Your browser does not support JavaScript!
http://iet.metastore.ingenta.com
1887

IEE Proceedings - Science, Measurement and Technology

Volume 147, Issue 6, November 2000

Volumes & issues:

Volume 147, Issue 6

November 2000

Show / Hide details

  • Medical signal processing

    • Editorial: Medical signal processing
      Robust methodology for the discrimination of brain tumours from in vivo magnetic resonance spectra
      • Author(s): Y.Y.B. Lee ; Y. Huang ; W. El-Deredy ; P.J.G. Lisboa ; C. Arus ; P. Harris
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 309 –314
      • DOI:  10.1049/ip-smt:20000850
      • Type: Article
      • + Show details - Hide details
      • p. 309 –314 (6)
        Magnetic resonance (MR) spectroscopy provides a direct non-invasive measure of tissue biochemistry, but tissue heterogeneity causes considerable mixing between tissue categories. A systematic methodology for variable selection and performance estimation, applied to 98 in vivo spectra from cysts and five categories of brain tumour is proposed. The selection of predictive variables from the spectra, and the estimation of misclassification errors, are made robust by pre-filtering the irrelevant spectral components and repeatedly applying bootstrap resampling. Three alternative approaches to the methodology were investigated, with reference to pairwise discriminant models. The first approach is applied directly to the spectral intensity values, treated as independent covariates that are interpreted as metabolite indicators, proceeding to search for the smallest number of metabolites necessary for class discrimination. The two other approaches use independent component analysis (ICA) to separate the heterogeneous spectra into a small number of independent spectral sources of intrinsic tissue types. Given the six classes with strong inter-class mixing, the most accurate classifier based on linear discriminant models is obtained by first optimising the discrimination between class pairs, then combining their outcome using a pairwise coupling method. Finally, the statistical and ICA pre-processing methods are compared in a retrospective study for the first class assignment pair, to separate low- and medium-grade from high-grade astrocytic tumours.
        Add to favourites
        Export citations
      Using dynamical embedding to isolate seizure components in the ictal EEG
      • Author(s): C. James  and  D. Lowe
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 315 –320
      • DOI:  10.1049/ip-smt:20000849
      • Type: Article
      • + Show details - Hide details
      • p. 315 –320 (6)
        A system for isolating seizure components in the epileptic electroencephalogram (EEG) is presented. The method of independent component analysis (ICA) is implemented to decompose multichannel recordings of scalp EEG known to contain epileptic seizures into their underlying independent components (ICs). The ICs are treated as a convenient expansion basis and in order to identify the relevant seizure components amongst the ICs, a series of dynamical embedding matrices are first constructed along each IC. By observing the change in structure of the singular spectra obtained by performing a singular value decomposition on each consecutive dynamical embedding matrix, it is possible to track changes in the underlying complexity of each IC with time. The change in complexity is linked to the change in entropy that can be calculated from each consecutive singular spectrum. The change in complexity, coupled with the topographical distribution of each IC, allows seizure-related components extracted by the ICA process to be subjectively identified. The method has been applied to four seizure EEG segments, and in each case probable seizure components were identified subjectively. As a proof-of-principle study, the method indicates that ICA coupled with dynamical embedding may be useful as a tool in pre-processing seizure EEG segments.
        Add to favourites
        Export citations
      Prospects for routine detection of dementia using the fractal dimension of the human electroencephalogram
      • Author(s): G.T. Henderson ; E.C. Ifeachor ; H.S.K. Wimalaratna ; E.M. Allen ; N.R. Hudson
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 321 –326
      • DOI:  10.1049/ip-smt:20000862
      • Type: Article
      • + Show details - Hide details
      • p. 321 –326 (6)
        The paper datails research which aims to improve the contribution made by electroencephalogram (EEG) analysis to the diagnosis and care of patients with brain disease; dementia in particular. Previous attempts to automate EEG analysis have concentrated on separating patient groups from control groups, often on the basis of a single neurophysiological index derived from a short, isolated segment of EEG. The authors seek to develop, and test, a novel technique for the analysis of changes in serial EEG recordings on individuals (subject-specific analysis) which may serve as a basis for routine early detection of dementia. The objectives of the reported study were to examine the feasibility of applying appropriate fractal dimension (FD) (complexity) measures to the human EEG, and to examine whether methods using the subject specific variability of these measures are likely to be useful for detecting patients who develop dementia. The reason for undertaking the study was to establish a ‘proof of concept’ and determine whether research should concentrate in this area. Existing EEG analysis methods were reviewed and four FD measures suitable for EEG analysis were developed. These four measures were applied to a total of 21 EEG recordings (from seven subjects with various dementias, eight age matched controls and two young subjects who gave three recordings each). The results were analysed and the following conclusions were drawn: it is possible to measure the complexity of the human EEG using the FD, and the subject specific variability of the FD is an important candidate method for identifying patients with dementia. Therefore, further work in this area is justified.
        Add to favourites
        Export citations
      Nonlinear analysis of carotid artery echographic images
      • Author(s): V. Positano ; R. Mammoliti ; M.F. Santarelli ; L. Landini ; A. Benassi
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 327 –332
      • DOI:  10.1049/ip-smt:20000848
      • Type: Article
      • + Show details - Hide details
      • p. 327 –332 (6)
        Nonlinear analysis is applied to identifying complex spatial patterns in echographic images of normal and pathologic carotid arteries. Complexity and entropy measures of normal and atherosclerotic plaques are evaluated to characterise the space-temporal evolution of biological patterns. They are: correlation dimension, Lyapunov exponent and Kolmogorov entropy. The application of principal component analysis to such measures clusters data according to different atherosclerosis severity degrees, which are confirmed by histologic analysis.
        Add to favourites
        Export citations
      Automated assessment of vigilance using committees of radial basis function analysers
      Visualisation and categorisation of respiratory mechanism using self organising maps
      • Author(s): M. Steuer ; P. Caleb ; G.B. Drummond ; A.M.S. Black
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 339 –344
      • DOI:  10.1049/ip-smt:20000856
      • Type: Article
      • + Show details - Hide details
      • p. 339 –344 (6)
        In post-operative patients it is sometimes necessary to push morphine-like analgesics to their limits for pain relief. Unfortunately, this can sometimes bring a significant risk of disrupting the control of breathing, and of precipitating life-threatening conditions. A possible way of monitoring patients is by studying the correlation between analgesia, airway obstruction and hypoxia. The first step towards achieving this objective is by visualising the relationship between different pairs of signals involved in respiratory mechanics. Based on previous work, where self-organising maps (SOMs) were used for representing these relationships on a breath by breath basis, it is demonstrated how it is now possible to automatically label nodes in the SOMs based on classification of the signals by a clinician. The use of a majority voting configuration of SOMs enables results to be presented with a confidence measure which enhances the medical applicability of the system. In addition, the ability to now visualise the transition between categories will enable further research into the significance of transition between the categories and the presence of possible new sub-categories.
        Add to favourites
        Export citations
      Learning interaction dynamics with coupled hidden Markov models
      • Author(s): I. Rezek ; P. Sykacek ; S.J. Roberts
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 345 –350
      • DOI:  10.1049/ip-smt:20000851
      • Type: Article
      • + Show details - Hide details
      • p. 345 –350 (6)
        An analysis of interactions between different physiological control systems may only be possible with correlation functions if the signals have similar spectral distributions. Interactions between such signals can be modelled in state space rather than observation space, i.e. interactions are modelled after first translating the observations into a common domain. Coupled hidden Markov models (CHMM) are such state-space models. They form a natural extension to standard hidden Markov models. The authors perform CHMM parameter estimation under a Bayesian paradigm, using Gibbs sampling, and in a maximum likelihood framework, using the expectation maximisation algorithm. The performance differences between the estimators are demonstrated on simulated data as well as biomedical data. It is shown that the proposed method gives meaningful results when comparing two different signals, such as respiration and EEG.
        Add to favourites
        Export citations
      Robust ECG R-R wave detection using evolutionary programming-based fuzzy inference system (EPFIS), and application to assessing brain-gut interaction
      • Author(s): Z.S. Wang  and  J.D.Z. Chen
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 351 –356
      • DOI:  10.1049/ip-smt:20000852
      • Type: Article
      • + Show details - Hide details
      • p. 351 –356 (6)
        Some recent studies have shown that gastric electrical stimulation can entrain gastric dysrhythmia, reduce chronic symptoms and accelerate gastric emptying. However, possible mechanisms involved remain unknown. It is investigated whether or not electrical stimulation is vagally mediated by assessing the heart rate variability (HRV). The study is performed in six healthy female hound dogs implanted with four pairs of bipolar serosal electrodes, which are used to measure gastric myoelectrical activity. A special fuzzy neural network, which is called the evolutionary programming-based fuzzy inference system (EPFIS), is developed to identify the R-R wave to precisely extract the R-R interval and derive the HRV data. A high-resolution adaptive time-frequency analysis method based on ARMA modelling previously developed by the author is used to obtain high quality HRV spectral parameters.
        Add to favourites
        Export citations
      Aspects of diagnostic schemes for biomedical and engineering systems
      • Author(s): N.B. Jones ; S.K. Spurgeon ; M.J. Pont ; J.A. Twiddle ; C.L. Lim ; C.R. Parikh ; K.B. Goh
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 357 –362
      • DOI:  10.1049/ip-smt:20000859
      • Type: Article
      • + Show details - Hide details
      • p. 357 –362 (6)
        The paper considers aspects of work carried out on projects in medical diagnostics and engineering fault detection and control, and shows the development of some ideas in signal processing and diagnostic techniques; in particular regarding models, estimators and decision schemes. The fact that analysis of human biological systems is restricted by powerful ethical and legal constraints would seem to lead to fundamental differences of approach to medical and engineering problems. In the research considered this did not seem to be a dominant issue. The use of system simulators as reference standards, and blackboard schemes for knowledge integration, has shown to advantage in both fields. Methods for increasing the usefulness of limited data have been successfully applied to decreasing the cost associated with sensors for a given standard of diagnostic reliability. The science of engineering control and diagnostics has benefited from the use of observers based on state-space descriptions of the system. An observer provides an alternative reference model which has been used to generate residuals when faults occur. A new technique employing sliding-mode observers to recreate fault signals is introduced. This idea is extended for use in a medical context.
        Add to favourites
        Export citations
      Extreme value statistics for novelty detection in biomedical data processing
      • Author(s): S.J. Roberts
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 363 –367
      • DOI:  10.1049/ip-smt:20000841
      • Type: Article
      • + Show details - Hide details
      • p. 363 –367 (5)
        Extreme value theory (EVT) is a branch of statistics which concerns the distributions of data of unusually low or high value i.e. in the tails of some distribution. These extremal points are important in many applications as they represent the outlying regions of normal events against which we may wish to define novel events. The use of such novelty detection approaches is useful for analysis of data for which few exemplars of some important class exist, for example in medical screening. It is shown that a principled approach to the issue of novelty detection may be taken using extreme value statistics.
        Add to favourites
        Export citations
      Time-frequency analysis of short segments of biomedical data
      • Author(s): C. Bigan  and  M.S. Woolfson
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 368 –373
      • DOI:  10.1049/ip-smt:20000840
      • Type: Article
      • + Show details - Hide details
      • p. 368 –373 (6)
        An evaluation is made of two novel methods in the tracking of variations in frequency of the various components present in an electroencephalogram (EEG) signal, for example when the subject is undergoing an epileptic seizure. In one method, the polynomial modelling method, the EEG is broken down into its constituent components using repeated polynomial modelling and zero crossing analysis is employed to characterise the time variation of the frequency of each component. In the second method, the phase compensation method, the signal is modelled as several cosines and the individual components are estimated and subtracted off the total signal; the phase derivative of each component is used to estimate the frequency. These methods are then compared with the conventional short time Fourier transform (STFT) and the high-order Yule–Walker (HOYW) methods. Using simulated data it is shown that the polynomial modelling method has the best performance in terms of breaking down the EEG into its constituent components. However, the HOYW and phase compensation methods provide estimates of the frequencies with lower variance. The three non-Fourier based methods are sensitive to the presence of noise and give similar estimates of the frequencies when applied to experimental non-pathological EEG data. The paper concludes with suggestions for combining the two novel methods to obtain a better frequency tracking performance.
        Add to favourites
        Export citations
      Parametric bispectral estimation of EEG signals in different functional states of the brain
      • Author(s): M. Shen ; L. Sun ; F.H.Y. Chan ; P.J. Beadle
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 374 –377
      • DOI:  10.1049/ip-smt:20000847
      • Type: Article
      • + Show details - Hide details
      • p. 374 –377 (4)
        Higher-order statistics is applied to the analysis of electroencephalograms (EEG) in order to investigate the non-Gaussianality and nonlinearity of EEG signals. The parametric bispectral estimate is proposed for the purpose of extracting more information beyond second order statistics. The actual EEGs, with normal subjects in several different functional states of the brain, are analysed in terms of the parametric bispectral estimate. The experimental results show that all kinds of spontaneous EEG exhibit obvious quadratic nonlinear interactions of EEG signals, but the bispectral pattern of normal EEG changes with different functional states of the brain. It is suggested that the bispectrum could be regarded as the main feature in the study of EEG signals, and an effective quantitative measure for analysing and processing electroencephalography in different physiological states of the brain is provided.
        Add to favourites
        Export citations
      Effect of cross-correlation on auditory evoked response detection
      • Author(s): D. Sprevak ; H.G. McAllister ; P.J. McCullagh
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 378 –381
      • DOI:  10.1049/ip-smt:20000845
      • Type: Article
      • + Show details - Hide details
      • p. 378 –381 (4)
        A procedure to quantify the relationship between the signal-to-noise ratio (SNR) and the number of observed series of the auditory brainstem response is discussed. The commonly held belief that the SNR increases as the number of terms being averaged originates from the assumption that repeated series of the evoked potentials are perfectly correlated. It is shown that in practice the correlation is far from perfect and this leads to an underestimate of the number of observations which are required to obtain a desired SNR. The presented formula quantifies this dependency for all values of the cross-correlation between series, and reduces to the accepted form for perfectly correlated signals.
        Add to favourites
        Export citations
      Extracting single trial event related potentials
      • Author(s): J. Britton ; B.W. Jervis ; R.A. Grünewald
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 382 –388
      • DOI:  10.1049/ip-smt:20000842
      • Type: Article
      • + Show details - Hide details
      • p. 382 –388 (7)
        Established techniques for the analysis of event related potentials (ERPs) involve averaging of time-locked sections of the EEG signal over many trials to extract the ERP waveform from the ongoing EEG noise. Several methods have been developed to enable extraction of single trial ERPs. A quantitative method was developed for testing the accuracy of single trial ERP estimates. ERP signals were simulated by using a piece-wise model of a well known ERP. A large number of unique ERPs were generated by randomly varying parameters of the model. Each of these was embedded in simulated EEG noise modelled as an auto-regressive process driven by white noise. Both stationary and nonstationary noise was simulated. The known simulated ERPs were then compared to the corresponding estimates produced by single trial ERP extraction techniques in terms of the amount of distortion introduced. The techniques tested were time sequence adaptive filtering, singularity detection using wavelets, adaptive multi-resolution analysis and a modification of the multi-resolution analysis technique. None of the methods was found to extract sufficiently accurate waveforms from single trial ERPs contaminated with realistic EEG noise. Improved, but still unsatisfactory, ERP estimates were obtained when the AR EEG noise was replaced by Gaussian noise.
        Add to favourites
        Export citations
      Novel decision strategy for P-wave detection utilising nonlinearly synthesised ECG components and their enhanced pseudospectral resonances
      • Author(s): M. Sabry-Rizk ; E.R. Carson ; W. Zgallai ; C. Morgan ; S. El-Khafif ; K.T.V. Grattan
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 389 –396
      • DOI:  10.1049/ip-smt:20000855
      • Type: Article
      • + Show details - Hide details
      • p. 389 –396 (8)
        A novel, robust, patient adaptive and potentially on-line P-wave detection decision strategy is presented. It is particularly well-suited to the noisy ECG environment, and highly effective with abnormal atrial activations which can occur simultaneously with the QRS complex or the T-wave. Essentially, it is based on the exploitation of the unique structural properties of the raw and nonlinearly synthesised high-resolution pseudospectral multi-resonance signatures of the P-wave (or the entire PR interval including the P-wave), QRS complex, and the combined S-T segment and the T-wave. Clinical trials involving 59 healthy volunteers, equivalent to 1600 hours of ECG recordings, sampled at 500 Hz, have confirmed that most of the atrium and ventricle depolarisation-wave spectral energies (approximately 90%) are confined to two overlapping clusters inscribed by the top of the principal pseudospectral peaks (PPPs) and centred at 12 and 15 Hz, respectively. The degree of overlapping varies from one person to another and can, to some extent, be reduced by using a temporal interleaving technique or sub-sampling. Normal sinus records obtained from the MIT/BIH database and sampled at a lower rate of 128 Hz have confirmed the above, thus facilitating an easy decision strategy for all the P-waves that bear relationships to the following QRS complexes. For the P-on-QRS and the P-on-T episodes, the linear and quadratic PPPs are used in the detection of independent P-waves, respectively. The results presented here include a representative of 59 cases of sinus rhythm and three cases of challenging arrhythmias. One observation of mechanical contraction of the atrium and using echocardiographic imaging coinciding with the P-wave is presented to demonstrate the clinical fall-outs resulting from detection of dependent as well as independent P-waves. In arrhythmias cases, the standard 12-lead or the 3-lead machines were used as appropriate. The multi-faceted detection routine employs temporal and frequency interleaving, the LMS-based cubic Volterra synthesiser in conjunction with Kaiser-filtered state-enhanced spectral MUSIC, and subsequent thresholding. An adaptive multi-level thresholding with first derivative calculations is only activated in suspected cases of several independent P-waves perturbing the QRS complex.
        Add to favourites
        Export citations
      Virtual instrumentation for clinical assessment of cardiovascular and autonomic function
      • Author(s): D.S. Benitez ; A. Zaidi ; A. Fitchet ; P.A. Gaydecki ; A.P. Fitzpatrick
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 397 –402
      • DOI:  10.1049/ip-smt:20000843
      • Type: Article
      • + Show details - Hide details
      • p. 397 –402 (6)
        Information about the activity of the autonomic nervous system can be derived by analysis of key physiological signals, including ECG and blood pressure, on a beat-to-beat basis. Due to the large amount of data and mathematical processing involved, analysis by hand is prohibitive. The solution developed facilitates this procedure by performing the processing on a PC-based instrument incorporating existing medical equipment, a data acquisition board and a computer program to acquire and process the physiological signals. The automated virtual instrument was written using LabVIEW© as the main platform. The techniques implemented include impedance cardiography, time- and frequency-domain analysis, invasive and non-invasive baroreflex sensitivity assessment, and forearm blood flow measurements. The system was designed to study patients suffering vasovagal blackouts, where the changes that occur in the heart and circulation during an attack could increase understanding of the physiological processes that underlie their blackouts.
        Add to favourites
        Export citations
      Variability of photoplethysmography peripheral pulse measurements at the ears, thumbs and toes
      • Author(s): J. Allen  and  A. Murray
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 403 –407
      • DOI:  10.1049/ip-smt:20000846
      • Type: Article
      • + Show details - Hide details
      • p. 403 –407 (5)
        Pulses radiating to the periphery after each heart beat can be detected by photoplethysmography (PPG). A multi-site PPG measurement and analysis system is described for beat-to-beat analysis of the pulse waveforms. Waveforms were collected from the right and left ear lobes, thumb pads, and toe pads of 20 normal subjects whilst they performed paced breathing. The median coefficients of variation for pulse amplitude were; ears and thumbs 11%, and toes 14%. Significant differences in pulse transit time of 4 ms (to foot of pulse), and 14 ms (to peak of pulse) between the right and left toes were detected (right lagging left, p < 0.01). The median pulse transit times (to foot) were; ears (0.133 s), thumbs (0.199 s), and toes (0.301 s), with coefficients of variation; 2.8%, 2.2%, and 1.6%, respectively. The median pulse transit times (to peak) were; ears (0.397 s), thumbs (0.436 s), and toes (0.515 s), with coefficients of variation; 1.4%, 2.0%, and 1.5%, respectively. The authors have utilised signal processing algorithms to calculate beat-to-beat measures of pulsatility, and determined normative data for multi-site PPG pulse assessments.
        Add to favourites
        Export citations
      New range-based neighbourhood operator for extracting edge and texture information from mammograms for subsequent image segmentation and analysis
      • Author(s): R. Chandrasekhar  and  Y. Attikiouzel
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 408 –413
      • DOI:  10.1049/ip-smt:20000853
      • Type: Article
      • + Show details - Hide details
      • p. 408 –413 (6)
        A new, spatially isotropic, range-based, neighbourhood operator is described, and its use for extracting edge and texture information from mammograms is illustrated. It is an extension of an operator, first introduced by J.C. Russ, and founded on the work of H.E. Hurst. An octagonal neighbourhood is defined, centred on each pixel in an image, and the difference between the maximum and minimum pixel values in each set of pixels at a given Euclidean distance from the centre pixel is computed to be the range. The logarithm of the range is plotted against the logarithm of the distance, and a straight line fitted to the data. The y-axis intercept c and the square of the correlation coefficient, η2, are associated with the position of the centre pixel. Preliminary experiments suggest that c is edge-sensitive and could be useful in detecting weak edges, with good noise immunity, while η2 is a promising texture measure that may be used to detect the edge of the pectoral muscle, define the boundary of the mammographic parenchyma, and in conjunction with other features, possibly detect circumscribed lesions.
        Add to favourites
        Export citations
      Adaptive combination of linear predictors for lossless image compression
      • Author(s): G. Deng ; H. Ye ; L.W. Cahill
      • Source: IEE Proceedings - Science, Measurement and Technology, Volume 147, Issue 6, p. 414 –419
      • DOI:  10.1049/ip-smt:20000854
      • Type: Article
      • + Show details - Hide details
      • p. 414 –419 (6)
        Lossless image coding is an essential requirement for medical imaging applications. Lossless image compression techniques usually have two major components: adaptive prediction and adaptive entropy coding. The paper is concerned with adaptive prediction. Recently, several researchers have studied prediction schemes in which the final prediction is formed by a combination of a group of subpredictors. The authors present an overview of this new type of prediction technique. They show that the basic principle of adaptive predictor combination has been extensively studied and applied to many science and engineering problems. They then describe their own combination scheme, which is based on the estimation of the local prediction error variance. Experimental results show that the compression performance of the algorithms that employ this new type of predictor is consistently better than that of state-of-the-art algorithms.
        Add to favourites
        Export citations

Most viewed content for this Journal

Article
content/journals/ip-smt
Journal
5
Loading

Most cited content for this Journal

We currently have no most cited data available for this content.

Print this page
This is a required field
Please enter a valid email address