Year: 2007
Type: dissertation
Abstract: The brain requires a continuous supply of oxygen and even a short period of\nreduced oxygen supply risks severe and lifelong consequences for the\naffected individual. The delivery is a vulnerable period for a baby who may\nexperience for example hypoxia (lack of oxygen) that can damage the brain.\nBabies who experience problems are placed in an intensive care unit where\ntheir vital signs are monitored, but there is no reliable way to monitor the\nbrain directly. Monitoring the brain would provide valuable information\nabout the processes going on in it and could influence the treatment and help\nto improve the quality of neonatal care. The scope of this project is to\ndevelop methods that eventually can be put together to form a monitoring\nsystem for the brain that can function as decision-support for the physician in\ncharge of treating the patient.\nThe specific technical problem that is the topic of this thesis is detection of\nburst and suppression in the electroencephalogram (EEG) signal. The thesis\nstarts with a brief description of the brain, with a focus on where the EEG\noriginates, what types of activity can be found in this signal and what they\nmean. The data that have been available for the project are described,\nfollowed by the signal processing methods that have been used for preprocessing,\nand the feature functions that can be used for extracting certain\ntypes of characteristics from the data are defined. The next section describes\nclassification methodology and how it can be used for making decisions\nbased on combinations of several features extracted from a signal. The\nclassification methods Fisher’s Linear Discriminant, Neural Networks and\nSupport Vector Machines are described and are finally compared with respect\nto their ability to discriminate between burst and suppression. An experiment\nwith different combinations of features in the classification has also been\ncarried out. The results show similar results for the three methods but it can\nbe seen that the SVM is the best method with respect to handling multiple\nfeatures.
Author: Johan Löfhede
Institution: University of Borås
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2106896081