A new feature extraction method for quantitative EEG based on relative source temporal (RST) approach /
Electroencephalogram (EEG) is one of neuroimaging tools that measures electrical activity of the brain. EEG devices have played an important role in brain studies and understanding brain functionality. For more than two decades quantitative EEG (QEEG) feature extraction methods, utilized different m...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Information and Communication Technology, International Islamic University Malaysia,
2015
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Electroencephalogram (EEG) is one of neuroimaging tools that measures electrical activity of the brain. EEG devices have played an important role in brain studies and understanding brain functionality. For more than two decades quantitative EEG (QEEG) feature extraction methods, utilized different mathematical and statistical techniques to extract relevant information of the EEG signals have been used in analysing and understanding the brain functionality prior to classification process. Some of the major contributions of EEG signals classification is in brain computer interface (BCI) and also neurophysiology and psychology diagnosis. Even though EEG signal classification algorithms represent very useful techniques especially for BCI, there is a high variation in methods' efficiencies among subjects and for the same subject. Further classification algorithms did not show good performance to identify high cognitive task such as emotional states. As it has known in classification system, extracted relevant features help to increase the classification performance. Therefore it is important to identify EEG signal by appropriate feature extraction method. Most of EEG feature extraction methods concentrate on frequency and time domain values. However, very little information is available for techniques relying on using EEG data for spatial feature information from the brain activity. In addition, understanding the underlying process of brain function is important to identify brain activity. Therefore, there is a significant need to extract the spatial information from the EEG signal measurement. Recent studies have shown the results of combining functional magnetic resonance imaging (fMRI) with EEG signal to extract the spatial domain information however it is expensive and cannot be affordable by a large number of people. Also, its applications are limited due to the size and structure of the system. This study aims to introduce a new technique to quantify EEG signal to extract the dynamic spatial information in terms of relative source temporal feature related to the brain activity and utilize it in classification process to identify different brain functions. The proposed model is based on applying one of the well-known wireless techniques called time difference of arrival (TDOA) with the EEG signal. The new feature extraction model has been evaluated in classification process to detect different categories of EEG signals that are 'rest conditions' and 'emotion responses' of visual stimuli of normal children of ages (4-6) years. It has also been used to detect the abnormal EEG activity of a group of children of ages (6-9) years with Autism Spectrum Disorder (ASD). Further comparative study has been done using other relevant EEG feature extraction methods. Results indicate the possibility to identify brain function by dynamic source distribution and utilize it as features in classification process. Hence results show significant performance of this technique, as compared to others, with less variance among subjects. In addition, it also shows high accuracy in detecting the ASD instances. However, it shows slightly inferior performance in case of blind tests. The knowledge gained by this study can help in designing better adaptable robust systems for BCI and Affective Computing areas. Indeed, introducing a novel approach for extracting the temporal spatial domain of EEG signal can be utilized in a variety of application of EEG signal classification system. |
|---|---|
| Physical Description: | xvi,148 leaves : ill. ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves |