Temporal complexity of EEG encodes human alertness.

Objective. Automatic human alertness monitoring has recently become an important research topic with important applications in many areas such as the detection of drivers' fatigue, monitoring of monotonous tasks that require a high level of alertness such as traffic control and nuclear power plant monitoring, and sleep staging. In this study, we propose that balanced dynamics of Electroencephalography (EEG) (so called EEG temporal complexity) is a potentially useful feature for identifying human alertness states. Recently, a new signal entropy measure, called range entropy (RangeEn), was proposed to overcome some limitations of two of the most widely used entropy measures, namely approximate entropy (ApEn) and Sample Entropy (SampEn), and showed its relevance for the study of time domain EEG complexity. In this paper, we investigated whether the RangeEn holds discriminating information associated with human alertness states, namely awake, drowsy, and sleep and compare its performance against those of SampEn and ApEn.Approach. We used EEG data from 60 healthy subjects of both sexes and different ages acquired during whole night sleeps. Using a 30 s sliding window, we computed the three entropy measures of EEG and performed statistical analyses to evaluate the ability of these entropy measures to discriminate among the different human alertness states.Main results. Although the three entropy measures contained useful information about human alertness, RangeEn showed a higher discriminative capability compared to ApEn and SampEn especially when using EEG within the beta frequency band.Significance. Our findings highlight the EEG temporal complexity evolution through the human alertness states. This relationship can potentially be exploited for the development of automatic human alertness monitoring systems and diagnostic tools for different neurological and sleep disorders, including insomnia.

Neonatal EEG seizure detection using a new signal structural complexity measure based on matching pursuit decomposition with nonstationary dictionary.

BACKGROUND AND OBJECTIVE: In newborns, it is often difficult to accurately differentiate between seizure and non-seizure based solely on clinical manifestations. This highlights the importance of electroencephalogram (EEG) in the recognition and management of neonatal seizures. This paper proposes an effective algorithm for the detection of neonatal seizure using multichannel EEG. METHODS: Neonatal EEG changes morphology as it alternates between seizure and non-seizure states. A new signal complexity measure based on matching pursuit (MP) decomposition is proposed and used to detect transitions between these two states. The new measure, referred to as weighted structural complexity (WSC), was used for the detection of seizures in 30 newborn EEG records. Multiple IIR filters and an MP-based filter were designed and used to remove artifacts from the EEG data. Geometrical correlation between the EEG data channels was applied to reduce the number of false detections caused by remnant artifacts. The seizure detector's performance was assessed using several epoch-based (e.g., accuracy) and event-based (GDR = good detection rate and FD/h = false detections per hour) metrics. RESULTS: Compared to the neurologist marking, the proposed detector was able to detect EEG seizures with 94% accuracy, 90.9% GDR, and 0.14 FD/h (95% CI: [0.06, 0.34]). CONCLUSIONS: The high performance of the MP-based detector may have significant implications for the accurate diagnosis of neonatal seizures and the appropriate use of anticonvulsants and ongoing clinical assessment and care of the newborn.

Automatic fetal movement recognition from multi-channel accelerometry data.

BACKGROUND AND OBJECTIVE: Significant health care resources are allocated to monitoring high risk pregnancies to minimize growth compromise, reduce morbidity and prevent stillbirth. Fetal movement has been recognized as an important indicator of fetal health. Studies have shown that 25% of pregnancies with decreased fetal movement in the third trimester led to poor outcomes at birth. The studies have also shown that maternal perception of fetal movement is highly subjective and varies from person to person. A non-invasive system for fetal movement detection that can be used outside hospital would represent an advance in at-home monitoring of at-risk pregnancies. This is a challenging task that requires the use of advanced signal processing techniques to differentiate genuine fetal movements from contaminating artefacts. METHODS: This manuscript proposes a novel algorithm for automatic fetal movement recognition using data collected from wearable tri-axial accelerometers strategically placed on the maternal abdomen. The novelty of the work resides in the efficient removal of artefacts and in distinctive feature extraction. The proposed algorithm used independent component analysis (ICA) for dimensionality reduction and artefact removal. A supplemental technique based on discrete wavelet transform (DWT) was also used to remove artefacts. RESULTS: To identify fetal movements, 31 features were extracted from the acceleration data. Based on these features, several classifiers were used to distinguish fetal from non-fetal movements. Robustness of the classifiers was tested for various concentrations of artefacts in the classification data. The best performance was achieved by Bagging classifier algorithm, with random forest as its basis classifier, yielding an accuracy ranging from 87.6% to 95.8% depending on the artefact concentration level. CONCLUSIONS: A high performance detection of fetal movements can be achieved using accelerometery-based systems suitable for long-term monitoring.

Range Entropy: A Bridge between Signal Complexity and Self-Similarity.

Approximate entropy (ApEn) and sample entropy (SampEn) are widely used for temporal complexity analysis of real-world phenomena. However, their relationship with the Hurst exponent as a measure of self-similarity is not widely studied. Additionally, ApEn and SampEn are susceptible to signal amplitude changes. A common practice for addressing this issue is to correct their input signal amplitude by its standard deviation. In this study, we first show, using simulations, that ApEn and SampEn are related to the Hurst exponent in their tolerance r and embedding dimension m parameters. We then propose a modification to ApEn and SampEn called range entropy or RangeEn. We show that RangeEn is more robust to nonstationary signal changes, and it has a more linear relationship with the Hurst exponent, compared to ApEn and SampEn. RangeEn is bounded in the tolerance r-plane between 0 (maximum entropy) and 1 (minimum entropy) and it has no need for signal amplitude correction. Finally, we demonstrate the clinical usefulness of signal entropy measures for characterisation of epileptic EEG data as a real-world example.

Ventricular ectopic beats classification using Sparse Representation and Gini Index.

In this study, we consider using sparse representation and the Gini Index (GI) for Arrhythmia classification. Our approach involves, first, designing a separate dictionary for each Arrhythmia class using a set of labeled training QRS complexes. Sparse representations, based on the designed dictionaries, of each new test QRS complex are then calculated. Its class is finally predicted using the winner-takes-all principle; that is, the class associated with the highest GI is chosen. Our experiments showed promising results for the classification of premature ventricular contractions using a patient-specific approach. For many of the subjects considered, our classifier attained accuracies close to 100 % on the test set.

A Transform-Based Feature Extraction Approach for Motor Imagery Tasks Classification.

In this paper, we present a new motor imagery classification method in the context of electroencephalography (EEG)-based brain-computer interface (BCI). This method uses a signal-dependent orthogonal transform, referred to as linear prediction singular value decomposition (LP-SVD), for feature extraction. The transform defines the mapping as the left singular vectors of the LP coefficient filter impulse response matrix. Using a logistic tree-based model classifier; the extracted features are classified into one of four motor imagery movements. The proposed approach was first benchmarked against two related state-of-the-art feature extraction approaches, namely, discrete cosine transform (DCT) and adaptive autoregressive (AAR)-based methods. By achieving an accuracy of 67.35%, the LP-SVD approach outperformed the other approaches by large margins (25% compared with DCT and 6 % compared with AAR-based methods). To further improve the discriminatory capability of the extracted features and reduce the computational complexity, we enlarged the extracted feature subset by incorporating two extra features, namely, Q- and the Hotelling's [Formula: see text] statistics of the transformed EEG and introduced a new EEG channel selection method. The performance of the EEG classification based on the expanded feature set and channel selection method was compared with that of a number of the state-of-the-art classification methods previously reported with the BCI IIIa competition data set. Our method came second with an average accuracy of 81.38%.

Neonatal EEG at scalp is focal and implies high skull conductivity in realistic neonatal head models.

The potential improvements in spatial resolution of neonatal EEG used in source localization have been challenged by the insufficiencies in realistic neonatal head models. Our present study aimed at using empirical methods to indirectly estimate skull conductivity; the model parameter that is known to significantly affect the behavior of newborn scalp EEG and cause it to be markedly different from that of an adult. To this end, we used 64 channel EEG recordings to study the spatial specificity of scalp EEG by assessing the spatial decays in focal transients using both amplitudes and between-c'hannels linear correlations. The findings showed that these amplitudes and correlations decay within few centimeters from the reference channel/electrode, and that the nature of the decay is independent of the scalp area. This decay in newborn infants was found to be approximately three times faster than the corresponding decay in adult EEG analyzed from a set of 256 channel recordings. We then generated realistic head models using both finite and boundary element methods along with a manually segmented magnetic resonance images to study the spatial decays of scalp potentials produced by single dipole in the cortex. By comparing the spatial decays due to real and simulated EEG for different skull conductivities (from 0.003 to 0.3S/m), we showed that a close match between the empirical and simulated decays was obtained when the selected skull conductivity for newborn was around 0.06-0.2S/m. This is over an order of magnitude higher than the currently used values in adult head modeling. The results also showed that the neonatal scalp EEG is less smeared than that of an adult and this characteristic is the same across the entire scalp, including the fontanel region. These results indicate that a focal cortical activity is generally only registered by electrodes within few centimeters from the source. Hence, the conventional 10 to 20 channel neonatal EEG acquisition systems give a significantly spatially under sampled scalp EEG and may, consequently, give distorted pictures of focal brain activities. Such spatial specificity can only be reconciled by appreciating the anatomy of the neonatal head, especially the still unossified skull structure that needs to be modeled with higher conductivities than conventionally used in the adults.

Newborn seizure detection based on heart rate variability.

In this paper, we investigate the use of heart rate variability (HRV) for automatic newborn seizure detection. The proposed method consists of a sequence of processing steps, namely, obtaining HRV from the ECG, extracting a discriminating HRV feature set, selecting an optimal subset from the full feature set, and, finally, classifying the HRV into seizure/nonseizure using a supervised statistical classifier. Due to the fact that HRV signals are nonstationary, a set of time-frequency features from the newborn HRV is proposed and extracted. In order to achieve efficient HRV-based automatic newborn seizure detection, a two-phase wrapper-based feature selection technique is used to select the feature subset with minimum redundancy and maximum class discriminability. Tested on ECG recordings obtained from eight newborns with identified EEG seizure, the proposed HRV-based neonatal seizure detection algorithm achieved 85.7% sensitivity and 84.6% specificity. These results suggest that the HRV is sensitive to changes in the cardioregulatory system induced by the seizure, and therefore, can be used as a basis for an automatic seizure detection.

Robust time-frequency analysis of newborn EEG seizure corrupted by impulsive artefacts.

The newborn EEG seizure is a nonstationary signal. The time-varying nature of the newborn EEG seizure can be characterized by time-frequency representations (TFRs) such as quadratic time-frequency distributions. The underlying time-frequency signatures of newborn EEG seizure, however, can be severely masked by short-time and high amplitude (STHA), or impulsive, artefacts. This type of artefact can be modelled as heavy-tailed noise. Robust time-frequency distributions (RTFDs) have been proposed as methods for TFRs which are robust to heavy-tailed noise. In this paper, we investigate the use of RTFDs for representing the underlying time-frequency characteristics of newborn EEG seizure in the presence of STHA artefacts.

A nonstationary model of newborn EEG.

The detection of seizure in the newborn is a critical aspect of neurological research. Current automatic detection techniques are difficult to assess due to the problems associated with acquiring and labelling newborn electroencephalogram (EEG) data. A realistic model for newborn EEG would allow confident development, assessment and comparison of these detection techniques. This paper presents a model for newborn EEG that accounts for its self-similar and nonstationary nature. The model consists of background and seizure submodels. The newborn EEG background model is based on the short-time power spectrum with a time-varying power law. The relationship between the fractal dimension and the power law of a power spectrum is utilized for accurate estimation of the short-time power law exponent. The newborn EEG seizure model is based on a well-known time-frequency signal model. This model addresses all significant time-frequency characteristics of newborn EEG seizure which include; multiple components or harmonics, piecewise linear instantaneous frequency laws and harmonic amplitude modulation. Estimates of the parameters of both models are shown to be random and are modelled using the data from a total of 500 background epochs and 204 seizure epochs. The newborn EEG background and seizure models are validated against real newborn EEG data using the correlation coefficient. The results show that the output of the proposed models have a higher correlation with real newborn EEG than currently accepted models (a 10% and 38% improvement for background and seizure models, respectively).

Time-frequency analysis of heart rate variability for neonatal seizure detection.

The ECG has been much neglected in automatic seizure detection in the newborn. Changes in heart rate and ECG rhythm are often found in animal and adult patients with seizure. However, little is known about heart rate variability (HRV) changes in human neonate during seizure. Results of ongoing time-frequency research are presented here with the aim to compare the performance of various time-frequency distributions (TFDs) when applied to HRV time series for non-seizure and seizure newborns. The TFDs studied are the Wigner-Ville (WVD), the Spectrogram (SP), the Choi-Williams (CWD) and the Modified B (MBD) distributions. Based on our preliminary results, our current conclusion is MBD outperforms other TFDs in terms of time-frequency resolution, cross-terms suppression and to represent the newborn HRV signals of non-seizure and seizure which are closely-spaced components in the time-frequency domain.

Time-frequency based newborn EEG seizure detection using low and high frequency signatures.

The nonstationary and multicomponent nature of newborn EEG seizures tend to increase the complexity of the seizure detection problem. In dealing with this type of problem, time-frequency based techniques were shown to outperform classical techniques. Neonatal EEG seizures have signatures in both low frequency (lower than 10 Hz) and high frequency (higher than 70 Hz) areas. Seizure detection techniques have been proposed that concentrate on either low frequency or high frequency signatures of seizures. They, however, tend to miss seizures that reveal themselves only in one of the frequency areas. To overcome this problem, we propose a detection method that uses time-frequency seizure features extracted from both low and high frequency areas. Results of applying the proposed method on five newborn EEGs are very encouraging.