Please use this identifier to cite or link to this item:
|Title: ||基於獨立EEG成分的運動執行、運動想像和運動觀察的源識別及比較乾/濕電極的獨立成分差異;Source identification for the motor execution, motor imagery and motor observation conditions based on independent EEG components, and comparison of independent components of the EEG data obtained using dry and wet electrodes|
|Authors: ||黃育瑜;Huang, Yu-Yu|
|Keywords: ||腦機介面;動作想像;Mu節律;乾電極腦波圖;brain computer interface;motor imagery;mu rhythm;dry electrode EEG|
|Issue Date: ||2019-09-03 15:52:04 (UTC+8)|
;Motor imagery (MI) is one of the brain wave features used in the brain-computer interface(BCI). It is often discussed together with the motor execution (ME) and motor observation (MO), because these motor conditions all induce the event-related desychronization(ERD) of the mu rhythm in the motor-related cortex. The relationship between the source locations of where the mu-rhythm of the three motor conditions have been originated has been largely discussed in the literature. Therefore, the first purpose of this study was to identify and compare the sources of the three motor tasks. In addition, EEG acquisition equipment has developed toward more convenient dry electrodes in recent years. To facilitate the application and development of the BCI, the second purpose of this study was to compare the ability of the wet and dry electrodes to collect signals using the mu component characteristics.
First experiment collected 13 participants with wet-electrode EEG. The within subjects experimental design included motor execution, motor imagery and motor observation. Source identification of different motor tasks was performed using two different sequences of independent component analysis. First, the ICA was applied to the preprocessed EEG data with the data from the three different motor condition altogether for the single ICA decomposition(thus, SD); second, the ICA was separately applied to the preprocessed EEG data of each motor condition to obtain ME, MI and MO with ICA being applied to the data of each motor condition separately. Then, the motor-related independent components, maximally projecting to surrounding the C4 channel location with the maximal mu-suppression feature, of each individual subject would be selected and subjected to source localization process using the DIPFIT2 extension under the EEGLAB. If the sources of the three motor conditions were co-located, there should not be any difference in the sources identified with different ICA steps. The results show that the SD ICA is more lateral than the MO ICA, and the MI ICA is more anterior than the ME ICA. Therefore, the mu ERD activity seemed to be originated from different brain regions for the motor execution, motor imagery, and motor observation conditions. In addition, to compare the mu rhythm characteristics obtained using the dry and wet electrodes, the dry electrode EEG of 16 participants were further collected with the performance of the three motor conditions combined. To compare the result of the dry EEG system to that of the wet EEG system, the number of wet EEG channels was reduced from the original 64 channels to 30 channels to match the setting of the dry EEG system. The results indicated that the number of channels had less influence on signal quality. However, the higher impedance inherently associated with the dry EEG electrodes might play more an important role in affecting the signal quality of the acquired EEG signals. Nonetheless, the dry EEG electrodes could still pick up some of the mu-suppression characteristics associated with the MI and other motor conditions.
|Appears in Collections:||[認知與神經科學研究所 ] 博碩士論文|
Files in This Item:
All items in NCUIR are protected by copyright, with all rights reserved.