FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4205543438> ?p ?o ?g. }

• W4205543438 abstract "Anesthesia delivery yields dynamic electroen-cephalogram (EEG) signatures during moments of consciousness and unconsciousness. Consequently, real-time analysis of EEG signals during anesthesia administration is a promising technique to monitor patient states. Moreover, EEG has the potential to be the prime brain-monitoring technique for the creation of a closed-loop anesthesia delivery (CLAD) system. Machine learning (ML) software that can classify EEG time samples as conscious or unconscious is necessary for further CLAD system development. Burst suppression, a common metabolic phenomenon in deeply unconscious patients, presents a unique challenge to engineering an effective classifier. In this project, a previously engineered EEG classifier was replicated using band wise power (BWP) and principal component analysis (PCA) as features in a logistic regression classifier. The classifier performed well on the training data (BWP AUC <tex>${=0.94428}$</tex>, PCA AUC <tex>${=0.96096}$</tex>) and validation sets (BWP Mean AUC <tex>${=0.94426}$</tex>, SEM <tex>${=0.0010468}$</tex>, PCA Mean AUC <tex>${=0.96099}$</tex>, SEM <tex>${=0.0010535}$</tex>) but performed poorly when burst suppression samples were isolated for performance (BWP Hit rate <tex>${=0.4062}$</tex>, PCA Hit rate <tex>${=0.4330}$</tex>). Two new supervised ML approaches, support vector machines (SVM) and single-layer artificial neural networks (ANN), were utilized to improve classifier performance on burst suppression samples. As a result, the best performing classifier, PCA trained SVM, generated a burst suppression hit rate of 0.9598 on the training set and 0.8660 on the testing set while maintaining an AUC of 0.94714. While this project demonstrates that a classifier utilizing PCA with SVM is a promising avenue for robustly characterizing EEG signals in the presence of burst suppression, we conclude that dynamic preprocessing normalization techniques need to be implemented for optimal performance without overfitting." @default.
• W4205543438 created "2022-01-25" @default.
• W4205543438 creator A5030321614 @default.
• W4205543438 creator A5054462593 @default.
• W4205543438 creator A5073437420 @default.
• W4205543438 date "2020-10-09" @default.
• W4205543438 modified "2023-10-18" @default.
• W4205543438 title "Classifying EEG of Propofol-Induced Unconsciousness in the Presence of Burst Suppression" @default.
• W4205543438 doi "https://doi.org/10.1109/urtc51696.2020.9668878" @default.
• W4205543438 hasPublicationYear "2020" @default.
• W4205543438 type Work @default.
• W4205543438 citedByCount "0" @default.
• W4205543438 crossrefType "proceedings-article" @default.
• W4205543438 hasAuthorship W4205543438A5030321614 @default.
• W4205543438 hasAuthorship W4205543438A5054462593 @default.
• W4205543438 hasAuthorship W4205543438A5073437420 @default.
• W4205543438 hasConcept C118552586 @default.
• W4205543438 hasConcept C12267149 @default.
• W4205543438 hasConcept C153180895 @default.
• W4205543438 hasConcept C154945302 @default.
• W4205543438 hasConcept C27438332 @default.
• W4205543438 hasConcept C2776470351 @default.
• W4205543438 hasConcept C41008148 @default.
• W4205543438 hasConcept C42219234 @default.
• W4205543438 hasConcept C50644808 @default.
• W4205543438 hasConcept C522805319 @default.
• W4205543438 hasConcept C71924100 @default.
• W4205543438 hasConcept C95623464 @default.
• W4205543438 hasConceptScore W4205543438C118552586 @default.
• W4205543438 hasConceptScore W4205543438C12267149 @default.
• W4205543438 hasConceptScore W4205543438C153180895 @default.
• W4205543438 hasConceptScore W4205543438C154945302 @default.
• W4205543438 hasConceptScore W4205543438C27438332 @default.
• W4205543438 hasConceptScore W4205543438C2776470351 @default.
• W4205543438 hasConceptScore W4205543438C41008148 @default.
• W4205543438 hasConceptScore W4205543438C42219234 @default.
• W4205543438 hasConceptScore W4205543438C50644808 @default.
• W4205543438 hasConceptScore W4205543438C522805319 @default.
• W4205543438 hasConceptScore W4205543438C71924100 @default.
• W4205543438 hasConceptScore W4205543438C95623464 @default.
• W4205543438 hasFunder F4320306076 @default.
• W4205543438 hasLocation W42055434381 @default.
• W4205543438 hasOpenAccess W4205543438 @default.
• W4205543438 hasPrimaryLocation W42055434381 @default.
• W4205543438 hasRelatedWork W1579270119 @default.
• W4205543438 hasRelatedWork W2041636156 @default.
• W4205543438 hasRelatedWork W2160451891 @default.
• W4205543438 hasRelatedWork W2358824780 @default.
• W4205543438 hasRelatedWork W2380927352 @default.
• W4205543438 hasRelatedWork W2538979170 @default.
• W4205543438 hasRelatedWork W3004377704 @default.
• W4205543438 hasRelatedWork W3160838270 @default.
• W4205543438 hasRelatedWork W4225691219 @default.
• W4205543438 hasRelatedWork W4242764575 @default.
• W4205543438 isParatext "false" @default.
• W4205543438 isRetracted "false" @default.
• W4205543438 workType "article" @default.