Accession Number : AD1032001

Title :   Analysis of Small Muscle Movement Effects on EEG Signals

Descriptive Note : Technical Report,01 Aug 2015,01 Dec 2016


Personal Author(s) : Yanteri,Erhan E

Full Text :

Report Date : 22 Dec 2016

Pagination or Media Count : 90

Abstract : Developments in the biomedical signal processing have led the electroencephalography (EEG) to be a critical tool for the Brain Computer Interface (BCI) systems and Human Machine Teams (HMTs). Both of them strongly rely on the EEG signals in order to evaluate the neural activity and the cognitive state. But the physiological and non-physiological artifacts distort the EEG signals and make the interpretation of cognitive state harder or may cause misinterpretations. While interacting with computers, humans perform small motor muscle movements such as operating a keyboard and mouse. On the other side, the computer agent needs to know the cognitive state of the human teammate in order to make decisions and the EEG signals are the only information source of cognitive state. In this thesis, the artefactual effects of the small muscle movements were investigated. Upper frequency bands (30 Hz) of the EEG signal were extracted in order to investigate the artefactual effects of the small muscle movements. When the contamination level is high, the detection of the small muscle artifact can be made with the 92.2% accuracy. If these artifacts are really small such as a single finger movement, the detection accuracy decreases to 64%. But, the detection accuracy increases to 72% after removing the eye blink artifacts. The results of the classification support our hypothesis about the artefactual effects of the small muscle movements

Descriptors :   machine learning , feature extraction , detection , electroencephalography , brain waves , artificial neural networks , Muscles

Subject Categories : Medicine and Medical Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE