Time-frequency Analysis Of EEG Signals Research Articles

A new hybrid and optimized algorithm for drivers’ drowsiness detection

When the roads are monotonous, especially on the highways, the state of vigilance decreases and the state of drowsiness appears. Drowsiness is defined as the transitional phase from the awake to the sleepy state. However, In Morocco, the majority of fatal accidents on the highway are caused by drowsiness at the wheel, reaching 33.33% rate. Therefore, we proposed the conception and realization of an automatic method based on electroencephalogram (EEG) signals that can predict drowsiness in real time. The proposed work is based on time-frequency analysis of EEG signals from a single channel (FP1-Ref), and drowsiness is predicted using a personalized and optimized machine learning model (optimized decision tree classification method) under Python. The results are much significant and optimized, improving the accuracy from 95.7% to 96.4% and a time consuming from 0.065 to 0.053 seconds.

Evaluating the Different Stages of Parkinson's Disease Using Electroencephalography With Holo-Hilbert Spectral Analysis.

Electroencephalography (EEG) can reveal the abnormalities of dopaminergic subcortico-cortical circuits in patients with Parkinson’s disease (PD). However, conventional time-frequency analysis of EEG signals cannot fully reveal the non-linear processes of neural activities and interactions. A novel Holo-Hilbert Spectral Analysis (HHSA) was applied to reveal non-linear features of resting state EEG in 99 PD patients and 59 healthy controls (HCs). PD patients demonstrated a reduction of β bands in frontal and central regions, and reduction of γ bands in central, parietal, and temporal regions. Compared with early-stage PD patients, late-stage PD patients demonstrated reduction of β bands in the posterior central region, and increased θ and δ2 bands in the left parietal region. θ and β bands in all brain regions were positively correlated with Hamilton depression rating scale scores. Machine learning algorithms using three prioritized HHSA features demonstrated “Bag” with the best accuracy of 0.90, followed by “LogitBoost” with an accuracy of 0.89. Our findings strengthen the application of HHSA to reveal high-dimensional frequency features in EEG signals of PD patients. The EEG characteristics extracted by HHSA are important markers for the identification of depression severity and diagnosis of PD.

Predicting seizure onset based on time-frequency analysis of EEG signals

Abstract Epilepsy is one of the chronic diseases of brain that occurs as a result of sudden and abnormal change in electric waves of brain. A lot of research has been carried out to predict the epileptic seizures. However, the literature shows that there are opportunities for further improvement in early seizure prediction. Therefore, in this paper, we aim to predict a seizure in advance with high prediction accuracy from EEG signals by using time-frequency feature extraction and classification techniques. An early prediction of an epileptic seizure will help clinicians and carers to properly intervene through medication or other preventive measures. Features from EEG signals were extracted using Discrete Wavelet Transformation (DWT) and then Support Vector Machine (SVM) and post-processing technique were applied to predict seizure in advance. We evaluated the performance of our proposed method in terms of sensitivity, specificity, and classification/prediction accuracy. The average prediction accuracy of our proposed method on ten patients was observed to be 96.38% and our method was able to predict a seizure 26.1 min before the actual seizure occurrence on average. The performance comparison of our proposed method with some existing techniques applied on similar datasets demonstrates that our proposed method can be an effective technique for the prediction of seizures. The findings project the towards this method’s possibility of future application in seizure prediction.

Time-Frequency Analysis of EEG Signals and GLCM Features for Depth of Anesthesia Monitoring.

One of the important tasks in the operating room is monitoring the depth of anesthesia (DoA) during surgery, and noninvasive techniques are very popular. Hence, we propose a new scheme for DoA monitoring considering the time-frequency analysis of electroencephalography (EEG) signals and GLCM features extracted from them. To this end, at first, the time-frequency map (TFM) of each channel of each EEG is computed by smoothed pseudo-Wigner–Ville distribution (SPWVD), where the EEG signal used in this paper is recorded in 15 channels. After that, we consider the gray-level co-occurrence matrix (GLCM) to obtain the content of TFM, and after that, four features such as homogeneity, correlation, energy, and contrast are obtained for each GLCM. Finally, after the selection of efficient features using the minimum redundancy maximum relevance (MRMR) method, the K-nearest neighbor (KNN) classifier is utilized to determine the DoA. Here, we consider the three states, namely, deep hypnotic, surgical anesthesia, and sedation and awake states according to bispectral index (BIS), and each EEG epoch is classified to these states. We also employ data augmentation to enhance the training phase and increase accuracy. We obtain the accuracy and confusion matrix of the proposed method. We also analyze the effects of a number of gray levels of GLCM, distance measure in KNN classifier, and parameters of data augmentation on the performance of the proposed method. Results indicate the efficiency of the proposed method to determine the DoA during surgery.

Time-frequency Analysis of Electroencephalogram Signals in a Cognitive Decision-making Task.

Choices and decisions involve a series of complex cognitive processes, and the time-frequency analysis of electroencephalogram (EEG) signals can help understand the brain activities in different cognitive tasks. In this study, a decision-making cognitive task of rock-paper-scissors was designed, and the complex decision-making task was divided into three stages (decision planning, confirmation, and feedback). 64 channels of EEG signals were simultaneously recorded using the Neuroscan QuikCap system during the whole task. The average spectral power and phase synchronization values of each frequency band (delta, theta, alpha and beta band) were extracted and compared within and across different stages. The results showed the desire to win or not to lose within the first stage might be accompanied by the increase of alpha and theta components. In the second stage, the spectral power inhibition of alpha wave and phase synchronization increase of delta wave indicated that subjects would improve their attentions when they confirmed their choice. In the third stage, the theta increased effect and alpha decreased effect were related to different feedback. This pilot study suggests that the time-frequency analysis of EEG signals could be a great tool to visualize the brain activities in response to different cognitive tasks.

Does the motor system contribute to the perception of changes in objects visual attributes? The neural dynamics of sensory binding by action

The contribution of the motor system to perception has been highlighted in research investigating the effect of performing an action on the conscious processing of information received from the sensory systems. For example, the perceptual temporal asynchrony observed when passively reporting changes in visual object attributes (e.g., colour and position) was found to disappear virtually when the changes resulted from a voluntary motor action. Although the spatio-temporal constraints of sensory binding by action have been broadly investigated, the underlying neural correlates are still largely unknown. In the present study, participants performed temporal order judgments of changes in the colour and position of a visual stimulus, while adapting to a 750 ms delay between a sound (perceptual condition) or the end of a manual reaching action (motor condition), and the visual changes. Behavioural observations indicated that temporal asynchrony (-30.2 ms) decreased in the motor condition (2.7 ms), as a result of sensorimotor adaptation, but not in the perceptual condition (-29.6 ms). EEG-evoked potentials on posterior visual regions showed that early components were altered by sensorimotor adaptation, with in particular a broad reduction in the amplitude of the early P1 component. Furthermore, time-frequency analysis of EEG signals during the 350 ms period preceding the visual changes revealed an increase of the 15–25 Hz frequency band amplitude in the central region and a decrease of the 8–12 Hz frequency band amplitude in the posterior region. Overall the results suggest that sensory binding by action depends on an early top-down modulation of the visual regions by the motor system - in agreement with the pre-activation theory of action-perception coupling - associated with an increase of attentional resources.

Time-frequency series based movement imagery classification

This paper proposes a novel pre-processing and feature extraction method for movement imagery classification based on Electroencephalogram (EEG) signal for Brain-Computer Interface (BCI). The proposed pre-processing method considers the time-frequency analysis of EEG signal. The Short Time Fourier Transforms (STFT) based time-frequency analysis is applied to generate the Time-Frequency Series (TFS) of Electrooculogram (EOG) corrected EEG signal. Further, the generated TFS of particular frequency bin is transformed into time domain signal by Inverse Short Time Fourier Transforms (ISTFT). The generated time domain signal of particular frequency bin is passed through feature extraction method. An asymmetry coefficient based on Hjorth parameter as a feature extracted for each frequency bin and further subjected to Support Vector Machine (SVM) classifier for their classification. The proposed method of EEG signal pre-processing and feature extraction outperforms the conventional method of EEG signal pre-processing like Discriminative Frequency Band Common Spatial Pattern (DFBCSP).

High-resolution time-frequency analysis of EEG signals using multiscale radial basis functions

An efficient time-varying autoregressive (TVAR) modeling approach using the multiscale radial basis functions (MRBF) method is presented for nonstationary signal processing, with applications to time-frequency analysis of electroencephalogram (EEG). In this new parametric modeling framework, the time-varying coefficients in the TVAR model are approximated by using MRBF that can better identify time-varying parameters with a variety of dynamic processes in nonstationary signals. Thus, the time-varying modeling problem is simplified to optimal scale determination of MRBF and parameter estimation, which can be effectively resolved by a modified particle swarm optimization (PSO) method and an ordinary least square (OLS) algorithm, respectively. To evaluate the performance of the proposed approach, a comparison with recursive least squares (RLS) and the Legendre polynomials expansion method for a synthesized EEG signal is performed. Results demonstrated that the proposed approach could indeed provide optimal time–frequency resolution as compared to RLS and Legendre polynomials expansion. The new TVAR modeling approach was also applied to the analysis of experimental EEG signals to demonstrate the performance of the proposed method.

Time-frequency analysis of EEG signals evoked by voluntary, stimulated and imaginary motions

In order to investigate the characteristics of sensorimotor cortex during motor execution (ME), voluntary, stimulated and imaginary finger flexions were performed by ten volunteer subjects. Electroencephalogram (EEG) data were recorded according to the modified 10–20 International EEG System. The patterns were compared by the analysis of the motion-evoked EEG signals focusing on the contralateral (C3) and ipsilateral (C4) channels for hemispheric differences. The EEG energy distributions at alpha (8–13 Hz), beta (14–30 Hz) and gamma (30–50 Hz) bands were computed by wavelet transform (WT) and compared by the analysis of variance (ANOVA). The timefrequency (TF) analysis indicated that there existed a contralateral dominance of alpha post-movement event-related synchronization (ERS) pattern during the voluntary task, and that the energy of alpha band increased in the ipsilateral area during the stimulated (median nerve of wrist) task. Besides, the contralateral alpha and beta event-related desynchronization (ERD) patterns were observed in both stimulated and imaginary tasks. Another significant difference was found in the mean power values of gamma band(p<0.01)between the imaginary and other tasks. The results show that significant hemispheric differences such as alpha and beta band EEG energy distributions and TF changing phenomena (ERS/ERD) were found between C3 and C4 areas during all of the three patterns. The largest energy distribution was always at the alpha band for each task.