ORCID Profile
0000-0003-0939-9145
Current Organisation
University of Southern Queensland
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
In Research Link Australia (RLA), "Research Topics" refer to ANZSRC FOR and SEO codes. These topics are either sourced from ANZSRC FOR and SEO codes listed in researchers' related grants or generated by a large language model (LLM) based on their publications.
Biomedical Engineering Not Elsewhere Classified | Clinical Engineering | Biomedical Engineering | Engineering/Technology Instrumentation |
Publisher: IEEE
Date: 12-2010
Publisher: Springer International Publishing
Date: 2017
Publisher: IEEE
Date: 12-2010
Publisher: IEEE
Date: 2013
Publisher: Informa UK Limited
Date: 15-05-2014
DOI: 10.1080/10255842.2012.683428
Abstract: The biological microenvironment is interrupted when tumour masses are introduced because of the strong competition for oxygen. During the period of avascular growth of tumours, capillaries that existed play a crucial role in supplying oxygen to both tumourous and healthy cells. Due to limitations of oxygen supply from capillaries, healthy cells have to compete for oxygen with tumourous cells. In this study, an improved Krogh's cylinder model which is more realistic than the previously reported assumption that oxygen is homogeneously distributed in a microenvironment, is proposed to describe the process of the oxygen diffusion from a capillary to its surrounding environment. The capillary wall permeability is also taken into account. The simulation study is conducted and the results show that when tumour masses are implanted at the upstream part of a capillary and followed by normal tissues, the whole normal tissues suffer from hypoxia. In contrast, when normal tissues are ahead of tumour masses, their pO2 is sufficient. In both situations, the pO2 in the whole normal tissues drops significantly due to the axial diffusion at the interface of normal tissues and tumourous cells. As the existence of the axial oxygen diffusion cannot supply the whole tumour masses, only these tumourous cells that are near the interface can be partially supplied, and have a small chance to survive.
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 11-2014
Publisher: IEEE
Date: 05-2007
Publisher: Elsevier BV
Date: 2009
DOI: 10.1016/J.COMPBIOMED.2008.10.007
Abstract: In this paper, we present a new method to identify anesthetic states based on routinely recorded electroencephalogram (EEG). The identification of anesthesia stages are conducted using fast Fourier transform (FFT) and modified detrended fluctuation analysis (DFA) method. Simulation results demonstrate that this new method can clearly discriminate all five anesthesia states: very deep anesthesia, deep anesthesia, moderate anesthesia, light anesthesia and awake.
Publisher: The Hong Kong Institution of Engineers
Date: 2010
Publisher: Elsevier BV
Date: 2010
Publisher: Informa UK Limited
Date: 2005
Publisher: Springer Science and Business Media LLC
Date: 09-2003
DOI: 10.1007/BF03178780
Publisher: IEEE
Date: 07-2012
Publisher: IEEE
Date: 12-2008
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2016
Publisher: Elsevier BV
Date: 11-2021
Publisher: IEEE
Date: 03-2007
Publisher: Hindawi Limited
Date: 08-09-2022
DOI: 10.1155/2022/1992596
Abstract: Schizophrenia (SZ) is a severe and prolonged disorder of the human brain where people interpret reality in an abnormal way. Traditional methods of SZ detection are based on handcrafted feature extraction methods (manual process), which are tedious and unsophisticated, and also limited in their ability to balance efficiency and accuracy. To solve this issue, this study designed a deep learning-based feature extraction scheme involving the GoogLeNet model called “SchizoGoogLeNet” that can efficiently and automatically distinguish schizophrenic patients from healthy control (HC) subjects using electroencephalogram (EEG) signals with improved performance. The proposed framework involves multiple stages of EEG data processing. First, this study employs the average filtering method to remove noise and artifacts from the raw EEG signals to improve the signal-to-noise ratio. After that, a GoogLeNet model is designed to discover significant hidden features from denoised signals to identify schizophrenic patients from HC subjects. Finally, the obtained deep feature set is evaluated by the GoogleNet classifier and also some renowned machine learning classifiers to find a sustainable classification method for the obtained deep feature set. Experimental results show that the proposed deep feature extraction model with a support vector machine performs the best, producing a 99.02% correct classification rate for SZ, with an overall accuracy of 98.84%. Furthermore, our proposed model outperforms other existing methods. The proposed design is able to accurately discriminate SZ from HC, and it will be useful for developing a diagnostic tool for SZ detection.
Publisher: Inderscience Publishers
Date: 2011
Publisher: Informa UK Limited
Date: 2009
Publisher: American Scientific Publishers
Date: 09-2011
Publisher: Inderscience Publishers
Date: 2011
Publisher: Elsevier BV
Date: 07-2014
DOI: 10.1016/J.CMPB.2014.04.001
Abstract: This paper proposes a fast weighted horizontal visibility graph constructing algorithm (FWHVA) to identify seizure from EEG signals. The performance of the FWHVA is evaluated by comparing with Fast Fourier Transform (FFT) and s le entropy (S En) method. Two noise-robustness graph features based on the FWHVA, mean degree and mean strength, are investigated using two chaos signals and five groups of EEG signals. Experimental results show that feature extraction using the FWHVA is faster than that of S En and FFT. And mean strength feature associated with ictal EEG is significant higher than that of healthy and inter-ictal EEGs. In addition, an 100% classification accuracy for identifying seizure from healthy shows that the features based on the FWHVA are more promising than the frequency features based on FFT and entropy indices based on S En for time series classification.
Publisher: Elsevier BV
Date: 04-2018
Publisher: Springer International Publishing
Date: 2013
Publisher: IEEE
Date: 05-2007
Publisher: IEEE
Date: 2012
Publisher: Elsevier BV
Date: 12-2017
Publisher: Elsevier BV
Date: 11-2009
Publisher: Elsevier BV
Date: 09-2016
DOI: 10.1016/J.CMPB.2016.05.022
Abstract: This paper focuses on electroconvulsive therapy (ECT) and head models to investigate temperature profiles arising when anisotropic thermal and electrical conductivities are considered in the skull layer. The aim was to numerically investigate the threshold for which this therapy operates safely to the brain, from the thermal point of view. A six-layer spherical head model consisting of scalp, fat, skull, cerebro-spinal fluid, grey matter and white matter was developed. Later on, a realistic human head model was also implemented. These models were built up using the packages from COMSOL Inc. and Simpleware Ltd. In these models, three of the most common electrode montages used in ECT were applied. Anisotropic conductivities were derived using volume constraint and included in both spherical and realistic head models. The bio-heat transferring problem governed by Laplace equation was solved numerically. The results show that both the tensor eigenvalues of electrical conductivity and the electrode montage affect the maximum temperature, but thermal anisotropy does not have a significant influence. Temperature increases occur mainly in the scalp and fat, and no harm is caused to the brain by the current applied during ECT. The work assures the thermal safety of ECT and also provides a numerical method to investigate other non-invasive therapies.
Publisher: National Taiwan University
Date: 17-03-2014
DOI: 10.4015/S1016237214500409
Abstract: This article reports on a comparative study to identify electroencephalography (EEG) signals during motor imagery (MI) for motor area EEG and all-channels EEG in the brain–computer interface (BCI) application. In this paper, we present two algorithms: CC-LS-SVM and CC-LR for MI tasks classification. The CC-LS-SVM algorithm combines the cross-correlation (CC) technique and the least square support vector machine (LS-SVM). The CC-LR algorithm assembles the CC technique and binary logistic regression (LR) model. These two algorithms are implemented on the motor area EEG and the all-channels EEG to investigate how well they perform and also to test which area EEG is better for the MI classification. These two algorithms are also compared with some existing methods which reveal their competitive performance during classification. Results on both datasets, IVa and IVb from BCI Competition III, show that the CC-LS-SVM algorithm performs better than the CC-LR algorithm on both the motor area EEG and the all-channels EEG. The results also demonstrate that the CC-LS-SVM algorithm performs much better for the all-channels EEG than for the motor area EEG. Furthermore, the LS-SVM-based approach can correctly identify the discriminative MI tasks, demonstrating the algorithm's superiority in classification performance over some existing methods.
Publisher: IEEE
Date: 08-2011
Publisher: Springer Science and Business Media LLC
Date: 21-06-2016
DOI: 10.1007/S13246-016-0459-5
Abstract: This paper presents a new method to apply timing characteristics of electroencephalograph (EEG) beta frequency bands to assess the depth of anaesthesia (DoA). Firstly, the measured EEG signals are denoised and decomposed into 20 different frequency bands. The Mobility (M), permutation entropy (PE) and Lempel-Ziv complexity (LCZ) of each frequency band are calculated. The M, PE and LCZ values of beta frequency bands (21.5-30 Hz) are selected to derive a new index. The new index is evaluated and compared with measured bispectral (BIS). The results show that there is a very close correlation between the proposed index and the BIS during different anaesthetic states. The new index also shows a 25-264 s earlier time response than BIS during the transient period of anaesthetic states. In addition, the proposed index is able to continuously assess the DoA when the quality of signal is poor and the BIS does not have any valid outputs.
Publisher: Elsevier BV
Date: 2023
Publisher: Elsevier BV
Date: 02-2019
Publisher: Springer Berlin Heidelberg
Date: 2012
Publisher: IEEE
Date: 07-2010
Publisher: IEEE
Date: 05-2011
Publisher: IEEE
Date: 07-2010
Publisher: IEEE
Date: 07-2010
Publisher: IOP Publishing
Date: 16-04-2014
DOI: 10.1088/1741-2560/11/3/036002
Abstract: Computational methods are increasingly used to optimize transcranial direct current stimulation (tDCS) dose strategies and yet complexities of existing approaches limit their clinical access. Since predictive modelling indicates the relevance of subject athology based data and hence the need for subject specific modelling, the incremental clinical value of increasingly complex modelling methods must be balanced against the computational and clinical time and costs. For ex le, the incorporation of multiple tissue layers and measured diffusion tensor (DTI) based conductivity estimates increase model precision but at the cost of clinical and computational resources. Costs related to such complexities aggregate when considering in idual optimization and the myriad of potential montages. Here, rather than considering if additional details change current-flow prediction, we consider when added complexities influence clinical decisions. Towards developing quantitative and qualitative metrics of value/cost associated with computational model complexity, we considered field distributions generated by two 4 × 1 high-definition montages (m1 = 4 × 1 HD montage with anode at C3 and m2 = 4 × 1 HD montage with anode at C1) and a single conventional (m3 = C3-Fp2) tDCS electrode montage. We evaluated statistical methods, including residual error (RE) and relative difference measure (RDM), to consider the clinical impact and utility of increased complexities, namely the influence of skull, muscle and brain anisotropic conductivities in a volume conductor model. Anisotropy modulated current-flow in a montage and region dependent manner. However, significant statistical changes, produced within montage by anisotropy, did not change qualitative peak and topographic comparisons across montages. Thus for the ex les analysed, clinical decision on which dose to select would not be altered by the omission of anisotropic brain conductivity. Results illustrate the need to rationally balance the role of model complexity, such as anisotropy in detailed current flow analysis versus value in clinical dose design. However, when extending our analysis to include axonal polarization, the results provide presumably clinically meaningful information. Hence the importance of model complexity may be more relevant with cellular level predictions of neuromodulation.
Publisher: Institution of Engineering and Technology (IET)
Date: 09-2007
Publisher: ACM
Date: 21-08-2020
Publisher: IEEE
Date: 02-2011
Publisher: Institution of Engineering and Technology (IET)
Date: 04-2017
Publisher: Springer Science and Business Media LLC
Date: 12-2012
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 07-2011
Publisher: Springer Science and Business Media LLC
Date: 26-05-2010
DOI: 10.1007/S13246-010-0015-7
Abstract: The accuracy of an electroencephalography (EEG) forward problem partially depends on the head tissue conductivities. These conductivities are anisotropic and inhomogeneous in nature. This paper investigates the effects of conductivity uncertainty and analyses its sensitivity on an EEG forward problem for a spherical and a realistic head models. We estimate the uncertain conductivities using an efficient constraint based on an optimization method and perturb it by means of the volume and directional constraints. Assigning the uncertain conductivities, we construct spherical and realistic head models by means of a stochastic finite element method for fixed dipolar sources. We also compute EEG based on the constructed head models. We use a probabilistic sensitivity analysis method to determine the sensitivity indexes. These indexes characterize the conductivities with the most or the least effects on the computed outputs. These results demonstrate that conductivity uncertainty has significant effects on EEG. These results also show that the uncertain conductivities of the scalp, the radial direction of the skull and transversal direction in the white matter are more sensible.
Publisher: IEEE
Date: 08-2016
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2000
DOI: 10.1109/51.887250
Publisher: Elsevier BV
Date: 03-2018
Publisher: Fuji Technology Press Ltd.
Date: 20-11-2011
DOI: 10.20965/JACIII.2011.P1221
Abstract: EEG-based applications have faced the challenge of multi-modal integrated analysis problems. In this paper, a greedy maximal weight matching approach is used to measure the functional connectivity in alcoholics datasets with EEG and EOG signals. The major discovery is that the processing of the repeated and unrepeated stimuli in the γ band in control drinkers is significantly more different than that in alcoholic subjects. However, the EOGs are always stable in the case of visual tasks, except for a weakly wave when subjects make an error response to the stimuli.
Publisher: Elsevier BV
Date: 2013
DOI: 10.1016/J.CMPB.2012.09.001
Abstract: The study investigates the impact of white matter directional conductivity on brain current density under the influence of Transcranial direct current stimulation (tDCS). The study employed different conductivity estimation algorithms to represent conductivity distribution in the white matter (WM) of the brain. Two procedures, one mathematically driven and the second one based on the Diffusion tensor imaging (DTI) are considered. The finite element method has been applied to estimate the current density distribution across the head models. Strengths and weaknesses of these algorithms have been compared by analyzing the variation in current density magnitude and distribution patterns with respect to the isotropic case. Results indicate that anisotropy has a profound influence on the strength of current density (up to ≈50% in WM) as it causes current flow to deviate from its isotropically defined path along with diffused distribution patterns across the gray and WM. The extent of this variation is highly correlated with the degree of the anisotropy of the regions. Regions of high anisotropy and models of fixed anisotropic ratio displayed higher and wider degree of variations across the structures (topographic variations up to 48%), respectively. In contrast, models, which are correlated with the magnitude of local diffusion tensor behaved in a less exacerbated manner (≈10% topographic changes in WM). Anisotropy increased the current density strength across the cortical gyri under and between the stimulating electrodes, whereas a significant drop has been recorded in deeper regions of the GM (max % difference ≈±10). In addition, it has been observed that Equivalent isotropic trace algorithm is more suitable to incorporate directional conductivity under tDCS paradigm, than other considered approaches, as this algorithm is computationally less expensive and insensitive to the limiting factor imposed by the volume constraint.
Publisher: Frontiers Media SA
Date: 28-06-2019
Publisher: Elsevier BV
Date: 12-2019
DOI: 10.1016/J.NEUROSCIENCE.2019.10.034
Abstract: K-complexes are important transient bio-signal waveforms in sleep stage 2. Detecting k-complexes visually requires a highly qualified expert. In this study, an efficient method for detecting k-complexes from electroencephalogram (EEG) signals based on fractal and frequency features coupled with an ensemble model of three classifiers is presented. EEG signals are first partitioned into segments, using a sliding window technique. Then, each EEG segment is decomposed using a dual-tree complex wavelet transform (DT-CWT) to a set of real and imaginary parts. A total of 10 sub-bands are used based on four levels of decomposition, and the high sub-bands are considered in this research for feature extraction. Fractal and frequency features based on DT-CWT and Higuchi's algorithm are pulled out from each sub-band and then forwarded to an ensemble classifier to detect k-complexes. A twelve-feature set is finally used to detect the sleep EEG characteristics using the ensemble model. The ensemble model is designed using a combination of three classification techniques including a least square support vector machine (LS-SVM), k-means and Naïve Bayes. The proposed method for the detection of the k-complexes achieves an average accuracy rate of 97.3 %. The results from the ensemble classifier were compared with those by in idual classifiers. Comparisons were also made with existing k-complexes detection approaches for which the same datasets were used. The results demonstrate that the proposed approach is efficient in identifying the k-complexes in EEG signals it yields optimal results with a window size 0.5 s. It can be an effective tool for sleep stages classification and can be useful for doctors and neurologists for diagnosing sleep disorders.
Publisher: Elsevier BV
Date: 2013
Publisher: IEEE
Date: 12-2008
Publisher: IEEE
Date: 12-2008
Publisher: MDPI AG
Date: 29-02-2020
DOI: 10.3390/APP10051622
Abstract: Adaptive traffic signal control (ATSC) based on deep reinforcement learning (DRL) has shown promising prospects to reduce traffic congestion. Most existing methods keeping traffic signal phases fixed adopt two agent actions to match a four-phase suffering unstable performance and undesirable operation in a four-phase signalized intersection. In this paper, a Double Deep Q-Network (DDQN) with a dual-agent algorithm is proposed to obtain a stable traffic signal control policy. Specifically, two agents are denoted by two different states and shift the control of green lights to make the phase sequence fixed and control process stable. State representations and reward functions are presented by improving the observability and reducing the leaning difficulty of two agents. To enhance the feasibility and reliability of two agents in the traffic control of the four-phase signalized intersection, a network structure incorporating DDQN is proposed to map states to rewards. Experiments under Simulation of Urban Mobility (SUMO) are carried out, and results show that the proposed traffic signal control algorithm is effective in improving traffic capacity.
Publisher: Institution of Engineering and Technology (IET)
Date: 11-2008
Publisher: IEEE
Date: 03-2007
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: IEEE
Date: 10-2012
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: Springer Science and Business Media LLC
Date: 06-2008
DOI: 10.1007/BF03178586
Publisher: IET
Date: 2011
DOI: 10.1049/CP.2011.0295
Publisher: Springer Science and Business Media LLC
Date: 31-08-2010
DOI: 10.1007/S13246-010-0027-3
Abstract: In this study, we consider different conductivity values based on tissue location in a human head model. We implement local conductivity (LC) to compute head surface potentials in three-, four-layered spherical and realistic head models using finite element method (FEM). Implementing LC for all head models, we obtain significant scalp potential variations in the term of relative difference measurement (RDM) and magnification (MAG) values with a maximum of 2.03±1.81 and 8.27±6.36, respectively. We also investigate the effects of conductivity variations (CVs) of head tissue layer on scalp potentials and find a maximum of 2.15±1.93 RDM and 8.57±6.61 MAG values. Our study concludes that it is important to assign LC to each tissue and it is also important to assign appropriate conductivity value in the construction of a head model for achieving accurate scalp potentials.
Publisher: Springer Science and Business Media LLC
Date: 13-09-2014
Publisher: Springer Science and Business Media LLC
Date: 05-01-2016
DOI: 10.1007/S13246-015-0414-X
Abstract: The alcoholism can be detected by analyzing electroencephalogram (EEG) signals. However, analyzing multi-channel EEG signals is a challenging task, which often requires complicated calculations and long execution time. This paper proposes three data selection methods to extract representative data from the EEG signals of alcoholics. The methods are the principal component analysis based on graph entropy (PCA-GE), the channel selection based on graph entropy (GE) difference, and the mathematic combinations channel selection, respectively. For comparison purposes, the selected data from the three methods are then classified by three classifiers: the J48 decision tree, the K-nearest neighbor and the Kstar, separately. The experimental results show that the proposed methods are successful in selecting data without compromising the classification accuracy in discriminating the EEG signals from alcoholics and non-alcoholics. Among them, the proposed PCA-GE method uses only 29.69% of the whole data and 29.5% of the computation time but achieves a 94.5% classification accuracy. The channel selection method based on the GE difference also gains a 91.67% classification accuracy by using only 29.69% of the full size of the original data. Using as little data as possible without sacrificing the final classification accuracy is useful for online EEG analysis and classification application design.
Publisher: IEEE
Date: 2003
Publisher: IEEE
Date: 07-2010
Publisher: IEEE
Date: 05-2011
Publisher: Springer Science and Business Media LLC
Date: 21-11-2014
DOI: 10.1007/S13246-014-0309-2
Abstract: This paper introduces a new method addressing depth of anaesthesia (DoA) assessment for real-time monitoring. The new method uses a combination of phase and litude of electroencephalogram (EEG) signals to assess the DoA level. A strong analytical signal transform is applied to extract the phase and litude information of the recorded EEG signals. Based on the extracted features from the EEG signal in each different frequency band, a new DoA index is developed. The proposed new DoA index is evaluated using data from adult patients in an age range from 22 to 75 years. The results show that the new DoA index is able to detect the changing pattern of EEG signals early and agree with the clinical notes of an attending anaesthetist. The results are also closely correlated with the popular BIS index. Furthermore, the proposed new DoA index is able to detect the state changes earlier than the BIS index.
Publisher: IEEE
Date: 05-2007
Publisher: IEEE
Date: 12-2010
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: Institution of Engineering and Technology (IET)
Date: 2012
Publisher: AIP
Date: 2013
DOI: 10.1063/1.4824993
Publisher: Elsevier BV
Date: 06-2012
DOI: 10.1016/J.COMPBIOMED.2012.03.004
Abstract: This paper presents a new index to measure the hypnotic depth of anaesthesia (DoA) using EEG signals. This index is derived from applying combined Wavelet transform, eigenvector and normalisation techniques. The eigenvector method is first applied to build a feature function for six levels of coefficients in a discrete wavelet transform (DWT). The best Daubechies wavelet and their ranking value p are optimally determined to identify different states of anaesthesia. A statistic normalisation process is then carried out to re-scale data and compute the hypnotic depth of anaesthesia. Finally, a new function ZDoA is proposed to compute a DoA index which corresponds one of the five depths of anaesthesia states to very deep anaesthesia, deep anaesthesia, moderate anaesthesia, light anaesthesia and awake. Simulation results based on real anaesthetised EEGs demonstrate that the new index generally parallels the BIS index. In particular, the ZDoA index is often faster than the BIS index to react to the transition period between consciousness and unconsciousness for this data set. A Bland-Altman plot indicates a 95.23% agreement between the ZDoA and BIS indices. The ZDoA trend is responsive, and its movement is consistent with the clinically observed and recorded changes of the patients.
Publisher: Springer Berlin Heidelberg
Date: 2007
Publisher: Institution of Engineering and Technology (IET)
Date: 12-2014
Publisher: IEEE
Date: 2012
Publisher: Springer International Publishing
Date: 2018
Publisher: Elsevier BV
Date: 04-2023
Publisher: Springer Berlin Heidelberg
Date: 2009
Publisher: Elsevier BV
Date: 03-2014
DOI: 10.1016/J.CMPB.2013.12.020
Abstract: Motor imagery (MI) tasks classification provides an important basis for designing brain-computer interface (BCI) systems. If the MI tasks are reliably distinguished through identifying typical patterns in electroencephalography (EEG) data, a motor disabled people could communicate with a device by composing sequences of these mental states. In our earlier study, we developed a cross-correlation based logistic regression (CC-LR) algorithm for the classification of MI tasks for BCI applications, but its performance was not satisfactory. This study develops a modified version of the CC-LR algorithm exploring a suitable feature set that can improve the performance. The modified CC-LR algorithm uses the C3 electrode channel (in the international 10-20 system) as a reference channel for the cross-correlation (CC) technique and applies three erse feature sets separately, as the input to the logistic regression (LR) classifier. The present algorithm investigates which feature set is the best to characterize the distribution of MI tasks based EEG data. This study also provides an insight into how to select a reference channel for the CC technique with EEG signals considering the anatomical structure of the human brain. The proposed algorithm is compared with eight of the most recently reported well-known methods including the BCI III Winner algorithm. The findings of this study indicate that the modified CC-LR algorithm has potential to improve the identification performance of MI tasks in BCI systems. The results demonstrate that the proposed technique provides a classification improvement over the existing methods tested.
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 11-2016
Publisher: IEEE
Date: 07-2010
Publisher: IEEE
Date: 07-2010
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 10-2010
Publisher: Springer Science and Business Media LLC
Date: 03-2010
DOI: 10.1007/S13246-010-0009-5
Abstract: In this study, we propose a stochastic method to analyze the effects of inhomogeneous anisotropic tissue conductivity on electroencephalogram (EEG) in forward computation. We apply this method to an inhomogeneous and anisotropic spherical human head model. We apply stochastic finite element method based on Legendre polynomials, Karhunen-Loeve expansion and stochastic Galerkin methods. We apply Volume and Wang's constraints to restrict the anisotropic conductivities for both the white matter (WM) and the skull tissue compartments. The EEGs resulting from deterministic and stochastic FEMs are compared using statistical measurement techniques. Based on these comparisons, we find that EEGs generated by incorporating WM and skull inhomogeneous anisotropic tissue properties in idually result in an average of 56.5 and 57.5% relative errors, respectively. Incorporating these tissue properties for both layers together generate 43.5% average relative error. Inhomogeneous scalp tissue causes 27% average relative error and a full inhomogeneous anisotropic model brings in an average of 45.5% relative error. The study results demonstrate that the effects of inhomogeneous anisotropic tissue conductivity are significant on EEG.
Publisher: Springer International Publishing
Date: 2019
Publisher: Springer International Publishing
Date: 13-10-2015
DOI: 10.1007/978-3-319-10984-8_8
Abstract: Most epileptic EEG classification algorithms are supervised and require large training datasets, that hinder their use in real time applications. This chapter proposes an unsupervised Multi-Scale K-means (MSK-means) MSK-means algorithm to distinguish epileptic EEG signals and identify epileptic zones. The random initialization of the K-means algorithm can lead to wrong clusters. Based on the characteristics of EEGs, the MSK-means MSK-means algorithm initializes the coarse-scale centroid of a cluster with a suitable scale factor. In this chapter, the MSK-means algorithm is proved theoretically superior to the K-means algorithm on efficiency. In addition, three classifiers: the K-means, MSK-means MSK-means and support vector machine (SVM), are used to identify seizure and localize epileptogenic zone using delay permutation entropy features. The experimental results demonstrate that identifying seizure with the MSK-means algorithm and delay permutation entropy achieves 4. 7 % higher accuracy than that of K-means, and 0. 7 % higher accuracy than that of the SVM.
Publisher: Institution of Engineering and Technology (IET)
Date: 08-2015
Publisher: Springer Science and Business Media LLC
Date: 03-2001
DOI: 10.1007/BF03178283
Publisher: Springer Science and Business Media LLC
Date: 29-03-2014
DOI: 10.1007/S13246-014-0263-Z
Abstract: This paper applies the nonlocal mean (NLM) method to denoise the simulated and real electroencephalograph signals. As a patch-based method, the NLM method calculates the weighted sum of a patch. The weight of each point is determined by the similarity between the points of the own patch and its neighbor. Based on the weighted sum, the noise is filtered out. In this study, the NLM denoising method is applied to signals with additive Gaussian white noise, spiking noise and specific frequency noise and the results are compared with that of the popular sym8 and db16 Wavelet threshold denoising (WTD) methods. The outcomes show that the NLM on average achieves 2.70 dB increase in improved signal to noise ratio (SNRimp) and 0.37 % drop in improved percentage distortion ratio compared with WTD. The moving adaptive shape patches-NLM performs better than the original NLM when the signals change dramatically. In addition, the performance of combined NLMWTD denoising method is also better than original WTD method (0.50-4.89 dB higher in SNRimp), especially, when the signal quality is poor.
Publisher: IEEE
Date: 2003
Publisher: Springer Science and Business Media LLC
Date: 03-2003
DOI: 10.1007/BF03178689
Publisher: Institution of Engineering and Technology (IET)
Date: 2011
Publisher: Inderscience Publishers
Date: 2013
Publisher: IEEE
Date: 2008
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 06-2013
Publisher: IEEE
Date: 03-2009
Publisher: IEEE
Date: 09-2019
Publisher: Springer Science and Business Media LLC
Date: 08-1999
Publisher: Elsevier BV
Date: 12-2011
DOI: 10.1016/J.CMPB.2010.11.014
Abstract: This paper presents a new approach called clustering technique-based least square support vector machine (CT-LS-SVM) for the classification of EEG signals. Decision making is performed in two stages. In the first stage, clustering technique (CT) has been used to extract representative features of EEG data. In the second stage, least square support vector machine (LS-SVM) is applied to the extracted features to classify two-class EEG signals. To demonstrate the effectiveness of the proposed method, several experiments have been conducted on three publicly available benchmark databases, one for epileptic EEG data, one for mental imagery tasks EEG data and another one for motor imagery EEG data. Our proposed approach achieves an average sensitivity, specificity and classification accuracy of 94.92%, 93.44% and 94.18%, respectively, for the epileptic EEG data 83.98%, 84.37% and 84.17% respectively, for the motor imagery EEG data and 64.61%, 58.77% and 61.69%, respectively, for the mental imagery tasks EEG data. The performance of the CT-LS-SVM algorithm is compared in terms of classification accuracy and execution (running) time with our previous study where simple random s ling with a least square support vector machine (SRS-LS-SVM) was employed for EEG signal classification. We also compare the proposed method with other existing methods in the literature for the three databases. The experimental results show that the proposed algorithm can produce a better classification rate than the previous reported methods and takes much less execution time compared to the SRS-LS-SVM technique. The research findings in this paper indicate that the proposed approach is very efficient for classification of two-class EEG signals.
Publisher: Springer Science and Business Media LLC
Date: 09-2006
DOI: 10.1007/BF03178571
Start Date: 01-2006
End Date: 11-2010
Amount: $135,000.00
Funder: Australian Research Council
View Funded Activity