| Summary: | 碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 101 === Schizophrenia is a chronic brain disorder. 1% population worldwide have this illness approximately and it typically occurred in young adulthood (15~45 year). Generally, schizophrenia patients usually could not identify reality and hallucination and it would make social and occupational dysfunctions. Until now, diagnosis of schizophrenia is based on observed behavior and the patient's reported experiences (Diagnostic and Statistical Manual of Mental Disorders, DSM-V).
The aim of this study is to evaluate the feasibility of classifying schizophrenia and healthy people by analyzing their objective physiological information, and construct an efficient classification procedure. In this thesis, the auditory even-related potentials (AEP) experiment were recorded from 12 schizophrenia patients and 12 healthy people. The stimuli was presented randomly by chord stimuli and interval stimuli, by observing the statistical difference of ERPs results in N1 and P2 components of each musical stimuli type in each group and selected the discrepant electrode sites for different combination, the amplitude of N1 and P2 of the specific electrode sites were taken as features for classification, and some specific electrode sites were found to be important through the classification result. The related locations of those electrode sites are roughly at frontal and temporal lobes in brain, which may be connected with brain functions impairment in Schizophrenia. Finally, the extracted features were fed into the LDA for classification. Accuracy of the proposed method can reach 91.67% through leave-one-out cross validation when combining N1 amplitude at statistically discrepant electrode sites (i.e. F3, F4, T3; F3, F7, T4; F3, F8, T4 three combination modes) as feature. The result showed that N1 amplitude of specific electrode sites supply useful features in Schizophrenia classification. In this thesis, statistical techniques combined with machine learning were used to select discriminating electrode sites and features for classification, and to develop an efficient and effective classification system. Furthermore, information observed through the classification result can be provided for the clinical diagnosis and deeper analysis in the future.
|
|---|
