firing rate, which may account for the occurrence of multiple stochastic resonances. In addition, the effect of synapses on firing rate oscillation and network resonance is investigated. As the strength of gap-junction and inhibitory-inhibitory chemical coupling is increased, the maximal resonant value increases, while the resonant frequency is unchanged. However, the resonant frequency and peak value increase with the coupling strength of excitatory-inhibitory chemical synapses. This difference may result from the interaction of excitation and inhibition within the cortical network. We further apply mean-field theory to time-delayed network model to validate the obtained numerical results. Both time delay and electrical-chemical synapses play an important role in firing rate oscillation and stochastic resonance within the cortical network, determining the ability to enhance the transmission of information in neural systems.

Accurate identification of Alzheimer's disease (AD) with electroencephalograph (EEG) is crucial in the clinical diagnosis of neurological disorders. However, the effectiveness and accuracy of manually labeling EEG signals are barely satisfactory, due to lacking effective biomarkers. In this paper, we propose a novel machine learning method network-based Takagi-Sugeno-Kang (N-TSK) for AD identification which employs the complex network theory and TSK fuzzy system. With the construction of functional network of AD subjects, the topological features of weighted and unweighted networks are extracted. Taken the network parameters as independent inputs, a fuzzy-system-based TSK model is established and further trained to identify AD EEG signals. Experimental results demonstrate the effectiveness of the proposed scheme in AD identification and ability of N-TSK fuzzy classifiers. The highest accuracy can achieve 97.3% for patients with closed eyes and 94.78% with open eyes. In addition, the performance of weighted N-TSK largely exceeds unweighted N-TSK. By further optimizing the network features utilized in the N-TSK fuzzy classifiers, it is found that local efficiency and clustering coefficient are the most effective factors in AD identification. This work provides a potential tool for identifying neurological disorders from the perspective of functional networks with EEG signal, especially contributing to the diagnosis and identification of AD.

Recent studies have demonstrated the effectiveness of supervised learning in spiking neural networks (SNNs). A trainable SNN provides a valuable tool not only for engineering applications but also for theoretical neuroscience studies. Here, we propose a modified SpikeProp learning algorithm, which ensures better learning stability for SNNs and provides more diverse network structures and coding schemes. Specifically, we designed a spike gradient threshold rule to solve the well-known gradient exploding problem in SNN training. In addition, regulation rules on firing rates and connection weights are proposed to control the network activity during training. Based on these rules, biologically realistic features such as lateral connections, complex synaptic dynamics, and sparse activities are included in the network to facilitate neural computation. We demonstrate the versatility of this framework by implementing three well-known temporal codes for different types of cognitive tasks, namely, handwritten digit recognition, spatial coordinate transformation, and motor sequence generation. Several important features observed in experimental studies, such as selective activity, excitatory-inhibitory balance, and weak pairwise correlation, emerged in the trained model. This agreement between experimental and computational results further confirmed the importance of these features in neural function. This work provides a new framework, in which various neural behaviors can be modeled and the underlying computational mechanisms can be studied.

Multicompartment emulation is an essential step to enhance the biological realism of neuromorphic systems and to further understand the computational power of neurons. In this paper, we present a hardware efficient, scalable, and real-time computing strategy for the implementation of large-scale biologically meaningful neural networks with one million multi-compartment neurons (CMNs). The hardware platform uses four Altera Stratix III field-programmable gate arrays, and both the cellular and the network levels are considered, which provides an efficient implementation of a large-scale spiking neural network with biophysically plausible dynamics. At the cellular level, a cost-efficient multi-CMN model is presented, which can reproduce the detailed neuronal dynamics with representative neuronal morphology. A set of efficient neuromorphic techniques for single-CMN implementation are presented with all the hardware cost of memory and multiplier resources removed and with hardware performance of computational speed enhanced by 56.59% in comparison with the classical digital implementation method. At the network level, a scalable network-on-chip (NoC) architecture is proposed with a novel routing algorithm to enhance the NoC performance including throughput and computational latency, leading to higher computational efficiency and capability in comparison with state-of-the-art projects. The experimental results demonstrate that the proposed work can provide an efficient model and architecture for large-scale biologically meaningful networks, while the hardware synthesis results demonstrate low area utilization and high computational speed that supports the scalability of the approach.

The benefit of noise in improving the basal ganglia (BG) dysfunctions, especially Parkinsonian state, is explored in this paper. High frequency (≥ 100 Hz) deep brain stimulation (DBS), as a clinical effective stimulation method, has compelling and fantastic results in alleviating the motor symptoms of Parkinson's disease (PD). However, the mechanism of DBS is still unclear. And the selection of the DBS waveform parameters faces great challenges to further optimize the stimulation effects and to reduce its energy expenditure. Considering that the desynchronization of the BG neuronal activities is benefited from the forced high frequency regular spikes driven by standard high frequency DBS, we expect to explore a novel stimulation method that has capability of restoring the BG physiological firing patterns without introducing artificial high-frequency fires. In this paper, a colored noise stimulation is used as a neuromodulation method to disrupt the firing patterns of the pathological neuronal activities. A computational model of the BG that exhibits the intrinsic properties of the BG neurons and their interactions with the thalamic (Th) cells is employed. Based on the model, we investigate the effects of noise stimulation and explore the impacts of the noise stimulation parameters on both relay reliability of the Th neurons and energy expenditure of the stimulation. By comparison, it can be found that noise stimulation does not entrain the network to an artificial high-frequency firing state, but induces the pathological increased synchronous activities back to a normal physiological level. Moreover, besides the capability of restoring the neuronal state, the benefits of the noise also include its balanced waveform to avert potential tissue or electrode damage and its ability to reduce the energy expenditure to 50% less than that of the standard DBS, when the noise stimulation has low frequency (≤ 100 Hz) and appropriate intensity. Thus, the exploration of the optimal noise-induced improvement of the BG dysfunction is of great significance in treating symptoms of neurological disorders such as PD.

The investigation of the human intelligence, cognitive systems and functional complexity of human brain is significantly facilitated by high-performance computational platforms. In this paper, we present a real-time digital neuromorphic system for the simulation of large-scale conductance-based spiking neural networks (LaCSNN), which has the advantages of both high biological realism and large network scale. Using this system, a detailed large-scale cortico-basal ganglia-thalamocortical loop is simulated using a scalable 3-D network-on-chip (NoC) topology with six Altera Stratix III field-programmable gate arrays simulate 1 million neurons. Novel router architecture is presented to deal with the communication of multiple data flows in the multinuclei neural network, which has not been solved in previous NoC studies. At the single neuron level, cost-efficient conductance-based neuron models are proposed, resulting in the average utilization of 95% less memory resources and 100% less DSP resources for multiplier-less realization, which is the foundation of the large-scale realization. An analysis of the modified models is conducted, including investigation of bifurcation behaviors and ionic dynamics, demonstrating the required range of dynamics with a more reduced resource cost. The proposed LaCSNN system is shown to outperform the alternative state-of-the-art approaches previously used to implement the large-scale spiking neural network, and enables a broad range of potential applications due to its real-time computational power.

One of the challenging problems in real-time control of movement disorders is the effective handling of time-variant brain activities that involve stochastic functional networks with nonlinear dynamics. For such challenges in neuromodulation tasks, fuzzy logic control (FLC) has shown significant potential. The objective of this paper is to present a FLC-based strategy to treat pathological symptoms of movement-disorder with higher performance. The strategy is two-fold: first, develop a design methodology for the FLC system that can robustly control pathological conditions and significantly improve control performance; and second, develop a hardware-efficient implementation for real-time neuromodulation applications. To enhance control performance, a hidden variable in the neural network that can be estimated using an unscented Kalman filter is identified as a feedback variable. In comparison with state-of-the-art schemes, the proposed design can adaptively optimize the control signals without requiring particular information of the controlled plant, thus avoiding repeated determinations of controller parameters. A field-programmable gate array is used for the reconfigurable realization of the entire control strategy based on a modification of the original neural network. The presented design, with enhanced control performance and higher hardware efficiency, has significant potential for clinical treatment of movement disorders and offers a new perspective on applications in the fields of neural control engineering and brain-machine interfaces.

Enhanced beta (12-30 Hz) oscillatory activity in the basal ganglia (BG) is a prominent feature of the Parkinsonian state in animal models and in patients with Parkinson's disease. Increased beta oscillations are associated with severe dopaminergic striatal depletion. However, the mechanisms underlying these pathological beta oscillations remain elusive. Inspired by the experimental observation that only subsets of neurons within each nucleus in the BG exhibit oscillatory activities, a computational model of the BG-thalamus neuronal network is proposed, which is characterized by subdivided nuclei within the BG. Using different currents externally applied to the neurons within a given nucleus, neurons behave according to one of the two subgroups, named "-N" and "-P," where "-N" and "-P" denote the normal and the Parkinsonian states, respectively. The ratio of "-P" to "-N" neurons indicates the degree of the Parkinsonian state. Simulation results show that if "-P" neurons have a high degree of connectivity in the subthalamic nucleus (STN), they will have a significant downstream effect on the generation of beta oscillations in the globus pallidus. Interestingly, however, the generation of beta oscillations in the STN is independent of the selection of the "-P" neurons in the external segment of the globus pallidus (GPe), despite the reciprocal structure between STN and GPe. This computational model may pave the way to revealing the mechanism of such pathological behaviors in a realistic way that can replicate experimental observations. The simulation results suggest that the STN is more suitable than GPe as a deep brain stimulation target.

Neuronal communication between different brain areas is achieved in terms of spikes. Consequently, spike-time regularity is closely related to many cognitive tasks and timing precision of neural information processing. A recent experiment on primate parietal cortex reports that spike-time regularity increases consistently from primary sensory to higher cortical regions. This observation conflicts with the influential view that spikes in the neocortex are fundamentally irregular. To uncover the underlying network mechanism, we construct a multilayered feedforward neural information transmission pathway and investigate how spike-time regularity evolves across subsequent layers. Numerical results reveal that despite the obviously irregular spiking patterns in previous several layers, neurons in downstream layers can generate rather regular spikes, which depends on the network topology. In particular, we find that collective temporal regularity in deeper layers exhibits resonance-like behavior with respect to both synaptic connection probability and synaptic weight, i.e., the optimal topology parameter maximizes the spike-timing regularity. Furthermore, it is demonstrated that synaptic properties, including inhibition, synaptic transient dynamics, and plasticity, have significant impacts on spike-timing regularity propagation. The emergence of the increasingly regular spiking (RS) patterns in higher parietal regions can, thus, be viewed as a natural consequence of spiking activity propagation between different brain areas. Finally, we validate an important function served by increased RS: promoting reliable propagation of spike-rate signals across downstream layers.

A generalized predictive closed-loop control strategy to improve the basal ganglia activity patterns in Parkinson's disease (PD) is explored in this paper. Based on system identification, an input-output model is established to reveal the relationship between external stimulation and neuronal responses. The model contributes to the implementation of the generalized predictive control (GPC) algorithm that generates the optimal stimulation waveform to modulate the activities of neuronal nuclei. By analyzing the roles of two critical control parameters within the GPC law, optimal closed-loop control that has the capability of restoring the normal relay reliability of the thalamus with the least stimulation energy expenditure can be achieved. In comparison with open-loop deep brain stimulation and traditional static control schemes, the generalized predictive closed-loop control strategy can optimize the stimulation waveform without requiring any particular knowledge of the physiological properties of the system. This type of closed-loop control strategy generates an adaptive stimulation waveform with low energy expenditure with the potential to improve the treatments for PD.

The objective here is to explore the use of adaptive input-output feedback linearization method to achieve an improved deep brain stimulation (DBS) algorithm for closed-loop control of Parkinson's state. The control law is based on a highly nonlinear computational model of Parkinson's disease (PD) with unknown parameters. The restoration of thalamic relay reliability is formulated as the desired outcome of the adaptive control methodology, and the DBS waveform is the control input. The control input is adjusted in real time according to estimates of unknown parameters as well as the feedback signal. Simulation results show that the proposed adaptive control algorithm succeeds in restoring the relay reliability of the thalamus, and at the same time achieves accurate estimation of unknown parameters. Our findings point to the potential value of adaptive control approach that could be used to regulate DBS waveform in more effective treatment of PD.

To provide insights into the modulation of neuronal activity by extremely low-frequency (ELF) magnetic field (MF), we present a conductance-based neuron model and introduce ELF sinusoidal MF as an additive voltage input. By analyzing spike times and spiking frequency, it is observed that neuron with distinct spiking patterns exhibits different response properties in the presence of MF exposure. For tonic spiking neuron, the perturbations of MF exposure on spike times is maximized at the harmonics of neuronal intrinsic spiking frequency, while it is maximized at the harmonics of bursting frequency for burst spiking neuron. As MF intensity increases, the perturbations also increase. Compared with tonic spiking, bursting dynamics are less sensitive to the perturbations of ELF MF exposure. Further, ELF MF exposure is more prone to perturb neuronal spike times relative to spiking frequency. Our finding suggests that the resonance may be one of the neural mechanisms underlying the modulatory effects of the low-intensity ELF MFs on neuronal activities. The results highlight the impacts of ELF MFs exposure on neuronal activity from the single cell level, and demonstrate various factors including ELF MF properties and neuronal spiking characteristics could determine the outcome of exposure. These insights into the mechanism of MF exposure may be relevant for the design of multi-intensity magnetic stimulus protocols, and may even contribute to the interpretation of MF effects on the central nervous systems.

A novel closed-loop control strategy is proposed to control Parkinsonian state based on a computational model. By modeling thalamocortical relay neurons under external electric field, a slow variable feedback control is applied to restore its relay functionality. Qualitative and quantitative analysis demonstrates the performance of feedback controller based on slow variable is more efficient compared with traditional feedback control based on fast variable. These findings point to the potential value of model-based design of feedback controllers for Parkinson's disease.