Epileptic Seizure Detection Using Geometric Features Extracted from SODP Shape of EEG Signals and AsyLnCPSO-GA.

Epilepsy is a neurological disorder that is characterized by transient and unexpected electrical disturbance of the brain. Seizure detection by electroencephalogram (EEG) is associated with the primary interest of the evaluation and auxiliary diagnosis of epileptic patients. The aim of this study is to establish a hybrid model with improved particle swarm optimization (PSO) and a genetic algorithm (GA) to determine the optimal combination of features for epileptic seizure detection. First, the second-order difference plot (SODP) method was applied, and ten geometric features of epileptic EEG signals were derived in each frequency band (δ, θ, α and ß), forming a high-dimensional feature vector. Secondly, an optimization algorithm, AsyLnCPSO-GA, combining a modified PSO with asynchronous learning factor (AsyLnCPSO) and the genetic algorithm (GA) was proposed for feature selection. Finally, the feature combinations were fed to a naïve Bayesian classifier for epileptic seizure and seizure-free identification. The method proposed in this paper achieved 95.35% classification accuracy with a tenfold cross-validation strategy when the interfrequency bands were crossed, serving as an effective method for epilepsy detection, which could help clinicians to expeditiously diagnose epilepsy based on SODP analysis and an optimization algorithm for feature selection.

Directed Brain Network Analysis for Fatigue Driving Based on EEG Source Signals.

Fatigue driving is one of the major factors that leads to traffic accidents. Long-term monotonous driving can easily cause a decrease in the driver's attention and vigilance, manifesting a fatigue effect. This paper proposes a means of revealing the effects of driving fatigue on the brain's information processing abilities, from the aspect of a directed brain network based on electroencephalogram (EEG) source signals. Based on current source density (CSD) data derived from EEG signals using source analysis, a directed brain network for fatigue driving was constructed by using a directed transfer function. As driving time increased, the average clustering coefficient as well as the average path length gradually increased; meanwhile, global efficiency gradually decreased for most rhythms, suggesting that deep driving fatigue enhances the brain's local information integration abilities while weakening its global abilities. Furthermore, causal flow analysis showed electrodes with significant differences between the awake state and the driving fatigue state, which were mainly distributed in several areas of the anterior and posterior regions, especially under the theta rhythm. It was also found that the ability of the anterior regions to receive information from the posterior regions became significantly worse in the driving fatigue state. These findings may provide a theoretical basis for revealing the underlying neural mechanisms of driving fatigue.

Brain Complex Network Characteristic Analysis of Fatigue during Simulated Driving Based on Electroencephalogram Signals.

Fatigued driving is one of the major causes of traffic accidents. Frequent repetition of driving behavior for a long time may lead to driver fatigue, which is closely related to the central nervous system. In the present work, we designed a fatigue driving simulation experiment and collected the electroencephalogram (EEG) signals. Complex network theory was introduced to study the evolution of brain dynamics under different rhythms of EEG signals during several periods of the simulated driving. The results show that as the fatigue degree deepened, the functional connectivity and the clustering coefficients increased while the average shortest path length decreased for the delta rhythm. In addition, there was a significant increase of the degree centrality in partial channels on the right side of the brain for the delta rhythm. Therefore, it can be concluded that driving fatigue can cause brain complex network characteristics to change significantly for certain brain regions and certain rhythms. This exploration may provide a theoretical basis for further finding objective and effective indicators to evaluate the degree of driving fatigue and to help avoid fatigue driving.

Effects of topologies on signal propagation in feedforward networks.

We systematically investigate the effects of topologies on signal propagation in feedforward networks (FFNs) based on the FitzHugh-Nagumo neuron model. FFNs with different topological structures are constructed with same number of both in-degrees and out-degrees in each layer and given the same input signal. The propagation of firing patterns and firing rates are found to be affected by the distribution of neuron connections in the FFNs. Synchronous firing patterns emerge in the later layers of FFNs with identical, uniform, and exponential degree distributions, but the number of synchronous spike trains in the output layers of the three topologies obviously differs from one another. The firing rates in the output layers of the three FFNs can be ordered from high to low according to their topological structures as exponential, uniform, and identical distributions, respectively. Interestingly, the sequence of spiking regularity in the output layers of the three FFNs is consistent with the firing rates, but their firing synchronization is in the opposite order. In summary, the node degree is an important factor that can dramatically influence the neuronal network activity.

Endogenous field feedback promotes the detectability for exogenous electric signal in the hybrid coupled population.

This paper presents the endogenous electric field in chemical or electrical synaptic coupled networks, aiming to study the role of endogenous field feedback in the signal propagation in neural systems. It shows that the feedback of endogenous fields to network activities can reduce the required energy of the noise and enhance the transmission of input signals in hybrid coupled populations. As a common and important nonsynaptic interactive method among neurons, particularly, the endogenous filed feedback can not only promote the detectability of exogenous weak signal in hybrid coupled neural population but also enhance the robustness of the detectability against noise. Furthermore, with the increasing of field coupling strengths, the endogenous field feedback is conductive to the stochastic resonance by facilitating the transition of cluster activities from the no spiking to spiking regions. Distinct from synaptic coupling, the endogenous field feedback can play a role as internal driving force to boost the population activities, which is similar to the noise. Thus, it can help to transmit exogenous weak signals within the network in the absence of noise drive via the stochastic-like resonance.

The estimation of neurotransmitter release probability in feedforward neuronal network based on adaptive synchronization.

In this paper, we proposed a new approach to estimate unknown parameters and topology of a neuronal network based on the adaptive synchronization control scheme. A virtual neuronal network is constructed as an observer to track the membrane potential of the corresponding neurons in the original network. When they achieve synchronization, the unknown parameters and topology of the original network are obtained. The method is applied to estimate the real-time status of the connection in the feedforward network and the neurotransmitter release probability of unreliable synapses is obtained by statistic computation. Numerical simulations are also performed to demonstrate the effectiveness of the proposed adaptive controller. The obtained results may have important implications in system identification in neural science.

Topology identification of uncertain nonlinearly coupled complex networks with delays based on anticipatory synchronization.

This paper presents an adaptive anticipatory synchronization based method for simultaneous identification of topology and parameters of uncertain nonlinearly coupled complex dynamical networks with time delays. An adaptive controller is proposed, based on Lyapunov stability theorem and Barbǎlat's Lemma, to guarantee the stability of the anticipatory synchronization manifold between drive and response networks. Meanwhile, not only the identification criteria of network topology and system parameters are obtained but also the anticipatory time is identified. Numerical simulation results illustrate the effectiveness of the proposed method.

A deep learning framework for identifying Alzheimer's disease using fMRI-based brain network.

Background: The convolutional neural network (CNN) is a mainstream deep learning (DL) algorithm, and it has gained great fame in solving problems from clinical examination and diagnosis, such as Alzheimer's disease (AD). AD is a degenerative disease difficult to clinical diagnosis due to its unclear underlying pathological mechanism. Previous studies have primarily focused on investigating structural abnormalities in the brain's functional networks related to the AD or proposing different deep learning approaches for AD classification. Objective: The aim of this study is to leverage the advantages of combining brain topological features extracted from functional network exploration and deep features extracted by the CNN. We establish a novel fMRI-based classification framework that utilizes Resting-state functional magnetic resonance imaging (rs-fMRI) with the phase synchronization index (PSI) and 2D-CNN to detect abnormal brain functional connectivity in AD. Methods: First, PSI was applied to construct the brain network by region of interest (ROI) signals obtained from data preprocessing stage, and eight topological features were extracted. Subsequently, the 2D-CNN was applied to the PSI matrix to explore the local and global patterns of the network connectivity by extracting eight deep features from the 2D-CNN convolutional layer. Results: Finally, classification analysis was carried out on the combined PSI and 2D-CNN methods to recognize AD by using support vector machine (SVM) with 5-fold cross-validation strategy. It was found that the classification accuracy of combined method achieved 98.869%. Conclusion: These findings show that our framework can adaptively combine the best brain network features to explore network synchronization, functional connections, and characterize brain functional abnormalities, which could effectively detect AD anomalies by the extracted features that may provide new insights into exploring the underlying pathogenesis of AD.

A novel framework of MOPSO-GDM in recognition of Alzheimer's EEG-based functional network.

Background: Most patients with Alzheimer's disease (AD) have an insidious onset and frequently atypical clinical symptoms, which are considered a normal consequence of aging, making it difficult to diagnose AD medically. But then again, accurate diagnosis is critical to prevent degeneration and provide early treatment for AD patients. Objective: This study aims to establish a novel EEG-based classification framework with deep learning methods for AD recognition. Methods: First, considering the network interactions in different frequency bands (δ, θ, α, ß, and γ), multiplex networks are reconstructed by the phase synchronization index (PSI) method, and fourteen topology features are extracted subsequently, forming a high-dimensional feature vector. However, in feature combination, not all features can provide effective information for recognition. Moreover, combining features by manual selection is time-consuming and laborious. Thus, a feature selection optimization algorithm called MOPSO-GDM was proposed by combining multi-objective particle swarm optimization (MOPSO) algorithm with Gaussian differential mutation (GDM) algorithm. In addition to considering the classification error rates of support vector machine, naive bayes, and discriminant analysis classifiers, our algorithm also considers distance measure as an optimization objective. Results: Finally, this method proposed achieves an excellent classification error rate of 0.0531 (5.31%) with the feature vector size of 8, by a ten-fold cross-validation strategy. Conclusion: These findings show that our framework can adaptively combine the best brain network features to explore network synchronization, functional interactions, and characterize brain functional abnormalities, which can improve the recognition efficiency of diseases. While improving the classification accuracy of application algorithms, we aim to expand our understanding of the brain function of patients with neurological disorders through the analysis of brain networks.

Parameter estimation of the FitzHugh-Nagumo model using noisy measurements for membrane potential.

This paper proposes an identification method to estimate the parameters of the FitzHugh-Nagumo (FHN) model for a neuron using noisy measurements available from a voltage-clamp experiment. By eliminating an unmeasurable recovery variable from the FHN model, a parametric second order ordinary differential equation for the only measurable membrane potential variable can be obtained. In the presence of the measurement noise, a simple least squares method is employed to estimate the associated parameters involved in the FHN model. Although the available measurements for the membrane potential are contaminated with noises, the proposed identification method aided by wavelet denoising can also give the FHN model parameters with satisfactory accuracy. Finally, two simulation examples demonstrate the effectiveness of the proposed method.

Estimating and adjusting abnormal networks with unknown parameters and topology.

The changes of parameters and topology in a complex network often lead to unexpected accidents in complex systems, such as diseases in neural systems and unexpected current in circuit system, so the methods of adjusting the abnormal network back to its normal conditions are necessary to avoid these problems. However, it is not easy to detect the structures and information of each network, even if we can find a network which has the same function as the abnormal network, it is still hard to use it as a reference to adjust the abnormal network because a lot of network information is unknown. In this paper, we design a "bridging network" as an information bridge between a normal network and an abnormal network to estimate and control the abnormal network. Through the "bridging network" and some adaptive laws, the abnormal parameters and connections in abnormal network can be adjusted to the same conditions as those of the normal network which is chosen as a reference model. Finally, the "bridging network" and the abnormal network achieve synchronization with the normal network. Besides, the detailed inner information in normal network and abnormal network can be accurately estimated by this "bridging network." Finally, the nodes in the abnormal network will behave normally after the correction. In this paper, we use Hindmarsh-Rose model as an example to describe our method.

A Data Driven Experimental System for Individualized Brain Stimulation Design and Validation.

Deep brain stimulation (DBS) is an effective clinical treatment for epilepsy. However, the individualized setting and adaptive adjustment of DBS parameters are still facing great challenges. This paper investigates a data-driven hardware-in-the-loop (HIL) experimental system for closed-loop brain stimulation system individualized design and validation. The unscented Kalman filter (UKF) is utilized to estimate critical parameters of neural mass model (NMM) from the electroencephalogram recordings to reconstruct individual neural activity. Based on the reconstructed NMM, we build a digital signal processor (DSP) based virtual brain platform with real time scale and biological signal level scale. Then, the corresponding hardware parts of signal amplification detection and closed-loop controller are designed to form the HIL experimental system. Based on the designed experimental system, the proportional-integral controller for different individual NMM is designed and validated, which proves the effectiveness of the experimental system. This experimental system provides a platform to explore neural activity under brain stimulation and the effects of various closed-loop stimulation paradigms.

Spike-sorting analysis of neural electrical signals evoked by acupuncture based on model.

Acupuncturing the Zusanli (ST 36) point with different types of manual acupuncture manipulations (MAs) and different frequencies can evoke a lot of neural response activities in spinal dorsal root neurons. The action potential is the basic unit of communication in the neural response process. With the rapid development of the electrode acquisition technology, we can simultaneously obtain neural electrical signals of multiple neurons in the target area. So it is crucial to extract spike trains of each neuron from raw recorded data. To solve the problem of variability of the spike waveform, this paper adopts a optimization algorithm based on model to improve the wave-cluster algorithm, which can provide higher accuracy and reliability. Further, through this optimization algorithm, we make a statistical analysis on spike events evoked by different MAs. Results suggest that numbers of response spikes under reinforcing manipulations are far more than reducing manipulations, which mainly embody in synchronous spike activities.

Effects of network topologies on stochastic resonance in feedforward neural network.

The effects of network topologies on signal propagation are studied in noisy feedforward neural network in detail, where the network topologies are modulated by changing both the in-degree and out-degree distributions of FFNs as identical, uniform and exponential respectively. Stochastic resonance appeared in three FFNs when the same external stimuli and noise are applied to the three different network topologies. It is found that optimal noise intensity decreases with the increase of network's layer index. Meanwhile, the Q index of FFN with identical distribution is higher than that of the other two FFNs, indicating that the synchronization between the neuronal firing activities and the external stimuli is more obvious in FFN with identical distribution. The optimal parameter regions for the time cycle of external stimuli and the noise intensity are found for three FFNs, in which the resonance is more easily induced when the parameters of stimuli are set in this region. Furthermore, the relationship between the in-degree, the average membrane potential and the resonance performance is studied at the neuronal level, where it is found that both the average membrane potentials and the Q indexes of neurons in FFN with identical degree distribution is more consistent with each other than that of the other two FFNs due to their network topologies. In summary, the simulations here indicate that the network topologies play essential roles in affecting the signal propagation of FFNs.

Comments and Corrections Corrections to "Application of Reinforcement Learning to Deep Brain Stimulation in a Computational Model of Parkinson's Disease".

In the above article [1], financial support was incorrectly published. The correct information is as follows: This work was supported in part by the National Natural Science Foundation of China under Grant 61501330 and Grant 61771330, and in part by the Tianjin Municipal Special Program of Talents Development for Excellent Youth Scholars under Grant TJTZJH-QNBJRC-2-2.

Training Spiking Neural Networks for Cognitive Tasks: A Versatile Framework Compatible With Various Temporal Codes.

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.

Application of Reinforcement Learning to Deep Brain Stimulation in a Computational Model of Parkinson's Disease.

Deep brain stimulation (DBS) has been proven to be an effective treatment to deal with the symptoms of Parkinson's disease (PD). Currently, the DBS is in an open-loop pattern with which the stimulation parameters remain constant regardless of fluctuations in the disease state, and adjustments of parameters rely mostly on trial and error of experienced clinicians. This could bring adverse effects to patients due to possible overstimulation. Thus closed-loop DBS of which stimulation parameters are automatically adjusted based on variations in the ongoing neurophysiological signals is desired. In this paper, we present a closed-loop DBS method based on reinforcement learning (RL) to regulate stimulation parameters based on a computational model. The network model consists of interconnected biophysically-based spiking neurons, and the PD state is described as distorted relay reliability of thalamus (TH). Results show that the RL-based closed-loop control strategy can effectively restore the distorted relay reliability of the TH but with less DBS energy expenditure.

Model Predictive Control for Seizure Suppression Based on Nonlinear Auto-Regressive Moving-Average Volterra Model.

This article investigates a closed-loop brain stimulation method based on model predictive control strategy to suppress epileptic seizures. A neural mass model (NMM), exhibiting the normal and various epileptic seizures by changing physiologically meaningful parameters, is used as a black-box model of the brain. Based on system identification, an auto-regressive moving-average Volterra model is established to reveal the relationship between stimulation and neuronal responses. Then, the model predictive control strategy is implemented based the Volterra model, which can generate an optimal stimulation waveform to eliminate epileptiform waves. The computational simulation results indicate the proposed closed-loop control strategy can optimize the stimulation waveform without particular knowledge of the physiological properties of the system. The robustness of the proposed control strategy to system disturbances makes it more appropriate for future clinical application.

Spiking Correlation Analysis of Synchronous Spikes Evoked by Acupuncture Mechanical Stimulus.

Acupuncturing the ST36 acupoint can evoke the response of the sensory nervous system, which is translated into output electrical signals in the spinal dorsal root. Neural response activities, especially synchronous spike events, evoked by different acupuncture manipulations have remarkable differences. In order to identify these network collaborative activities, we analyze the underlying spike correlation in the synchronous spike event. In this paper, we adopt a log-linear model to describe network response activities evoked by different acupuncture manipulations. Then the state-space model and Bayesian theory are used to estimate network spike correlations. Two sets of simulation data are used to test the effectiveness of the estimation algorithm and the model goodness-of-fit. In addition, simulation data are also used to analyze the relationship between spike correlations and synchronous spike events. Finally, we use this method to identify network spike correlations evoked by four different acupuncture manipulations. Results show that reinforcing manipulations (twirling reinforcing and lifting-thrusting reinforcing) can evoke the third-order spike correlation but reducing manipulations (twirling reducing and lifting-thrusting reducing) does not. This is the main reason why synchronous spikes evoked by reinforcing manipulations are more abundant than reducing manipulations.

Scalable Digital Neuromorphic Architecture for Large-Scale Biophysically Meaningful Neural Network With Multi-Compartment Neurons.

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.