Luke Rankine

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)

A matching pursuit-based signal complexity measure for the analysis of newborn EEG

This paper presents a new relative measure of signal complexity, referred to here as relative structural complexity (RSC), which is based on the matching pursuit (MP) decomposition. By relative, we refer to the fact that this new measure is highly dependent on the decomposition dictionary used by MP. The structural part of the definition points to the fact that this new measure is related to the structure, or composition, of the signal under analysis. After a formal definition, the proposed RSC measure is used in the analysis of newborn electroencephalogram (EEG). To do this, firstly, a time–frequency decomposition dictionary is specifically designed to compactly represent the newborn EEG seizure state using MP. We then show, through the analysis of synthetic and real newborn EEG data, that the relative structural complexity measure can indicate changes in EEG structure as it transitions between the two EEG states; namely seizure and background (non-seizure).

Heart rate variability characterization using a time-frequency based instantaneous frequency estimation technique

In this paper, a new method for characterizing the newborn heart rate variability (HRV) is proposed. The central of the method is the newly proposed technique for instantaneous frequency (IF) estimation specifically designed for nonstationary multicomponen signals such as HRV. The new method attempts to characterize the newborn HRV using features extracted from the time–frequency (TF) domain of the signal. These features comprise the IF, the instantaneous bandwidth (IB) and instantaneous energy (IE) of the different TF components of the HRV. Applied to the HRV of both normal and seizure suffering newborns, this method clearly reveals the locations of the spectral peaks and their time-varying nature. The total energy of HRV components, ET and ratio of energy concentrated in the low-frequency (LF) to that in high frequency (HF) components have been shown to be significant features in identifying the HRV of newborn with seizures.

Time- frequency based renyi entropy of heart rate variability for newborn seizure detection

The time-frequency (TF) version of Renyi entropy, which measures the information content and complexity of a signal, is used here as a feature in the classification of the newborn heart rate variability (HRV) as either corresponding to seizure or non-seizure. The newborn HRV is initially mapped to the TF domain using the modified B distribution (MBD). The time-frequency distribution (TFD) of HRV is post-processed before the Renyi entropy is computed. This post-processing method uses an image processing technique called component linking to identify the true HRV components and localize them in the TF plane. The results obtained so far show that the HRV corresponding to non-seizure can be discriminated from those corresponding to seizure using TF-based Renyi entropy with 78.57% sensitivity and 83.33 % specificity.

Resolution analysis of the T-class time-frequency distributions

The T-class of time-frequency distributions (TFDs) is a newly proposed subclass of the general quadratic TFD class. The TFDs of this subclass are characterized by their time-lag kernels which are functions of time only. In this paper, we report the results of our investigation related to the time and frequency resolution of the T-class TFDs. It is shown, analytically, for the case of a linear chirp, that the frequency resolution is a function of both the smoothing parameter of the TFD kernel and the rate of change in instantaneous frequency. The results are then illustrated with a number of synthetic examples.

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.

Analysis of discretization errors in if estimation of polynomial phase signals

The peak of the polynomial Wigner-Ville distribution (PWVD) is a method for providing an unbiased estimate of the instantaneous frequency (IF) for polynomial phase signals. The theoretical lower variance bound, assuming a continuous frequency variable, has been studied previously. However, due to the discretization of the PWVD required for computer implementation, there is also another theoretical lower variance bound which is a result of the discretization error. In this paper, we study the relationship between the discretization error bound and the theoretical lower variance bound and determine the minimum number of frequency samples required such that the theoretical lower variance bound can be attained.