FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2802959148> ?p ?o ?g. }

• W2802959148 abstract "Neonatal Hypoxic Ischaemic Encephalopathy (HIE) is a major cause of morbidity and mortality in newborns throughout the world with an incidence of 1-8/1000 newborns in developed countries and up to 26/1000 newborns in developing countries. It causes one million neonatal deaths globally per year. Electroencephalography (EEG) is a useful method for assessment of brain activity in newborns with HIE. EEG performs well in the early diagnosis and classification of HIE severity and in predicting neurodevelopmental outcome. A normal EEG is highly predictive of a normal outcome and various abnormal EEG features are associated with neurological abnormalities or death.A systematic review and meta-analysis of the literature was undertaken to determine which specific background features of the EEG best predict outcome (Aim 1). Automatic detection and classification would be useful for clinicians and reduce workload and subjectivity. Signal processing approaches particularly in the joint time-frequency domain and machine learning techniques have been used to characterise (Aim 2), and to detect and classify (Aim 3) these identified features. The final goal (Aim 4) was to identify the combination of EEG signal features that best predict neurodevelopmental outcome in a cohort of term neonates with HIE.The first contribution of this project was to identify the specific EEG background patterns, in term neonates with HIE, that best predict neurodevelopmental outcome through a systemic review of the published literature. A meta-analysis was performed to establish the prognostic value of the identified EEG background patterns and the pooled sensitivity and specificity were calculated. A significant problem was identified: the use of different definitions of the different abnormal EEG features with respect to voltage level, phase or frequency. Agreement on definitions is necessary for the effective implementation and use of EEG in NICUs, I have called for the adoption of specific definitions. The second contribution was to the optimal use of time-frequency distributions (TFD) for the characterisation of EEG signals and other non-stationary signals. Two optimisation methods, one a global method using a hybrid genetic algorithm and a second, a local optimisation method using a locally optimised spectrogram, have been proposed to optimise the TFD. These contributions present a data-adaptive kernel and analysis-window that reduces the presence of cross-terms and enhances concentration and resolution. The global optimisation method uses single global parameters whereas the local optimisation method divides the TFD into small grids and performs local optimisation. Both methods are applied automatically, and no user input is necessary. Using both simulated signals and real EEG, both methods improved characterisation. The optimised TFDs are not only suitable for characterisation of non-stationary signals, but also have applications in signal detection and classification.The third contribution was to the design and optimisation of classification of multi-channel EEG background patterns in term neonates with HIE. Two classification methods utilising a single feature subset and a class-specific feature subset have been designed and optimised to classify five EEG background patterns: burst, suppression, normal, seizure and artefact. These patterns are clinically relevant to treatment and prediction of neurodevelopmental outcomes. Various time domain, frequency domain and joint time-frequency domain features are first extracted and then concatenated to produce a long feature set. A hybrid feature selection (HFS) algorithm was proposed to simultaneously select the prominent feature subset as well as the support vector machine (SVM) tuning parameters. Three fusion techniques, channel fusion, feature fusion and decision fusion, have been used in the multi-channel classification. Finally, two outputs of a decision support system (DSS), ‘EEG state’ and ‘EEG quality’, are presented as useful parameters to assist clinical staff in management of babies with HIE.The fourth contribution was the conduct of a study to test the relationship between the identified EEG features in term neonates with HIE and neurodevelopmental outcome at the age of 2 years. A highly discriminative and non-redundant feature subset as well as SVM parameters have been selected from this high dimensional feature set using a HFS algorithm. This algorithm is applied to the statistically most unbiased Leave-One-Subject-Out (LOSO) cross-validation and an optimised model is created using the most consistent feature subset. This model predicts good/poor neurodevelopmental outcome with 83.77% accuracy when applied to a separate dataset and improved the prediction accuracy of other approaches/studies by 5–10%. A DSS has been built as a potential application of the model to visualise the ‘probable long-term neurodevelopmental outcome’ in a continuous probabilistic fashion. These results provide strong support for a future objective decision support tool for the early prediction of neurodevelopmental outcome for babies with HIE. In summary, this project makes significant contributions to optimise use of TFDs and SVM for EEG signal characterisation and the detection and classification of abnormal EEG patterns used in the prediction of neurodevelopmental outcome and suggests future directions for research and translation to clinical practice." @default.
• W2802959148 created "2018-05-17" @default.
• W2802959148 creator A5020888402 @default.
• W2802959148 date "2018-03-27" @default.
• W2802959148 modified "2023-09-27" @default.
• W2802959148 title "Design and optimisation of time-frequency analysis for multichannel neonatal EEG background features in term neonates with hypoxic ischaemic encephalopathy: characterisation, classification and neurodevelopmental outcome prediction" @default.
• W2802959148 doi "https://doi.org/10.14264/uql.2018.225" @default.
• W2802959148 hasPublicationYear "2018" @default.
• W2802959148 type Work @default.
• W2802959148 sameAs 2802959148 @default.
• W2802959148 citedByCount "0" @default.
• W2802959148 crossrefType "dissertation" @default.
• W2802959148 hasAuthorship W2802959148A5020888402 @default.
• W2802959148 hasBestOaLocation W28029591482 @default.
• W2802959148 hasConcept C118552586 @default.
• W2802959148 hasConcept C126322002 @default.
• W2802959148 hasConcept C187212893 @default.
• W2802959148 hasConcept C2778080812 @default.
• W2802959148 hasConcept C2778621155 @default.
• W2802959148 hasConcept C522805319 @default.
• W2802959148 hasConcept C71924100 @default.
• W2802959148 hasConcept C72563966 @default.
• W2802959148 hasConceptScore W2802959148C118552586 @default.
• W2802959148 hasConceptScore W2802959148C126322002 @default.
• W2802959148 hasConceptScore W2802959148C187212893 @default.
• W2802959148 hasConceptScore W2802959148C2778080812 @default.
• W2802959148 hasConceptScore W2802959148C2778621155 @default.
• W2802959148 hasConceptScore W2802959148C522805319 @default.
• W2802959148 hasConceptScore W2802959148C71924100 @default.
• W2802959148 hasConceptScore W2802959148C72563966 @default.
• W2802959148 hasLocation W28029591481 @default.
• W2802959148 hasLocation W28029591482 @default.
• W2802959148 hasOpenAccess W2802959148 @default.
• W2802959148 hasPrimaryLocation W28029591481 @default.
• W2802959148 hasRelatedWork W1487068232 @default.
• W2802959148 hasRelatedWork W1997359198 @default.
• W2802959148 hasRelatedWork W2034384415 @default.
• W2802959148 hasRelatedWork W2320466601 @default.
• W2802959148 hasRelatedWork W2475148710 @default.
• W2802959148 hasRelatedWork W2479263883 @default.
• W2802959148 hasRelatedWork W2551054676 @default.
• W2802959148 hasRelatedWork W2590940923 @default.
• W2802959148 hasRelatedWork W2732013820 @default.
• W2802959148 hasRelatedWork W2734510146 @default.
• W2802959148 hasRelatedWork W2908003897 @default.
• W2802959148 hasRelatedWork W2926087227 @default.
• W2802959148 hasRelatedWork W2982627395 @default.
• W2802959148 hasRelatedWork W2985333787 @default.
• W2802959148 hasRelatedWork W3025788106 @default.
• W2802959148 hasRelatedWork W3098821978 @default.
• W2802959148 hasRelatedWork W3106067791 @default.
• W2802959148 hasRelatedWork W3159456745 @default.
• W2802959148 hasRelatedWork W3172536974 @default.
• W2802959148 hasRelatedWork W3188686187 @default.
• W2802959148 isParatext "false" @default.
• W2802959148 isRetracted "false" @default.
• W2802959148 magId "2802959148" @default.
• W2802959148 workType "dissertation" @default.