Home
Resource
Modeling classification network of electroencephalographic artifacts and signals associated with deep brain stimulation
PDF
PDF format is widely accepted and good for printing.
Plug-in required
PDF-1(9.64 MB)
Citation & Export
View Usage Statistics
Staff View

Citation & Export
Hide

Simple citation

Tong, Saichiu Nelson. Modeling classification network of electroencephalographic artifacts and signals associated with deep brain stimulation. Retrieved from https://doi.org/doi:10.7282/T3J67KCR

Export

Click here for information about Citation Management Tools at Rutgers.

Statistics
Hide

Description

TitleModeling classification network of electroencephalographic artifacts and signals associated with deep brain stimulation
NameTong, Saichiu Nelson (author); Pham, Hoang (chair); Albin, Susan L. (internal member); Wang, Honggang (internal member); Wong, Stephen (outside member); Rutgers University; Graduate School - New Brunswick
Date Created2017
Other Date2017-01 (degree)
SubjectIndustrial and Systems Engineering, Diagnostic imaging, Brain stimulation
Extent1 online resource (xvi, 236 p. : ill.)
DescriptionIn the last decade, the study of network has emerged as a useful tool to investigate complex systems with interdependent components because a new set of theoretical tools and practical techniques have been developed to analyze data that can be modeled as networks. The objective of this study is to develop a new reliable framework for classifying network data evolved over time. In particular, this research project conducts a study on EEG data as a form of multivariate time series modeled as networks during a period of time in which a treatment procedure called deep brain stimulation (DBS) is either set to be on or off. DBS is a procedure used to treat symptoms of neurological movement and neuropsychiatric disorders—most commonly for patients with Parkinson’s disease (PD) who do not respond well to medication. The procedure involves the use of a surgically implanted, battery operated neurostimulator to send electrical stimulation to targeted areas in the brain. Currently, conventional DBS therapy is not suitable to be individualized: DBS is only able to deliver stimulation continuously and the DBS setting, such as amount of stimulation delivered, is adjusted based on a trial-and-error process involving subjective clinical assessment. If DBS is able to deliver an adequate amount of stimulation only when it is necessary for an individual patient based on a quantitative metric, adjusting DBS setting will become a more objective process and the therapy will become more effective with less side effect. For such an adaptive DBS to be realized, a detectable feedback signal must be identified that can be used to adjust DBS setting in a feedback control system. Currently, there is lack of such definitive signals that are thought to be reliable as the basis for feedback. Electroencephalographic (EEG) recording is a clinical and electrophysiological monitoring tool used by physicians to diagnose and monitor patients with neurological conditions. EEG measures regional electrical activity of the brain of an individual in the form of a multivariate time series dataset. In this study, a classification framework is developed to identify whether or not a set of EEG signals and its frequency components can be used to develop classifier for identifying EEG signals belonging to a state when the DBS is on vs. a state when the DBS is off. In particular, the framework applies binary classifiers to features derived from networks constructed based on the EEG data. The first part of this study attempts to apply the classification framework by focusing on physiological frequency ranges with three different binary classifiers including logistic regression, least absolute shrinkage and selection operator (LASSO), and a new method known as principal component stepwise selection logistic regression (PCSSLR) that is developed in this thesis. The second part attempts to identify potential biomarkers to be used as feedback control signals for an adaptive DBS device. The results of both parts will help to address the question of whether or not EEG data can be used to detect feedback signals for use in adaptive DBS feedback control system and to establish a modeling framework to develop similar classifier for performing statistical classification based on network features to identify which group a temporal signal belongs to.
NoteM.S.
NoteIncludes bibliographical references
Noteby Saichiu Nelson Tong
Genretheses, ETD graduate
Persistent URLhttps://doi.org/doi:10.7282/T3J67KCR
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Version 8.5.5
Rutgers University Libraries - Copyright ©2024