چكيده به لاتين
Abstract:
Nowadays, EEG signals have been considered under a great attention because of their wide applications in vital roles such as, bio-medical science, medical industry and especially in brain computer interface (BCI) systems.
EEG signals are appearing in 5 frequency bands which are called Delta, Theta, Alpha, Beta and Gamma. The synchronization and relation between these bands are called frequency coupling and this coupling can be occurred between amplitude, phase and frequency.
In this thesis, it is used from two EEG recorded signals database in order to assess the functional of frequency-frequency coupling. The first EEG recorded signals is an One-Bakc task which was recorded by showing 3D images of 10 different categories involves clothes and towel, curtains, electrical devices, lamps, pots and glasses, sport equipments, bath countertops, tables and desks, trees and vehicles which are shown in both front and side view. The another EEG database is a selective task which users must reaction after showing 12 different categories of images involves animals, flowers, foods, fruits, buildings, stationeries, dolls, jewelleries, clothes, transportations, humans body organs and electrical devices which are shown just in front view. The frequency-frequency coupling and features were extracted in three ways:
1- Within each electrodes
2- Between five brain areas (Frontal, Cerebral, Temporal, Occipital and Parietal).
The results which are calculated by using Support Vector Machine (SVM) and K-Nearest neighbor (KNN) and scalar feature selction methods such as, Receiver Operating Characteristics (ROC) and variance test method (Anova) are showing there would be a meaningful information connectivity between brain areas and frequency bands.
Keywords: Cross-Frequency Coupling, Electroencephalogram Signal, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Frequency-Frequency Coupling (FFC), Scalar Feature Selection
