EEG Signal Analysis System for Finger Movement Detection

• Main Authors: Yung-Chun Liu, 劉勇均
• Other Authors: Yung-Nien Sun
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/23412415412673544100

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 92 === 　　Many neurological diseases, such as stroke and spinal cord injury, disrupt the connections between brain cortex and muscles. Besides, some other diseases may destruct the muscle and make it functionless. All these diseases interfere with the voluntary movements of the subjects and influence their ability to accomplish the attempted task. Brain-computer interface (BCI), which defines an artificial alternative output from the brain cortex to make communication with their surrounding targets, can improve above deficits. 　　The most common way of BCI is to give control signals based on the analysis of Electroencephalogram (EEG) signals. And the recognition of finger movements has been one of the most important issues in this field. In the previous researches, the length of EEG trials for analysis were usually between 4 to 10 seconds, therefore it would have difficulties in real-time applications. For this reason, we study the technique of analyzing the EEG signals which have the length of one second, and construct a real-time EEG recognition system based on it for detecting finger movements. We adopt the strategy, named Active Time Segment Selection, to pick the most appropriate time segment of the EEG trial for the recognition of finger movements. And the classifier is trained with the information of this segment in all trials. The integrated processes with the above-mentioned functions form a two-staged recognition system to classify the finger motions in real-time. Besides, we propose an automatic approach to provide statistical analysis on the results of recognition in each stage. 　　From the results of the experiment, it has shown that our system can distinguish a finger movement or a non-movement from the input EEG signal sequences, and further recognize the movement as a left or a right one successfully. We expect to use the system in controlling clinical assistive devices in the future, and benefit the subjects with neurological diseases or limb disabilities.