Primjena metode nezavisnih komponenata u analizi elektroencefalograma

undergraduate thesis

• Title (english): Independent Component Analysis of EEG data
• Author: Juraj Patalen
• Mentor: Igor Lacković (mentor)
• Committee member: Igor Lacković (predsjednik povjerenstva)
• Committee member: Sara Žulj (član povjerenstva)
• Committee member: Krešimir Friganović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-08-28, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Metoda nezavisnih komponenata (eng. Independent Componenet Analysis, skraćeno ICA) je matematička metoda razdvajanja izvora na slijepo, koja potječe iz područja obrade signala. U ovom radu prikazana je i opisana uporaba metode u elektrofiziologiji, odnosno elektroencefalografiji. Jedan od najvećih problema pri analizi elektroencefalograma je određivanje točne lokacije izvora EEG-a. Primjenom metode nezavisnih komponenata rješava se prva polovica inverznog problema lociranja izvora EEG-a. U radu je opisan i izveden matematički model ICA koji se koristi u mnogim ICA algoritmima. Nadalje, Infomax ICA algoritmom, u programu EEGLAB i MATLAB okružju, izvršena je dekompozicija snimljenih elektroencefalograma 13 osoba jednog javno dostupnog skupa podataka. Prikazane su i objašnjenje vrste nezavisnih komponenata koje se pojavljuju ICA dekompozicijom EEG podataka. Na jednom od 13 setova podataka prikazano je uklanjanje očnih artefakata uporabom ICA. Na kraju rada je korištenjem cijelog skupa podataka prikazana mogućnost grupiranja sličnih nezavisnih komponenti na temelju njihovih projekcija na skalp i sličnosti u ekvivalentnim strujnim dipolima.

Abstract (english)

Independent Component Analysis (ICA) is a mathematical method for blind source separation used in signal and data processing. In this paper, the use of ICA in electrophysiology, specifically electroencephalography is discussed with examples showing the results of ICA decomposition of EEG data. One of the main issues in analyzing EEG data is determining the locations of sources of EEG data in the brain. ICA is an effective tool for solving the first part of the inverse problem of locating EEG sources. A mathematical model of ICA, which is used in many ICA algorithms, is derived in the paper. An ICA decomposition with Infomax Algorithm is performed in EEGLAB and MATLAB on a publicly available EEG dataset collection consisting of 13 subjects' individual EEG recordings. Different types of independent components that ICA decomposition produces are shown. After that, removal of eye artifacts with ICA from one of the datasets is demonstrated. In the last part, the entire dataset collection is used in an example showing how clustering similar independent components with respect to their scalp maps and equivalent dipole locations can be performed.

• Keywords: Metoda nezavisnih komponenata; elektroencefalografija; EEG; izvori EEG-a; EEGLAB; elektrofiziologija
• Keywords (english): Independent Component Analysis; ICA; Electroencephalography; EEG; EEG Sources; EEGLAB; Electrophysiology
• Language: croatian
• URN:NBN: urn:nbn:hr:168:828709
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 09:38:42