Automatic recognition of alertness and drowsiness from EEG by an artificial neural network

The measurement of mental fatigue following an overnight on-call duty among doctors using electroencephalogram

This study aimed to measure the spectral power differences in the brain rhythms among a group of hospital doctors before and after an overnight on-call duty. Thirty-two healthy doctors who performed regular on-call duty in a tertiary hospital in Sarawak, Malaysia were voluntarily recruited into this study. All participants were interviewed to collect relevant background information, followed by a self-administered questionnaire using Chalder Fatigue Scale and electroencephalogram test before and after an overnight on-call duty. The average overnight sleep duration during the on-call period was 2.2 hours (p<0.001, significantly shorter than usual sleep duration) among the participants. The mean (SD) Chalder Fatigue Scale score of the participants were 10.8 (5.3) before on-call and 18.4 (6.6) after on-call (p-value < 0.001). The theta rhythm showed significant increase in spectral power globally after an overnight on-call duty, especially when measured at eye closure. In contrast, the alpha and beta rhythms showed reduction in spectral power, significantly at temporal region, at eye closure, following an overnight on-call duty. These effects are more statistically significant when we derived the respective relative theta, alpha, and beta values. The finding of this study could be useful for development of electroencephalogram screening tool to detect mental fatigue.

EEG driving fatigue detection based on log-Mel spectrogram and convolutional recurrent neural networks

Driver fatigue detection is one of the essential tools to reduce accidents and improve traffic safety. Its main challenge lies in the problem of how to identify the driver's fatigue state accurately. Existing detection methods include yawning and blinking based on facial expressions and physiological signals. Still, lighting and the environment affect the detection results based on facial expressions. In contrast, the electroencephalographic (EEG) signal is a physiological signal that directly responds to the human mental state, thus reducing the impact on the detection results. This paper proposes a log-Mel spectrogram and Convolution Recurrent Neural Network (CRNN) model based on EEG to implement driver fatigue detection. This structure allows the advantages of the different networks to be exploited to overcome the disadvantages of using them individually. The process is as follows: first, the original EEG signal is subjected to a one-dimensional convolution method to achieve a Short Time Fourier Transform (STFT) and passed through a Mel filter bank to obtain a logarithmic Mel spectrogram, and then the resulting logarithmic Mel spectrogram is fed into a fatigue detection model to complete the fatigue detection task for the EEG signals. The fatigue detection model consists of a 6-layer convolutional neural network (CNN), bi-directional recurrent neural networks (Bi-RNNs), and a classifier. In the modeling phase, spectrogram features are transported to the 6-layer CNN to automatically learn high-level features, thereby extracting temporal features in the bi-directional RNN to obtain spectrogram-temporal information. Finally, the alert or fatigue state is obtained by a classifier consisting of a fully connected layer, a ReLU activation function, and a softmax function. Experiments were conducted on publicly available datasets in this study. The results show that the method can accurately distinguish between alert and fatigue states with high stability. In addition, the performance of four existing methods was compared with the results of the proposed method, all of which showed that the proposed method could achieve the best results so far.

Two-Stream Spatial-Temporal Graph Convolutional Networks for Driver Drowsiness Detection

Convolutional neural networks (CNNs) have achieved remarkable performance in driver drowsiness detection based on the extraction of deep features of drivers' faces. However, the performance of driver drowsiness detection methods decreases sharply when complications, such as illumination changes in the cab, occlusions and shadows on the driver's face, and variations in the driver's head pose, occur. In addition, current driver drowsiness detection methods are not capable of distinguishing between driver states, such as talking versus yawning or blinking versus closing eyes. Therefore, technical challenges remain in driver drowsiness detection. In this article, we propose a novel and robust two-stream spatial-temporal graph convolutional network (2s-STGCN) for driver drowsiness detection to solve the above-mentioned challenges. To take advantage of the spatial and temporal features of the input data, we use a facial landmark detection method to extract the driver's facial landmarks from real-time videos and then obtain the driver drowsiness detection result by 2s-STGCN. Unlike existing methods, our proposed method uses videos rather than consecutive video frames as processing units. This is the first effort to exploit these processing units in the field of driver drowsiness detection. Moreover, the two-stream framework not only models both the spatial and temporal features but also models both the first-order and second-order information simultaneously, thereby notably improving driver drowsiness detection. Extensive experiments have been performed on the yawn detection dataset (YawDD) and the National TsingHua University drowsy driver detection (NTHU-DDD) dataset. The experimental results validate the feasibility of the proposed method. This method achieves an average accuracy of 93.4% on the YawDD dataset and an average accuracy of 92.7% on the evaluation set of the NTHU-DDD dataset.

The Application of Electroencephalogram in Driving Safety: Current Status and Future Prospects

The driver is one of the most important factors in the safety of the transportation system. The driver's perceptual characteristics are closely related to driving behavior, while electroencephalogram (EEG) as the gold standard for evaluating human perception is non-deceptive. It is essential to study driving characteristics by analyzing the driver's brain activity pattern, effectively acquiring driver perceptual characteristics, creating a direct connection between the driver's brain and external devices, and realizing information interchange. This paper first introduces the theories related to EEG, then reviews the applications of EEG in scenarios such as fatigue driving, distracted driving, and emotional driving. The limitations of existing research have been identified and the prospect of EEG application in future brain-computer interface automotive assisted driving systems have been proposed. This review provides guidance for researchers to use EEG to improve driving safety. It also offers valuable suggestions for future research.

EEG-Based Driving Fatigue Detection Using a Two-Level Learning Hierarchy Radial Basis Function

Electroencephalography (EEG)-based driving fatigue detection has gained increasing attention recently due to the non-invasive, low-cost, and potable nature of the EEG technology, but it is still challenging to extract informative features from noisy EEG signals for driving fatigue detection. Radial basis function (RBF) neural network has drawn lots of attention as a promising classifier due to its linear-in-the-parameters network structure, strong non-linear approximation ability, and desired generalization property. The RBF network performance heavily relies on network parameters such as the number of the hidden nodes, number of the center vectors, width, and output weights. However, global optimization methods that directly optimize all the network parameters often result in high evaluation cost and slow convergence. To enhance the accuracy and efficiency of EEG-based driving fatigue detection model, this study aims to develop a two-level learning hierarchy RBF network (RBF-TLLH) which allows for global optimization of the key network parameters. Experimental EEG data were collected, at both fatigue and alert states, from six healthy participants in a simulated driving environment. Principal component analysis was first utilized to extract features from EEG signals, and the proposed RBF-TLLH was then employed for driving status (fatigue vs. alert) classification. The results demonstrated that the proposed RBF-TLLH approach achieved a better classification performance (mean accuracy: 92.71%; area under the receiver operating curve: 0.9199) compared to other widely used artificial neural networks. Moreover, only three core parameters need to be determined using the training datasets in the proposed RBF-TLLH classifier, which increases its reliability and applicability. The findings demonstrate that the proposed RBF-TLLH approach can be used as a promising framework for reliable EEG-based driving fatigue detection.

Applications of EEG indices for the quantification of human cognitive performance: A systematic review and bibliometric analysis

BACKGROUND

Neuroergonomics combines neuroscience with ergonomics to study human performance using recorded brain signals. Such neural signatures of performance can be measured using a variety of neuroimaging techniques, including functional magnetic resonance imaging (fMRI), functional near-infrared spectroscopy (fNIRS), and electroencephalography (EEG). EEG has an excellent temporal resolution, and EEG indices are highly sensitive to human brain activity fluctuations.

OBJECTIVE

The focus of this systematic review was to explore the applications of EEG indices for quantifying human performance in a variety of cognitive tasks at the macro and micro scales. To identify trends and the state of the field, we examined global patterns among selected articles, such as journal contributions, highly cited papers, affiliations, and high-frequency keywords. Moreover, we discussed the most frequently used EEG indices and synthesized current knowledge regarding the EEG signatures of associated human performance measurements.

METHODS

In this systematic review, we analyzed articles published in English (from peer-reviewed journals, proceedings, and conference papers), Ph.D. dissertations, textbooks, and reference books. All articles reviewed herein included exclusively EEG-based experimental studies in healthy participants. We searched Web-of-Science and Scopus databases using specific sets of keywords.

RESULTS

Out of 143 papers, a considerable number of cognitive studies focused on quantifying human performance with respect to mental fatigue, mental workload, mental effort, visual fatigue, emotion, and stress. An increasing trend for publication in this area was observed, with the highest number of publications in 2017. Most studies applied linear methods (e.g., EEG power spectral density and the amplitude of event-related potentials) to evaluate human cognitive performance. A few papers utilized nonlinear methods, such as fractal dimension, largest Lyapunov exponent, and signal entropy. More than 50% of the studies focused on evaluating an individual's mental states while operating a vehicle. Several different methods of artifact removal have also been noted. Based on the reviewed articles, research gaps, trends, and potential directions for future research were explored.

CONCLUSION

This systematic review synthesized current knowledge regarding the application of EEG indices for quantifying human performance in a wide variety of cognitive tasks. This knowledge is useful for understanding the global patterns of applications of EEG indices for the analysis and design of cognitive tasks.

A Systemic Review of Available Low-Cost EEG Headsets Used for Drowsiness Detection

Drowsiness is a leading cause of traffic and industrial accidents, costing lives and productivity. Electroencephalography (EEG) signals can reflect awareness and attentiveness, and low-cost consumer EEG headsets are available on the market. The use of these devices as drowsiness detectors could increase the accessibility of safety and productivity-enhancing devices for small businesses and developing countries. We conducted a systemic review of currently available, low-cost, consumer EEG-based drowsiness detection systems. We sought to determine whether consumer EEG headsets could be reliably utilized as rudimentary drowsiness detection systems. We included documented cases describing successful drowsiness detection using consumer EEG-based devices, including the Neurosky MindWave, InteraXon Muse, Emotiv Epoc, Emotiv Insight, and OpenBCI. Of 46 relevant studies, ~27 reported an accuracy score. The lowest of these was the Neurosky Mindwave, with a minimum of 31%. The second lowest accuracy reported was 79.4% with an OpenBCI study. In many cases, algorithmic optimization remains necessary. Different methods for accuracy calculation, system calibration, and different definitions of drowsiness made direct comparisons problematic. However, even basic features, such as the power spectra of EEG bands, were able to consistently detect drowsiness. Each specific device has its own capabilities, tradeoffs, and limitations. Widely used spectral features can achieve successful drowsiness detection, even with low-cost consumer devices; however, reliability issues must still be addressed in an occupational context.

Recognition of cognitive load with a stacking network ensemble of denoising autoencoders and abstracted neurophysiological features

The safety of human-machine systems can be indirectly evaluated based on operator's cognitive load levels at each temporal instant. However, relevant features of cognitive states are hidden behind in multiple sources of cortical neural responses. In this study, we developed a novel neural network ensemble, SE-SDAE, based on stacked denoising autoencoders (SDAEs) which identify different levels of cognitive load by electroencephalography (EEG) signals. To improve the generalization capability of the ensemble framework, a stacking-based approach is adopted to fuse the abstracted EEG features from activations of deep-structured hidden layers. In particular, we also combine multiple K-nearest neighbor and naive Bayesian classifiers with SDAEs to generate a heterogeneous classification committee to enhance ensemble's diversity. Finally, we validate the proposed SE-SDAE by comparing its performance with mainstream pattern classifiers for cognitive load evaluation to show its effectiveness.

Generalizability of A Neural Network Model for Circadian Phase Prediction in Real-World Conditions

A neural network model was previously developed to predict melatonin rhythms accurately from blue light and skin temperature recordings in individuals on a fixed sleep schedule. This study aimed to test the generalizability of the model to other sleep schedules, including rotating shift work. Ambulatory wrist blue light irradiance and skin temperature data were collected in 16 healthy individuals on fixed and habitual sleep schedules, and 28 rotating shift workers. Artificial neural network models were trained to predict the circadian rhythm of (i) salivary melatonin on a fixed sleep schedule; (ii) urinary aMT6s on both fixed and habitual sleep schedules, including shift workers on a diurnal schedule; and (iii) urinary aMT6s in rotating shift workers on a night shift schedule. To determine predicted circadian phase, center of gravity of the fitted bimodal skewed baseline cosine curve was used for melatonin, and acrophase of the cosine curve for aMT6s. On a fixed sleep schedule, the model predicted melatonin phase to within ± 1 hour in 67% and ± 1.5 hours in 100% of participants, with mean absolute error of 41 ± 32 minutes. On diurnal schedules, including shift workers, the model predicted aMT6s acrophase to within ± 1 hour in 66% and ± 2 hours in 87% of participants, with mean absolute error of 63 ± 67 minutes. On night shift schedules, the model predicted aMT6s acrophase to within ± 1 hour in 42% and ± 2 hours in 53% of participants, with mean absolute error of 143 ± 155 minutes. Prediction accuracy was similar when using either 1 (wrist) or 11 skin temperature sensor inputs. These findings demonstrate that the model can predict circadian timing to within ± 2 hours for the vast majority of individuals on diurnal schedules, using blue light and a single temperature sensor. However, this approach did not generalize to night shift conditions.

Portable Drowsiness Detection through Use of a Prefrontal Single-Channel Electroencephalogram

Drowsiness detection has been studied in the context of evaluating products, assessing driver alertness, and managing office environments. Drowsiness level can be readily detected through measurement of human brain activity. The electroencephalogram (EEG), a device whose application relies on adhering electrodes to the scalp, is the primary method used to monitor brain activity. The many electrodes and wires required to perform an EEG place considerable constraints on the movement of users, and the cost of the device limits its availability. For these reasons, conventional EEG devices are not used in practical studies and businesses. Many potential practical applications could benefit from the development of a wire-free, low-priced device; however, it remains to be elucidated whether portable EEG devices can be used to estimate human drowsiness levels and applied within practical research settings and businesses. In this study, we outline the development of a drowsiness detection system that makes use of a low-priced, prefrontal single-channel EEG device and evaluate its performance in an offline analysis and a practical experiment. Firstly, for the development of the system, we compared three feature extraction methods: power spectral density (PSD), autoregressive (AR) modeling, and multiscale entropy (MSE) for detecting characteristics of an EEG. In order to efficiently select a meaningful PSD, we utilized step-wise linear discriminant analysis (SWLDA). Time-averaging and robust-scaling were used to fit the data for pattern recognition. Pattern recognition was performed by a support vector machine (SVM) with a radial basis function (RBF) kernel. The optimal hyperparameters for the SVM were selected by the grind search method so as to increase drowsiness detection accuracy. To evaluate the performance of the detections, we calculated classification accuracy using the SVM through 10-fold cross-validation. Our model achieved a classification accuracy of 72.7% using the PSD with SWLDA and the SVM. Secondly, we conducted a practical study using the system and evaluated its performance in a practical situation. There was a significant difference (* p < 0.05) between the drowsiness-evoked task and concentration-needed task. Our results demonstrate the efficacy of our low-priced portable drowsiness detection system in quantifying drowsy states. We anticipate that our system will be useful to practical studies with aims as diverse as measurement of classroom mental engagement, evaluation of movies, and office environment evaluation.

A contextual and temporal algorithm for driver drowsiness detection

This study designs and evaluates a contextual and temporal algorithm for detecting drowsiness-related lane. The algorithm uses steering angle, pedal input, vehicle speed and acceleration as input. Speed and acceleration are used to develop a real-time measure of driving context. These measures are integrated with a Dynamic Bayesian Network that considers the time dependencies in transitions between drowsiness and awake states. The Dynamic Bayesian Network algorithm is validated with data collected from 72 participants driving the National Advanced Driving Simulator. The algorithm has a significantly lower false positive rate than PERCLOS-the current gold standard-and baseline, non-contextual, algorithms under design parameters that prioritize drowsiness detection. Under these parameters, the algorithm reduces false positive rate in highway and rural environments, which are typically problematic for vehicle-based detection algorithms. This algorithm is a promising new approach to driver impairment detection and suggests contextual factors should be considered in subsequent algorithm development processes. It may be combined with comprehensive mitigation methods to improve driving safety.

A subject-transfer framework for obviating inter- and intra-subject variability in EEG-based drowsiness detection

Inter- and intra-subject variability pose a major challenge to decoding human brain activity in brain-computer interfaces (BCIs) based on non-invasive electroencephalogram (EEG). Conventionally, a time-consuming and laborious training procedure is performed on each new user to collect sufficient individualized data, hindering the applications of BCIs on monitoring brain states (e.g. drowsiness) in real-world settings. This study proposes applying hierarchical clustering to assess the inter- and intra-subject variability within a large-scale dataset of EEG collected in a simulated driving task, and validates the feasibility of transferring EEG-based drowsiness-detection models across subjects. A subject-transfer framework is thus developed for detecting drowsiness based on a large-scale model pool from other subjects and a small amount of alert baseline calibration data from a new user. The model pool ensures the availability of positive model transferring, whereas the alert baseline data serve as a selector of decoding models in the pool. Compared with the conventional within-subject approach, the proposed framework remarkably reduced the required calibration time for a new user by 90% (18.00 min-1.72 ± 0.36 min) without compromising performance (p = 0.0910) when sufficient existing data are available. These findings suggest a practical pathway toward plug-and-play drowsiness detection and can ignite numerous real-world BCI applications.

New complexity measures reveal that topographic loops of human alpha phase potentials are more complex in drowsy than in wake

A number of measures, stemming from nonlinear dynamics, exist to estimate complexity of biomedical objects. In most cases they are appropriate, but sometimes unconventional measures, more suited for specific objects, are needed to perform the task. In our present work, we propose three new complexity measures to quantify complexity of topographic closed loops of alpha carrier frequency phase potentials (CFPP) of healthy humans in wake and drowsy states. EEG of ten adult individuals was recorded in both states, using a 14-channel montage. For each subject and each state, a topographic loop (circular directed graph) was constructed according to CFPP values. Circular complexity measure was obtained by summing angles which directed graph edges (arrows) form with the topographic center. Longitudinal complexity was defined as the sum of all arrow lengths, while intersecting complexity was introduced by counting the number of intersections of graph edges. Wilcoxon's signed-ranks test was used on the sets of these three measures, as well as on fractal dimension values of some loop properties, to test differences between loops obtained in wake vs. drowsy. While fractal dimension values were not significantly different, longitudinal and intersecting complexities, as well as anticlockwise circularity, were significantly increased in drowsy. Graphical abstract An example of closed topographic carrier frequency phase potential (CFPP) loops, recorded in one of the subjects in the wake (A) and drowsy (C) states. Lengths of loop graph edges, r(c _j, c _j + 1), plotted against the series of EEG channels with decreasing CFPP values, c _j , in the wake (B) and drowsy (D) states. Conventional fractal analysis did not reveal any difference between them; therefore, three new complexity measures were introduced.

On-line EEG Denoising and Cleaning Using Correlated Sparse Recovery and Active Learning

We have developed two new methods that use sparse recovery and active learning techniques for near real-time artifact identification and removal in EEG recordings. The first algorithm, called Correlated Sparse Signal Recovery (CSSR) addresses the problem of structured sparse signal recovery when statistical rather than exact properties describing the structure of the signal are appropriate, as in the elimination of eye movement artifacts; such tasks cannot be done efficiently using structured models that assume a common sparsity profile of fixed groups of components. Our algorithm learns structured sparse coefficients in a Bayesian paradigm. Using it, we have successfully identified and subtracted eye movement (structured) artifacts in real EEG recordings resulting in minimal data loss. Our method outperforms ICA and standard sparse recovery algorithms by preserving both spectral and complexity properties of the denoised EEG. Our second method uses a new active selection algorithm that we call Output-based Active Selection (OAS). When applied to the task of detection of EEG epochs containing other non-structured artifacts from an ensemble of detectors, OAS boosts accuracy of the ensemble from 91% to 97.5% with only 10% active labels. Our methods can also be applied to real-time artifact removal in magnetoencephalography (MEG) and blood pressure signals.

Utilization of a combined EEG/NIRS system to predict driver drowsiness

The large number of automobile accidents due to driver drowsiness is a critical concern of many countries. To solve this problem, numerous methods of countermeasure have been proposed. However, the results were unsatisfactory due to inadequate accuracy of drowsiness detection. In this study, we introduce a new approach, a combination of EEG and NIRS, to detect driver drowsiness. EEG, EOG, ECG and NIRS signals have been measured during a simulated driving task, in which subjects underwent both awake and drowsy states. The blinking rate, eye closure, heart rate, alpha and beta band power were used to identify subject's condition. Statistical tests were performed on EEG and NIRS signals to find the most informative parameters. Fisher's linear discriminant analysis method was employed to classify awake and drowsy states. Time series analysis was used to predict drowsiness. The oxy-hemoglobin concentration change and the beta band power in the frontal lobe were found to differ the most between the two states. In addition, these two parameters correspond well to an awake to drowsy state transition. A sharp increase of the oxy-hemoglobin concentration change, together with a dramatic decrease of the beta band power, happened several seconds before the first eye closure.

APPLICATION OF S-TRANSFORM FOR AUTOMATED DETECTION OF VIGILANCE LEVEL USING EEG SIGNALS

This paper presents an S-transform-based Electroencephalogram channel optimization and feature extraction methodology for monitoring mental vigilance level of humans. Vigilance level detection methodology consists of four steps. In the first stage, two types of Electroencephalogram signals (alert and drowsy) are acquired from 30 healthy subjects and decomposed into sub-bands using the S-transform. In the second stage, permutation entropy of the S-transform coefficients is calculated and Electroencephalogram channel optimization is performed. S-transform-based statistical features are computed from the optimized Electroencephalogram channels, in the third stage. In the fourth stage, artificial intelligence techniques such as Least Square-Support Vector Machine, Artificial Neural Network and Naive Bayes Classifier are used for the classification of Electroencephalogram signals using extracted features. The performance of the feature extraction methodology is tested on the Electroencephalogram data of 30 healthy subjects. Experimental results ensured the effectiveness of proposed methodology for the estimation of mental vigilance level by using Electroencephalogram signals. It is observed that the Artificial Neural Network classifier is a good candidate for pre-emptive automatic vigilance level detection system for Brain-Computer Interface applications.

Mutual information measures applied to EEG signals for sleepiness characterization

Excessive daytime sleepiness (EDS) is one of the main symptoms of several sleep related disorders with a great impact on the patient lives. While many studies have been carried out in order to assess daytime sleepiness, the automatic EDS detection still remains an open problem. In this work, a novel approach to this issue based on non-linear dynamical analysis of EEG signal was proposed. Multichannel EEG signals were recorded during five maintenance of wakefulness (MWT) and multiple sleep latency (MSLT) tests alternated throughout the day from patients suffering from sleep disordered breathing. A group of 20 patients with excessive daytime sleepiness (EDS) was compared with a group of 20 patients without daytime sleepiness (WDS), by analyzing 60-s EEG windows in waking state. Measures obtained from cross-mutual information function (CMIF) and auto-mutual-information function (AMIF) were calculated in the EEG. These functions permitted a quantification of the complexity properties of the EEG signal and the non-linear couplings between different zones of the scalp. Statistical differences between EDS and WDS groups were found in β band during MSLT events (p-value < 0.0001). WDS group presented more complexity than EDS in the occipital zone, while a stronger nonlinear coupling between occipital and frontal zones was detected in EDS patients than in WDS. The AMIF and CMIF measures yielded sensitivity and specificity above 80% and AUC of ROC above 0.85 in classifying EDS and WDS patients.

Topographic distribution of EEG alpha attractor correlation dimension values in wake and drowsy states in humans

Organization of resting state cortical networks is of fundamental importance for the phenomenon of awareness, which is altered in the first part of hypnagogic period (Hori stages 1-4). Our aim was to investigate the change in brain topography pattern of EEG alpha attractor correlation dimension (CD) in the period of transition from Hori stage 1 to 4. EEG of ten healthy adult individuals was recorded in the wake and drowsy states, using a 14 channel average reference montage, from which 91 bipolar channels were derived and filtered in the wider alpha (6-14 Hz) range. Sixty 1s long epochs of each state and individual were subjected to CD calculation according to the Grassberger-Procaccia method. For such a collection of signals, two embedding dimensions, d={5, 10}, and 22 time delays τ=2-23 samples were explored. Optimal values were d=10 and τ=18, where both saturation and second zero crossing of the autocorrelation function occurred. Bipolar channel CD underwent a significant decrease during the transition and showed a positive linear correlation with electrode distance, stronger in the wake individuals. Topographic distribution of bipolar channels with above median CD changed from longitudinal anterior-posterior pattern (awake) to a more diagonal pattern, with localization in posterior regions (drowsiness). Our data are in line with the literature reporting functional segregation of neuronal assemblies in anterior and posterior regions during this transition. Our results should contribute to understanding of complex reorganization of the cortical part of alpha generators during the wake/drowsy transition.

Correntropy measures to detect daytime sleepiness from EEG signals

Excessive daytime sleepiness (EDS) is one of the main symptoms of several sleep related disorders and has a great impact on patients' lives. While many studies have been carried out in order to assess daytime sleepiness, automatic EDS detection still remains an open problem. In this work, a novel approach to this issue based on correntropy function analysis of EEG signals was proposed in order to detect patients suffering from EDS. Multichannel EEG signals were recorded during five Maintenance of Wakefulness Tests (MWT) and Multiple Sleep Latency Tests (MSLT) alternated throughout the day for patients suffering from sleep disordered breathing (SDB). A group of 20 patients with EDS was compared with a group of 20 patients without daytime sleepiness (WDS), by analyzing 60 s EEG windows in a waking state. Measures obtained from the cross-correntropy function (CCORR) and auto-correntropy function (ACORR) were calculated in the EEG frequency bands: δ, 0.1-4 Hz; θ, 4-8 Hz; α, 8-12 Hz; β, 12-30 Hz; total band TB, 0.1-45 Hz. These functions permitted the quantification of complex signal properties and the non-linear couplings between different areas of the scalp. Statistical differences between EDS and WDS groups were mainly found in the β band during MSLT events (p-value < 0.0001). The WDS group presented more complexity in the occipital zone than the EDS group, while a stronger nonlinear coupling between the occipital and frontal regions was detected in EDS patients than in the WDS group. At best, ACORR and CCORR measures yielded sensitivity and specificity above 80% and the area under ROC curve (AUC) was above 0.85 in classifying EDS and WDS patients. These performances represent an improvement with respect to classical EEG indices applied in the same database (sensitivity and specificity were never above 80% and AUC was under 0.75).

In-flight automatic detection of vigilance states using a single EEG channel

Sleepiness and fatigue can reach particularly high levels during long-haul overnight flights. Under these conditions, voluntary or even involuntary sleep periods may occur, increasing the risk of accidents. The aim of this study was to assess the performance of an in-flight automatic detection system of low-vigilance states using a single electroencephalogram channel. Fourteen healthy pilots voluntarily wore a miniaturized brain electrical activity recording device during long-haul flights ( 10 ±2.0 h, Atlantic 2 and Falcon 50 M, French naval aviation). No subject was disturbed by the equipment. Seven pilots experienced at least a period of voluntary ( 26.8 ±8.0 min, n = 4) or involuntary sleep (N1 sleep stage, 26.6 ±18.7 s, n = 7) during the flight. Automatic classification (wake/sleep) by the algorithm was made for 10-s epochs (O1-M2 or C3-M2 channel), based on comparison of means to detect changes in α, β, and θ relative power, or ratio [( α+θ)/β], or fuzzy logic fusion (α, β). Pertinence and prognostic of the algorithm were determined using epoch-by-epoch comparison with visual-scoring (two blinded readers, AASM rules). The best concordance between automatic detection and visual-scoring was observed within the O1-M2 channel, using the ratio [( α+θ )/β] ( 98.3 ±4.1% of good detection, K = 0.94 ±0.07, with a 0.04 ±0.04 false positive rate and a 0.87 ±0.10 true positive rate). Our results confirm the efficiency of a miniaturized single electroencephalographic channel recording device, associated with an automatic detection algorithm, in order to detect low-vigilance states during real flights.

Measuring neurophysiological signals in aircraft pilots and car drivers for the assessment of mental workload, fatigue and drowsiness

This paper reviews published papers related to neurophysiological measurements (electroencephalography: EEG, electrooculography EOG; heart rate: HR) in pilots/drivers during their driving tasks. The aim is to summarise the main neurophysiological findings related to the measurements of pilot/driver's brain activity during drive performance and how particular aspects of this brain activity could be connected with the important concepts of "mental workload", "mental fatigue" or "situational awareness". Review of the literature suggests that exists a coherent sequence of changes for EEG, EOG and HR variables during the transition from normal drive, high mental workload and eventually mental fatigue and drowsiness. In particular, increased EEG power in theta band and a decrease in alpha band occurred in high mental workload. Successively, increased EEG power in theta as well as delta and alpha bands characterise the transition between mental workload and mental fatigue. Drowsiness is also characterised by increased blink rate and decreased HR values. The detection of such mental states is actually performed "offline" with accuracy around 90% but not online. A discussion on the possible future applications of findings provided by these neurophysiological measurements in order to improve the safety of the vehicles will be also presented.

CLASSIFICATION OF EEG SIGNALS IN NORMAL AND DEPRESSION CONDITIONS BY ANN USING RWE AND SIGNAL ENTROPY

EEG is useful for the analysis of the functional activity of the brain and a detailed assessment of this non-stationary waveform can provide crucial parameters indicative of the mental state of patients. The complex nature of EEG signals calls for automated analysis using various signal processing methods. This paper attempts to classify the EEG signals of normal and depression patients using well-established signal processing techniques involving relative wavelet energy (RWE) and artificial feedForward neural network. High frequency noise present in the recorded signal is removed using total variation filtering (TVF). Classification of the frequency bands of EEG signals into appropriate detail levels and approximation level is carried out using an eight-level multiresolution decomposition method of discrete wavelet transform (DWT). Parseval's theorem is used for calculating the energy at different resolution levels. RWE analysis gives information about the signal energy distribution at different decomposition levels. Both RWE and feedforward Network are used to classify the signals from normal controls and depression patients. The performance of the artificial neural network was evaluated using the classification accuracy and its value of 98.11% indicates a great potential for classifying normal and depression signals.

Modeling the relationship between Higuchi's fractal dimension and Fourier spectra of physiological signals

The exact mathematical relationship between FFT spectrum and fractal dimension (FD) of an experimentally recorded signal is not known. In this work, we tried to calculate signal FD directly from its Fourier amplitudes. First, dependence of Higuchi's FD of mathematical sinusoids on their individual frequencies was modeled with a two-parameter exponential function. Next, FD of a finite sum of sinusoids was found to be a weighted average of their FDs, weighting factors being their Fourier amplitudes raised to a fractal degree. Exponent dependence on frequency was modeled with exponential, power and logarithmic functions. A set of 280 EEG signals and Weierstrass functions were analyzed. Cross-validation was done within EEG signals and between them and Weierstrass functions. Exponential dependence of fractal exponents on frequency was found to be the most accurate. In this work, signal FD was for the first time expressed as a fractal weighted average of FD values of its Fourier components, also allowing researchers to perform direct estimation of signal fractal dimension from its FFT spectrum.

EEG alpha phase shifts during transition from wakefulness to drowsiness

Phases of alpha oscillations recorded by EEG were typically studied in the context of event or task related experiments, rarely during spontaneous alpha activity and in different brain states. During wake-to-drowsy transition they change unevenly, depending on the brain region. To explore their dynamics, we recorded ten adult healthy individuals in these two states. Alpha waves were treated as stable frequency and variable amplitude signals with one carrier frequency (CF). A method for calculating their CF phase shifts (CFPS) and CF phase potentials (CFPP) was developed and verified on surrogate signals as more accurate than phase shifts of Fourier components. Probability density estimate (PDE) of CFPS, CFPP and CF phase locking showed that frontal and fronto-temporal areas of the cortex underwent more extensive changes than posterior regions. The greatest differences were found between pairs of channels involving F7, F8, F3 and F4 (PDE of CFPS); F7, F8, T3 and T4 (CFPP); F7, F8, F3, F4, C3, C4 and T3 (decrease in CF phase locking). A topographic distribution of channels with above the average phase locking in the wake state revealed two separate regions occupying anterior and posterior brain areas (with intra regional and inter hemispheric connections). These regions merged and became mutually phase locked longitudinally in the drowsy state. Changes occurring primarily in the frontal and fronto-temporal regions correlated with an early decrease of alertness. Areas of increased phase locking might be correlated with topography of synchronous neuronal assemblies conceptualized within neural correlates of consciousness.

Hybrid Intelligent Diagnosis Approach Based On Neural Pattern Recognition and Fuzzy Decision-Making

Fault diagnosis is a complex and fuzzy cognitive process, and soft computing methods and technologies based on Neural Networks (NN) and Fuzzy Logic (FL), have shown great potential in the development of Decision Support Systems (DSS). Dealing with expert (human) knowledge consideration, Computer Aided Diagnosis (CAD) dilemma is one of the most interesting, but also one of the most difficult problems. Among difficulties contributing to challenging nature of this problem, one can mention the need of fine pattern recognition (classification) and decision-making. This Chapter deals with classification and decision-making based on Artificial Intelligence using multiple model approaches under soft computing implying modular Neural Networks (NN) and Fuzzy Logic (FL) for biomedical and industrial applications. The aim of this Chapter is absolutely not to replace specialized human but to suggest decision support tools: hybrid intelligent diagnosis systems with a satisfactory reliability degree for CAD. In this Chapter, a methodology is given in order to design hybrid intelligent diagnosis systems for a large field of biomedical and industrial applications. For this purpose, first, a survey on diagnosis tasks in such applications is presented. Second, fault diagnosis systems are presented. Third, the main steps of hybrid intelligent diagnosis systems are developed, for each step emphasizing problems and suggesting solutions able to ensure the design of hybrid intelligent diagnosis systems with a satisfactory reliability degree. In fact, the main steps discussed are knowledge representation, classification, classifier issued information fusion, and decision-making. Then, the suggested approach is developed for a CAD in biomedicine, from Auditory Brainstem Response (ABR) test, and the prototype design and experimental results are presented. Finally, a discussion is given with regard to the reliability and large application field of the suggested approach.

EEG-derived estimators of present and future cognitive performance

Previous electroencephalography (EEG)-based fatigue-related research primarily focused on the association between concurrent cognitive performance and time-locked physiology. The goal of this study was to investigate the capability of EEG to assess the impact of fatigue on both present and future cognitive performance during a 20-min sustained attention task, the 3-choice active vigilance task (3CVT), that requires subjects to discriminate one primary target from two secondary non-target geometric shapes. The current study demonstrated the ability of EEG to estimate not only present, but also future cognitive performance, utilizing a single, combined reaction time (RT), and accuracy performance metric. The correlations between observed and estimated performance, for both present and future performance, were strong (up to 0.89 and 0.79, respectively). The models were able to consistently estimate "unacceptable" performance throughout the entire 3CVT, i.e., excessively missed responses and/or slow RTs, while acceptable performance was recognized less accurately later in the task. The developed models were trained on a relatively large dataset (n = 50 subjects) to increase stability. Cross-validation results suggested the models were not over-fitted. This study indicates that EEG can be used to predict gross-performance degradations 5-15 min in advance.

Drowsiness/alertness algorithm development and validation using synchronized EEG and cognitive performance to individualize a generalized model

A great deal of research over the last century has focused on drowsiness/alertness detection, as fatigue-related physical and cognitive impairments pose a serious risk to public health and safety. Available drowsiness/alertness detection solutions are unsatisfactory for a number of reasons: (1) lack of generalizability, (2) failure to address individual variability in generalized models, and/or (3) lack of a portable, un-tethered application. The current study aimed to address these issues, and determine if an individualized electroencephalography (EEG) based algorithm could be defined to track performance decrements associated with sleep loss, as this is the first step in developing a field deployable drowsiness/alertness detection system. The results indicated that an EEG-based algorithm, individualized using a series of brief "identification" tasks, was able to effectively track performance decrements associated with sleep deprivation. Future development will address the need for the algorithm to predict performance decrements due to sleep loss, and provide field applicability.

Off-line and on-line vigilance estimation based on linear dynamical system and manifold learning

For many human machine interaction systems, to ensure work safety, the techniques for continuously estimating the vigilance of operators are highly desirable. Up to now, various methods based on electroencephalogram (EEG) are proposed to solve this problem. However, most of them are static methods and are based on supervised learning strategy. The main deficiencies of the existing methods are that the label information is hard to get and the time dependency of vigilance changes are ignored. In this paper, we introduce the dynamic characteristics of vigilance changes into vigilance estimation and propose a novel model based on linear dynamical system and manifold learning techniques to implement off-line and online vigilance estimation. In this model, both spatial information of EEG and temporal information of vigilance changes are used. The label information what we need is merely to know which EEG indices are important for vigilance estimation. Experimental results show that the mean off-line and on-line correlation coefficients between estimated vigilance level and local error rate in second-scale without being averaged are 0.89 and 0.83, respectively.

A portable device for real time drowsiness detection using novel active dry electrode system

Electroencephalogram (EEG) signals give important information about the vigilance states of a subject. Therefore, this study constructs a real-time EEG-based system for detecting a drowsy driver. The proposed system uses a novel six channels active dry electrode system to acquire EEG non-invasively. In addition, it uses a TMS320VC5510 DSP chip as the algorithm processor, and a MSP430F149 chip as a controller to achieve a real-time portable system. This study implements stationary wavelet transform to extract two features of EEG signal: integral of EEG and zero crossings as the input to a back propagation neural network for vigilance states classification. This system can discriminate alertness and drowsiness in real-time. The accuracy of the system is 79.1% for alertness and 90.91% for drowsiness states. When the system detects drowsiness, it will warn drivers by using a vibrator and a beeper.