Pregled bibliografske jedinice broj: 1185970
Extraction of Movement Related Oscillatory Changes from Electroencephalogram with Application to Brain- Computer Interface
Extraction of Movement Related Oscillatory Changes from Electroencephalogram with Application to Brain- Computer Interface, 2015., doktorska disertacija, Fakultet elektrotehnike i računarstva, Zagreb
CROSBI ID: 1185970 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Extraction of Movement Related Oscillatory Changes
from Electroencephalogram with Application to Brain-
Computer Interface
Autori
Ana Branka Jerbić
Vrsta, podvrsta i kategorija rada
Ocjenski radovi, doktorska disertacija
Fakultet
Fakultet elektrotehnike i računarstva
Mjesto
Zagreb
Datum
29.04
Godina
2015
Stranica
157
Mentor
Mario Cifrek
Ključne riječi
brain computer interface (BCI) ; motor imagery (MI) ; sensorimotor rhythms (SMRs) ; time frequency (TF) analysis ; event related desynchronization (ERD) ; event related synchronization (ERS) ; Hilbert Huang transform (HHT) ; band power (BP) features
Sažetak
A brain computer interface (BCI) is a technology that decodes a user's intent from brain activity. Obtained information can be used to control devices in the user's environment. In this thesis, changes in sensorimotor rhythms (SMR) during motor imagery (MI) are extracted from electroencephalogram (EEG). These changes are reflected in amplitude attenuation (event related desynchronization) or enhancement (event related synchronization) of SMRs in certain brain regions. Extracted features are classified to left hand, right hand and feet MI, which is possible since MI of different parts of the body results in spatially distinct activation of motor cortex. The brain is a complex organ and interindividual differences in EEG are considerable. In this thesis, parameters influencing the classification accuracy of MI based BCI are identified and analyzed. It is demonstrated that for maximal classification accuracy a MI based BCI has to be custom designed. SMR changes can be detected using time frequency (TF) methods of analysis. A TF method called the Hilbert Huang transform (HHT) was utilized to facilitate the extraction of SMRs, while a well established band power (BP) method served as a reference method for comparison purposes. The HHT was successful in the extraction of SMR changes, but did not outperform the simple BP method. However, the HHT revealed the existence of a phenomenon that, to the author's knowledge, has not been reported previously in the literature. It concerns the increase of frequency in the MI period on the contralateral side of the brain in those intrinsic mode functions, obtained using the HHT, which contain SMRs.
Izvorni jezik
Engleski
Znanstvena područja
Elektrotehnika
POVEZANOST RADA
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb