基于脑功能网络和样本熵的脑电信号特征提取

针对脑-机接口(BCI)研究中采用单一特征对运动想象脑电信号(EEG)识别率不高的问题,该文提出一种结合脑功能网络和样本熵的特征提取方法。根据事件相关同步/去同步(ERS/ERD)现象以及皮层与肢体运动想象间的对侧映射机制,选取小波包变换消噪重构后的\begin{document}$ \mu$\end{document}节律脑电信号,用左侧27个通道、右侧27个通道分别对左半球脑区和右半球脑区构建脑功能网络,计算网络的平均节点度和平均聚集系数作为运动想象的脑功能网络特征,并结合C3, C4通道节律的样本熵构筑分布性和指向性相结合的特征向量。选用支持向量机(SVM)对左右手运动想象脑电信号进行分类,结果表明基于脑功能网络和样本熵的特征提取方法能够实现更优的分类效果,分类准确率最高可达90.27%。     

A comparison of algorithms for detection of spikes in the electroencephalogram

Identification of the short transient waveform, called a spike, in the cortical electroencephalogram (EEG) plays an important role during diagnosis of neurological disorders such as epilepsy. It has been suggested that artificial neural networks (ANN) can be employed for spike detection in the EEG, if suitable features are provided as input to an ANN. In this paper, we explore the performance of neural network-based classifiers using features selected by algorithms suggested by four previous investigators. Of these, three algorithms model the spike by mathematical parameters and use them as features for classification while the fourth algorithm uses raw EEG to train the classifier. The objective of this paper is to examine if there is any inherent advantage to any particular set of features, subject to the condition that the same data are used for all feature selection algorithms. Our results suggest that artificial neural networks trained with features selected using any one of the above three algorithms as well as raw EEG directly fed to the ANN will yield similar results.     

Brain source localization using reduced EEG sensors

Brain source activation is caused due to certain mental or physical task, and such activation is localized by using various optimization techniques. This localization has vital application for diagnoses of various brain disorders such as epilepsy, schizophrenia, Alzheimer, depression, Parkinson and stress. Various neuroimaging techniques (such as EEG, fMRI, MEG) are used to record brain activity for inference and estimation of active source locations. EEG employs set of sensors which are placed on scalp to measure electric potentials. These sensors have significant role in overall system complexity, computational time and system cost. Hence, sensor reduction for EEG source localization has been a topic of interest for researchers to develop a system with improved localization precision, less system complexity and reduced cost. This research work discusses and implements the brain source localization for real-time and synthetically generated EEG dataset with reduced number of sensors. For this, various optimization algorithms are used which include Bayesian framework-based multiple sparse priors (MSP), classical low-resolution brain electromagnetic tomography (LORETA), beamformer and minimum norm estimation (MNE). The results obtained are then compared in terms of negative variational free energy, localization error and computational time measured in seconds. It is observed that multiple sparse priors (MSP) with increased number of patches performed best even with reduced number of sensors, i.e., 7 instead of 74. The results are shown valid for synthetic EEG data at low SNR level, i.e., 5 dB and real-time EEG data, respectively.     

基于核典型相关分析的语音情感识别研究

语音是人类表达思想和感情交流最重要的工具,是人类文化的重要组成部分。语音情感识别作为情感计算中的重要课题已经成为国际上的研究热点,受到越来越多的关注。已有神经科学研究表明,大脑是产生调节情感的物质基础。因此,在语音情感的研究中,我们不能仅考虑语音信号自身,还应将大脑的活动信号融入语音情感识别中,以实现更高准确率的情感识别。基于上述思想,本文提出了一种基于核典型相关分析（KCCA）的语音特征提取方法。该方法将语音特征与脑电图（EEG）特征映射到高维希尔伯特空间,并计算二者的最大相关系数。KCCA将语音特征在高维希尔伯特空间上向与脑电特征相关性最大的方向投影,最终得到包含脑电信息的语音特征。本文方法将与语音情感相关的脑电信息融入语音情感特征提取中,所提特征能够更准确的表征情感。同时,本方法在理论上具有良好的可迁移性,当所提脑电特征足够准确与具有代表性时,KCCA建模得到的投影向量具有通用性,可直接用于新的语音情感数据集中而无需重新采集和计算相应的脑电信号。在自建语音情感数据库与公开语音情感数据库MSP-IMPROV上的实验结果表明,使用投影语音特征进行语音情感分类的方法优于使用原始音频特征...     

Predicting Reaching Targets from Human EEG

Extracting reach information from brain signals is of great interest to the fields of brain-computer interfaces (BCIs) and human motor control. To date, most work in this area has focused on invasive intracranial recordings; however, successful decoding of reach targets from noninvasive electroencephalogram (EEG) signals would be of great interest. In this article, we show that EEG signals contain sufficient information to decode target location during a reach (Experiment 1) and during the planning period before a reach (Experiment 2). We discuss the application of independent component analysis and dipole fitting for removing movement artifacts. With this technique we get similar classification accuracy for classifying EEG signals during a reach (Experiment 1) and during the planning period before a reach (Experiment 2). To the best of our knowledge, this is the first demonstration of decoding (planned) reach targets from EEG. These results lay the foundation for future EEG-based BCIs based on decoding of planned reaches.     

Automatic selection of frequency and time intervals for classification of EEG signals

A novel technique is presented for the automatic selection of time and frequency intervals to be used in feature extraction on multidimensional signals acquired by an electroencephalogram (EEG). This technique is completely automatic, adaptive (task independent), and does not require any specific prior domain knowledge. Experimental results obtained by integrating the proposed technique in a system for brain computer interface (BCI) confirm its effectiveness.     

Multiview Classification and Dimensionality Reduction of Scalp and Intracranial EEG Data through Tensor Factorisation

Electroencephalography (EEG) signals arise as mixtures of various neural processes which occur in particular spatial, frequency, and temporal brain locations. In classification paradigms, algorithms are developed that can distinguish between these processes. In this work, we apply tensor factorisation to a set of EEG data from a group of epileptic patients and factorise the data into three modes; space, time, and frequency with each mode containing a number of components or signatures. We train separate classifiers on various feature sets corresponding to complementary combinations of those modes and components and test the classification accuracy for each set. The relative influence on the classification accuracy of the respective spatial, temporal, or frequency signatures can then be analysed and useful interpretations can be made. Additionaly, we show that through tensor factorisation we can perform dimensionality reduction by evaluating the classification performance with regards to the number of components in each mode and also by rejecting components with insignificant contribution to the classification accuracy.     

A feature selection method for multilevel mental fatigue EEG classification

Two feature selection approaches for multilevel mental fatigue electroencephalogram (EEG) classification are presented in this paper, in which random forest (RF) is combined with the heuristic initial feature ranking scheme (INIT) or with the recursive feature elimination scheme (RFE). In a "leave-one-proband-out" evaluation strategy, both feature selection approaches are evaluated on the recorded mental fatigue EEG time series data from 12 subjects (each for a 25-h duration) after initial feature extractions. The latter of the two approaches performs better both in classification performance and more importantly in feature reduction. RF with RFE achieved its lowest test error rate of 12.3% using 24 top-ranked features, whereas RF with INIT reached its lowest test error rate of 15.1% using 64 top-ranked features, compared to a test error rate of 22.1% using all 304 features. The results also show that 17 key features (out of 24 top-ranked features) are consistent between the subjects using RF with RFE, which is superior to the set of 64 features as determined by RF with INIT.     

基于神经网络的光标上下移动EEG信号分类方法

脑电信号（EEG）是研究脑活动的一种重要的信息来源,基于脑电信号的人与计算机的通信已成为一种新的人机接口方式。在此主要通过时域回归方法对BCIⅡ竞赛数据进行EEG信号去噪预处理,运用6阶AR参数提取脑电特征作为神经网络的输入,最后用Matlab 7.0进行仿真,得到分类正确率为90%。实验表明,该方法可以达到很好的分类效果。     

抑郁症正性情绪加工脑电样本熵异常

本文通过研究抑郁症患者与正常人在处理不同情绪刺激时脑电信号样本熵的差异,探索抑郁症患者情绪加工异常的电生理机制。我们招募了16名抑郁症患者和14名健康对照组参与面部表情空间搜索任务,同时采集了他们完成任务时的头皮脑电信号。我们首先选用希尔伯特-黄变换获取脑电的各频段活动;然后通过比较抑郁症患者与健康对照组脑电的样本熵来研究两组受试者不同情绪加工的电生理差异;最后选取β频段样本熵作为特征,采用不同分类器和不同提取方式进行分类研究。结果反映,抑郁症患者在情绪加工上,尤其是正性情绪的认知加工上存在异常。同时也表明样本熵在一定程度上可以反映不同条件情绪加工脑电的特异性,可作为一种区分正常人与抑郁症患者的潜在的特征指标,用于抑郁症患者的辅助分类识别,为医生诊断抑郁症患者提供一种辅助方案。      

基于fMRI功能网络和贝叶斯矩阵分解的脑电源成像方法

脑电(EEG)是一种重要的脑功能成像技术,根据头皮记录的EEG信号重构皮层脑活动称为EEG源成像。然而脑源活动位置和尺寸的准确重构依然是一个挑战。为充分利用EEG和功能磁共振(fMRI)信号在时空分辨率上的互补信息,该文提出一个新的源成像方法——基于fMRI脑网络和时空约束的EEG源重构算法(FN-STCSI)。该方法在参数贝叶斯框架下,基于矩阵分解思想将源信号分解为若干时间基函数的线性组合。此外,为融合fMRI的高空间分辨率信息,FN-STCSI利用独立成分分析提取fMRI信号的功能网络,构建EEG源成像的空间协方差基,通过变分贝叶斯推断技术确定每个空间协方差基的相对贡献,实现EEG-fMRI融合。通过蒙特卡罗数值仿真和实验数据分析比较了FN-STCSI与现有算法在不同信噪比和不同先验条件下的性能,结果表明FN-STCSI能有效融合EEG-fMRI在时空上的互补信息,提高EEG弥散源成像的性能。     

基于深度时空特征融合的多通道运动想象EEG解码方法

脑电(EEG)是一种在临床上广泛应用的脑信息记录形式,其反映了脑活动中神经细胞放电产生的电场变化情况。脑电广泛应用于脑-机接口(BCI)系统。然而,研究表明脑电信息空间分辨率较低,这种缺陷可以综合分析多通道电极的脑电数据来弥补。为了从多通道数据中高效地获取到与运动想象任务相关的辨识特征,该文提出一种针对多通道脑电信息的卷积神经网络(MC-CNN)解码方法,先对预先选取好的多通道数据预处理后送入2维卷积神经网络(CNN)进行时间-空间特征提取,然后利用自动编码(AE)器把这些特征映射为具有辨识度的特征子空间,最后指导识别网络进行分类识别。实验结果表明,该文所提多通道空间特征提取和构建方法在运动想象脑电任务识别性能和效率上都具有较大优势。     

Single-shot MR imaging using trapezoidal-gradient-based Lissajous trajectories

A novel single-shot trapezoidal-gradient-based Lissajous trajectory is described and implemented on a 3-tesla magnetic resonance (MR) scanner. A feature of this trajectory is that its sampling points are located on a nonequidistant rectangular grid, which permits the usage of one-dimensional optimal algorithms to increase the robustness and speed of image reconstruction. Another advantage of the trajectory is that two images with different effective echo times can be obtained within a single excitation, which might be used for fast T2* mapping, in functional MR imaging scanning of brain activity associated with mental processes. Potential artifacts in reconstructed images were investigated and methods for suppressing these artifacts were developed. Experiments on normal subjects at rest and during brain activation were performed to demonstrate the feasibility of the new sequence.     

用核学习算法的意识任务特征提取与分类

介绍了核学习算法中核主分量分析(KPCA)和支持向量机(SVM)的基本原理,给出一种推广误差上界估计判据,实现了SVM核参数及惩罚因子的优化选取.根据多变量自回归模型理论对4个受试对象、三种不同意识任务的脑电信号进行特征提取,并利用KPCA方法进行降维预处理,对SVM进行训练和分类测试.结果表明,KPCA算法在高维特征空间具有较强的特征选择能力,优化核参数的SVM的分类正确率明显高于径向基函数网络,三种意识任务的平均分类正确率达78.6％.     

Conditional Mutual Information‐Based Feature Selection Analyzing for Synergy and Redundancy

Battiti's mutual information feature selector (MIFS) and its variant algorithms are used for many classification applications. Since they ignore feature synergy, MIFS and its variants may cause a big bias when features are combined to cooperate together. Besides, MIFS and its variants estimate feature redundancy regardless of the corresponding classification task. In this paper, we propose an automated greedy feature selection algorithm called conditional mutual information‐based feature selection (CMIFS). Based on the link between interaction information and conditional mutual information, CMIFS takes account of both redundancy and synergy interactions of features and identifies discriminative features. In addition, CMIFS combines feature redundancy evaluation with classification tasks. It can decrease the probability of mistaking important features as redundant features in searching process. The experimental results show that CMIFS can achieve higher best‐classification‐accuracy than MIFS and its variants, with the same or less (nearly 50%) number of features.     

基于相关性和稀疏表示的运动想象脑电通道选择方法

在基于运动想象(MI)的脑机接口(BCI)中,通常采用较多通道的脑电信号(EEG)来提高分类精度,但其中会有包含与MI任务无关或冗余信息的通道,从而影响BCI的性能提升。该文针对运动想象脑电分类中的通道选择问题,提出一种采用相关性和稀疏表示对通道进行选择的方法(CSR-CS)。首先计算训练样本每个通道的皮尔逊相关系数来选择显著通道,然后提取显著通道所在区域的滤波器组共空间模式特征拼接成字典,利用由字典所得到的非零稀疏系数的个数表征每个区域的分类能力,选出显著区域所包含的显著通道作为最优通道,最后采用共空间模式和支持向量机分别进行特征提取与分类。在对BCI第3次竞赛数据集IVa和BCI第4次竞赛数据集I两个二分类MI任务的分类实验中,平均分类精度达到了88.61%和83.9%,表明所提通道选择方法的有效性和鲁棒性。     

卷积神经网络在脑疲劳检测中的研究

脑疲劳是由于持续进行脑力劳动导致的一种状态，脑电被认为是脑疲劳状态检测的最佳工具。如何选取合适的脑疲劳特征成为脑疲劳检测的关键问题，传统模式识别中手动提取特征会产生信息损失，针对脑电的时空特性，本文设计了具有时域卷积核、空间域卷积核的深层卷积神经网络和浅层卷积神经网络两种网络结构，将特征提取和状态分类合二为一，对正常态与疲劳态脑电数据进行分类，可视化了卷积神经网络的空间域卷积核。结果表明，浅层卷积神经网络平均分类正确率为98.868%，深层卷积神经网络平均分类正确率为98.217%，均高于传统分类方法，通过空间域卷积核的可视化，能够了解不同导联在网络中的参与程度，验证了该模型在脑疲劳检测任务中具有很高的有效性，同时为脑疲劳检测提供了新思路。      

Maximally Informative Feature and Sensor Selection in Pattern Recognition Using Local and Global Independent Component Analysis

In pattern recognition, a suitable criterion for feature selection is the mutual information (MI) between feature vectors and class labels. Estimating MI in high dimensional feature spaces is problematic in terms of computation load and accuracy. We propose an independent component analysis based MI estimation (ICA-MI) methodology for feature selection. This simplifies the high dimensional MI estimation problem into multiple one-dimensional MI estimation problems. Nonlinear ICA transformation is achieved using piecewise local linear approximation on partitions in the feature space, which allows the exploitation of the additivity property of entropy and the simplicity of linear ICA algorithms. Number of partitions controls the tradeoff between more accurate approximation of the nonlinear data topology and small-sample statistical variations in estimation. We test the ICA-MI feature selection framework on synthetic, UCI repository, and EEG activity classification problems. Experiments demonstrate, as expected, that the selection of the number of partitions for local linear ICA is highly problem dependent and must be carried out properly through cross validation. When this is done properly, the proposed ICA-MI feature selection framework yields feature ranking results that are comparable to the optimal probability of error based feature ranking and selection strategy at a much lower computational load.     

Performance evaluation of spatially multiplexed MIMO systems with subset antenna transmission in interference limited environments

Transmit antenna selection in spatially multiplexed multiple-input multiple-output (MIMO) systems is a low complexity low-rate feedback technique, which involves transmission of a reduced number of streams from the maximum possible to improve the error rate performance of linear receivers. It has been shown to be effective in enhancing the performance of single-user interference-free point-to-point MIMO systems. However, performance of transmit antenna selection techniques in interference-limited environments and over frequency selective channels is less well understood. In this paper, we investigate the performance of transmit antenna selection in spatially multiplexed MIMO systems in the presence of co-channel interference. We propose a transmission technique for the downlink of a cellular MIMO system that employs transmit antenna selection to minimize the effect of co-channel interference from surrounding cells. Several transmit antenna selection algorithms are proposed and their performance is evaluated in both frequency flat and frequency selective channels. Various antenna selection algorithms proposed in the literature for single user MIMO links are extended to a cellular scenario, where each user experiences co-channel interference from the other cells (intercell interference) in the system. For frequency selective channels, we consider orthogonal frequency division multiplexing (OFDM) with MIMO. We propose a selection algorithm that maximizes the average output SINR over all subcarriers. A method to quantify selection gain in frequency selective channel is discussed. The effect of delay spread on the selection gain is studied by simulating practical fading environments with different delay spreads. The effect of the variable signal constellation sizes and the number of transmitted streams on the bit error rate (BER) performance of the proposed system is also investigated in conjunction with the transmit antenna selection. Simulation results show that for low to moderate interference power, significant improvement in the system performance is achievable with the use of transmit antenna selection algorithms. Even though the gain due to selection in frequency selective channels is reduced compared to that in flat fading channels due to the inherent frequency diversity, the performance improvement is significant when the system is interference limited. The performance improvement due to reduced number of transmit streams at larger signal constellation sizes is found to be more significant in spatially correlated scenarios, and the gain due to selection is found to be reduced with the increased delay spread. It is found that employing transmit antenna selection algorithms in conjunction with adaptation of the number of transmitted streams and the signal constellation sizes can significantly improve the performance of MIMO systems with co-channel interference.     

Spatiotemporal emotion recognition based on 3D time-frequency domain feature matrix

The research of emotion recognition based on electroencephalogram (EEG) signals often ignores the relatedinformation between the brain electrode channels and the contextual emotional information existing in EEG signals,which may contain important characteristics related to emotional states. Aiming at the above defects, aspatiotemporal emotion recognition method based on a 3-dimensional (3D) time-frequency domain feature matrixwas proposed. Specifically, the extracted time-frequency domain EEG features are first expressed as a 3D matrixformat according to the actual position of the cerebral cortex. Then, the input 3D matrix is processed successivelyby multivariate convolutional neural network (MVCNN) and long short-term memory (LSTM) to classify theemotional state. Spatiotemporal emotion recognition method is evaluated on the DEAP data set, and achievedaccuracy of 87.58% and 88.50% on arousal and valence dimensions respectively in binary classification tasks, aswell as obtained accuracy of 84.58% in four class classification tasks. The experimental results show that 3D matrixrepresentation can represent emotional information more reasonably than two-dimensional (2D). In addition,MVCNN and LSTM can utilize the spatial information of the electrode channels and the temporal context information of the EEG signal respectively.