An approach to EEG-based emotion recognition using combined feature extraction method

Electroencephalogram (EEG) signal-based emotion classification is vital in the ever-growing human-computer interface (HCI) applications. However, the chaotic, non-stationary, and person-dependent nature of EEG signals often limits such practical applications. These challenges reduce the ability of the state-of-the-art approaches to effectively distinguish between different emotional states from EEG data, resulting in sub-optimal emotion recognition performance. This work presented a time-frequency (T-F) analysis of EEG signals to localize different EEG frequency rhythms responsible for emotion-related information in the EEG signals. In particular, this work investigated two T-F analysis domains for multichannel EEG signals based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and local mean decomposition (LMD) to extract entropy features that capture specific emotion-relevant traits from the specific EEG channels that are highly responsive to emotions. The CEEMDAN and LMD decompose the EEG signals into different EEG frequency rhythms called mode functions, intrinsic mode functions (IMFs), and product functions (PFs), respectively. Further, various types of entropy feature of these two categories of mode functions, such as approximate entropy (ApEn), sample entropy (SaEn), permutation entropy (PeEn), and bubble entropy (BuEn), are computed for extracting emotion-relevant distinguishing features. Entropy features help quantify EEG’s non-linear behavior and eventually help classify EEG-based emotions precisely. Emotion classification has been achieved using a grid search cross-validation (GSCV) optimized XGboost classifier. Thorough experimentations are conducted to validate the efficacy of the proposed approach on publically accessible datasets named Database for Emotion Analysis of Physiological Signals (DEAP), SJTU emotion EEG dataset (SEED), and SEED-IV. The efficacy of the proposed emotion recognition approach is measured in terms of widely used performance metrics such as accuracy, confusion metrics, receiver operating characteristics (RoC), and area under the curve (AuC). The highest average accuracy is attained using the proposed LMD-domain BuEn features, i.e., 97.8%, 98.6%, and 95.7% using SEED, SEED-IV, and DEAP databases, respectively, outperforming the recent state-of-the-art emotion recognition algorithms.

Emotion feature extraction is the key to speech emotional recognition. And ensemble empirical mode decomposition(EEMD) is a newly developed method aimed at eliminating emotion mode mixing present in the original empirical mode decomposition(EMD). To evaluate the performance of this new method, this paper investigates the effect of a parameters pertinent to EEMD: speech emotional envelope. Firstly, a speaker emotional envelope features extraction based on EEMD is proposed in the paper. Using the piecewise power function in speech emotional envelope has a better effect in emotional identification. At the same time, the proposed technique has been utilized for classification of fo ur kinds of emotional(angry, happy, sad and neutral) speech signals. Emotional intrinsic mode functions(IMFe) are obtained by empirical mode decomposition on emotional speech signals, the fast fourier transform(FFT) of each intrinsic mode function is extracte d as the emotional feature coefficient which is used in speaker emotional identification applying by vector quantization. MATLAB is used to ca lculate the characteristic of emotional speech signals using empirical mode decomposition (EEMD). We obtain an emotional envelope by transforming the IMFe of emotional speech signals, and obtain a new method of emotion recognition according to different emotional envelop feature vectors. The results indicate the proposed method works well in speaker emotional identification.

Emotion recognition from single-channel EEG signals using a two-stage correlation and instantaneous frequency-based filtering method

Emotion recognition by artificial intelligence (AI) is a challenging task. A wide variety of research has been done, which demonstrated the utility of audio, imagery, and electroencephalography (EEG) data for automatic emotion recognition. This paper presents a new automated emotion recognition framework, which utilizes electroencephalography (EEG) signals. The proposed method is lightweight, and it consists of four major phases, which include: a reprocessing phase, a feature extraction phase, a feature dimension reduction phase, and a classification phase. A discrete wavelet transforms (DWT) based noise reduction method, which is hereby named multi scale principal component analysis (MSPCA), is utilized during the pre-processing phase, where a Symlets-4 filter is utilized for noise reduction. A tunable Q wavelet transform (TQWT) is utilized as feature extractor. Six different statistical methods are used for dimension reduction. In the classification step, rotation forest ensemble (RFE) classifier is utilized with different classification algorithms such as k-Nearest Neighbor (k-NN), support vector machine (SVM), artificial neural network (ANN), random forest (RF), and four different types of the decision tree (DT) algorithms. The proposed framework achieves over 93 % classification accuracy with RFE + SVM. The results clearly show that the proposed TQWT and RFE based emotion recognition framework is an effective approach for emotion recognition using EEG signals.

Sleep disorders has a vital effect on mental depression and many other diseases of human body. Diagnosing the sleep disorder in an early curable stage may help to provide better treatment and save the life. The EEG (Electroencephalogram) signal is one of the most uses bio signal for capturing brain activities to detect and diagnosis the sleep disorders. Empirical mode decomposition (EMD) is an efficient time-frequency data analysis technique for diagnosing disease by analyzing EEG signal. However, it is a challenging issue to select the optimal intrinsic mode functions (IMFs) of Empirical mode decomposition (EMD) for extracting discriminant properties of EEG signals to diagnosis the sleep disorder. From this point of view, this paper presents a model to select optimal IMF of EMD for diagnosing the sleep disorder using EEG brain signal. In this proposed model, EMD is applied to decompose and analyze EEG signal for extracting biomarker/feature of sleep disorders. During the EMD decomposition process, different levels of IMF are extracted and features, i.e., Shannon Entropy, Spectral Entropy, Standard deviation, Skewness and Kurtosis are calculated from those IMFs for detecting the sleep disorders. In identification process, the multiclass support vector machine (MC-SVM) classification algorithm is used and sleep disorders are classified based on trained knowledge. Finally, the performance of proposed model is evaluated for different IMFs of EMD and find the optimal IMF for sleep disorder diagnosis. For evaluating the proposed model, a benchmark dataset including 4 types of data such as Apnea, REM, PLM and healthy subjects are used in experiment. According to the experimental result, the proposed model achieves the optimal classification performance for IMF 8, i.e., 93.24% average classification accuracy.

In this paper, a statistical analysis of electroencephalogram (EEG) signals is carried out in the empirical mode decomposition (EMD) domain using a publicly available benchmark EEG database. First, the intrinsic mode functions (IMF) are extracted from the EEG signals in the EMD domain. Next, the investigation was carried whether the Bessel k-form (BKF) probability density function (pdf) can appropriately model the IMFs extracted in EMD domain of the EEG signals. It is shown that on an average, the BKF pdf is a suitable prior to model the first five IMFs extracted from various types of EEG recordings. After that, it is shown that the BKF parameters can distinguish among the EEG signals at those five IMF levels quite well. The analysis is further confirmed through the p-values obtained by one way analysis of variance (ANOVA). Thus, the BKF parameters in the EMD domain may be used to characterize EEG signals and help the electroencephalographers in developing fast and effective classifiers for seizure seizure detection.

Emotion is a general term for a series of subjective cognitive experiences. When the human brain reacts to the relationship between external things and the subject’s needs, it produces our experience and feelings of external things. Emotional information plays an important role in life today. The application of emotional information has spread to all aspects of life, such as medical, educational, psychological, business and military fields. This paper summarizes all the feature extraction methods currently used in the research field of EEG-based emotion recognition, and introduces the definition and calculation methods of various feature extraction from the time domain, frequency domain, time-frequency domain, and space domain, respectively. After summarizing and summarizing, a method suitable for EEG-based emotion recognition feature extraction is obtained. At the same time, it is pointed out that the combination of hybrid feature extraction and deep learning will become the mainstream trend. It is hoped that subsequent researchers can gain a preliminary understanding of feature extraction methods suitable for emotion recognition by studying this article when they engage in EEG-based emotion research, and provide new ideas for subsequent research.

Depth of anaesthesia (DoA) is an important measure for assessing the degree to which the central nervous system of a patient is depressed by a general anaesthetic agent, depending on the potency and concentration with which anaesthesia is administered during surgery. We can monitor the DoA by observing the patient’s electroencephalography (EEG) signals during the surgical procedure. Typically high frequency EEG signals indicates the patient is conscious, while low frequency signals mean the patient is in a general anaesthetic state. If the anaesthetist is able to observe the instantaneous frequency changes of the patient’s EEG signals during surgery this can help to better regulate and monitor DoA, reducing surgical and post-operative risks. This paper describes an approach towards the development of a 3D real-time visualization application which can show the instantaneous frequency and instantaneous amplitude of EEG simultaneously by using empirical mode decomposition (EMD) and the Hilbert–Huang transform (HHT). HHT uses the EMD method to decompose a signal into so-called intrinsic mode functions (IMFs). The Hilbert spectral analysis method is then used to obtain instantaneous frequency data. The HHT provides a new method of analyzing non-stationary and nonlinear time series data. We investigate this approach by analyzing EEG data collected from patients undergoing surgical procedures. The results show that the EEG differences between three distinct surgical stages computed by using sample entropy (SampEn) are consistent with the expected differences between these stages based on the bispectral index (BIS), which has been shown to be quantifiable measure of the effect of anaesthetics on the central nervous system. Also, the proposed filtering approach is more effective compared to the standard filtering method in filtering out signal noise resulting in more consistent results than those provided by the BIS. The proposed approach is therefore able to distinguish between key operational stages related to DoA, which is consistent with the clinical observations. SampEn can also be viewed as a useful index for evaluating and monitoring the DoA of a patient when used in combination with this approach.

Accurate recognition and understating of human emotions is an essential skill that can improve the collaboration between humans and machines. In this vein, electroencephalogram (EEG)-based emotion recognition is considered an active research field with challenging issues regarding the analyses of the nonstationary EEG signals and the extraction of salient features that can be used to achieve accurate emotion recognition. In this paper, an EEG-based emotion recognition approach with a novel time-frequency feature extraction technique is presented. In particular, a quadratic time-frequency distribution (QTFD) is employed to construct a high resolution time-frequency representation of the EEG signals and capture the spectral variations of the EEG signals over time. To reduce the dimensionality of the constructed QTFD-based representation, a set of 13 time- and frequency-domain features is extended to the joint time-frequency-domain and employed to quantify the QTFD-based time-frequency representation of the EEG signals. Moreover, to describe different emotion classes, we have utilized the 2D arousal-valence plane to develop four emotion labeling schemes of the EEG signals, such that each emotion labeling scheme defines a set of emotion classes. The extracted time-frequency features are used to construct a set of subject-specific support vector machine classifiers to classify the EEG signals of each subject into the different emotion classes that are defined using each of the four emotion labeling schemes. The performance of the proposed approach is evaluated using a publicly available EEG dataset, namely the DEAPdataset. Moreover, we design three performance evaluation analyses, namely the channel-based analysis, feature-based analysis and neutral class exclusion analysis, to quantify the effects of utilizing different groups of EEG channels that cover various regions in the brain, reducing the dimensionality of the extracted time-frequency features and excluding the EEG signals that correspond to the neutral class, on the capability of the proposed approach to discriminate between different emotion classes. The results reported in the current study demonstrate the efficacy of the proposed QTFD-based approach in recognizing different emotion classes. In particular, the average classification accuracies obtained in differentiating between the various emotion classes defined using each of the four emotion labeling schemes are within the range of –. Moreover, the emotion classification accuracies achieved by our proposed approach are higher than the results reported in several existing state-of-the-art EEG-based emotion recognition studies.

EEG-based emotion recognition: A deep learning approach to brain region analysis

The measurement of brain electrical activity recorded as EEG signals finds most application in epilepsy. EEG waveforms carry information about the underlying neural system dynamics and show different features amongst epilepsy syndromes. In this research, empirical mode decomposition (EMD) and power spectrum were employed to extract the features from EEG dataset of healthy participants, and epilepsy patients with seizure and seizure free conditions. The recorded EEG signals are represented by 500 signal segments from 5 sets of different conditions. The sum of Intrinsic Mode Function (IMF) power spectrum components gave 10 features for 50 components, and 20 features for 25 components, which were used as the classification inputs for artificial neural networks and random forest classifiers. The classifications were carried out for 3, 4, and 5 classes. From the experiments, the highest average accuracy was obtained for 3 classes using 20 features of power spectrum from the sum of 6 IMFs. For use of 6 IMFs, the accuracies had the maximum values of 92.4%, 90.4%, and 78.6% for 3, 4, and 5 classes respectively. It also improved the accuracy significantly for 5 classes.

Various faults inevitably occur in mechanical systems and may result in unexpected failures. Hence, fault detection is critical to reduce unscheduled downtime and costly breakdowns. Empirical mode decomposition (EMD) is an adaptive time-frequency domain signal processing method, potentially suitable for nonstationary and/or nonlinear processes. However, the EMD method suffers from several problems such as mode mixing, defined as intrinsic mode functions (IMFs) with incorrect scales. In this paper, an ensemble noise-reconstructed EMD method is proposed to ameliorate the mode mixing problem and denoise IMFs for enhancing fault signatures. The proposed method defines the IMF components as an ensemble mean of EMD trials, where each trial is obtained by sifting signals that have been reconstructed using the estimated noise present in the measured signal. Unlike traditional denoising methods, the noise inherent in the input data is reconstructed and used to reduce the background noise. Furthermore, the reconstructed noise helps to project different scales of the signal onto their corresponding IMFs, instrumental in alleviating the mode mixing problem. Two critical issues concerned in the method, i.e., the noise estimation strategy and the number of EMD trials required for denoising are discussed. Furthermore, a comprehensive noise-assisted EMD method is proposed, which includes the proposed method and ensemble EMD (EEMD). Numerical simulations and experimental case studies on accelerometer data collected from an industrial shaving process are used to demonstrate and validate the proposed method. Results show that the proposed method can both detect impending faults and isolate multiple faults. Hence, the proposed method can act as a promising tool for mechanical fault detection.

Epilepsy is a neurological disorder that affects people of any age, which can be detected by Electroencephalogram (EEG) signals. This paper proposes a novel method called Volume of Phase Space Representation (VOPSR) to classify seizure and seizure-free EEG signals automatically. Primarily, the recorded EEG signal is disintegrated into several Intrinsic Mode Functions (IMFs) using the Empirical Mode Decomposition (EMD) method and the three-dimensional phase space have been reconstructed for the obtained IMFs. The volume is measured for the obtained 3D-PSR for different IMFs called VOPSR, which is used as a feature set for the classification of Epileptic seizure EEG signals. Support vector machine (SVM) is used as a classifier for the classification of epileptic and epileptic-free EEG signals. The classification performance of the proposed method is evaluated under different kernels such as Linear, Polynomial and Radial Basis Function (RBF) kernels. Finally, the proposed method outperforms noteworthy state-of-the-art classification methods in the context of epileptic EEG signals, achieving 99.13% accuracy (average) with the Linear, Polynomial, and RBF kernels. The proposed technique can be used to detect epilepsy from the EEG signals automatically without human intervention.

Using filter bank property to simplify the calculations of Empirical Mode Decomposition

Electroencephalogram (EEG) based affective brain-computer interface (a-BCI) systems are gaining interest among researchers in recent decades. a-BCI systems interpret/recognize human emotions using features extracted from the EEG signals. Hence, features play a crucial role in building EEG based emotion recognition models. As a result, analysis of feature extraction models becomes inevitable. In this work, we have proposed an analysis model for analyzing the feature extraction models for EEG based emotion recognition with the help of the DEAP dataset. Features were extracted using manual feature extraction techniques and the convolutional neural network. Several combinations of feature sets were given as input to classifiers and the results obtained were analyzed with various evaluation metrics. The proposed analysis model will help the researchers to choose the feature extraction model for emotion recognition.