Repository logo
 

Feasibility Analysis of Using NeuCube 3D SNN Environment for Spatio-Temporal EEG Data Classification Related to Perception of Art

Files

Whole thesis(6.18 MB)

Date

2014

Authors

Supervisor

Kasabov, Nikola
Schliebs, Stefan

Item type

Thesis

Degree name

Master of Philosophy

Journal Title

Journal ISSN

Volume Title

Publisher

Auckland University of Technology

Abstract

This thesis is a feasibility study of using a Spiking Neural Network (SNN) architecture named NeuCube for the classification of electroencephalography (EEG) data related to the perception of art. We have performed classification of human brain perception EEG data obtained via a set of experiments on originally created audio, video, and audio & video mixed stimuli. The analysis of results confirms that the proposed method is feasible for further analysis and experimentation, and for the study of art perception and creativity. Massive amounts of complex Spatio-Temporal Brain Data (STBD) have been accumulated recently. As it is critical in many disciplines to rely on proper analysis, understanding and utilization of complex spatio-temporal brain data, such as that from an EEG, this is a great challenge which this study seeks to contribute to. In this study for classification purposes a new evolving Spiking Neural Network (SNN) architecture will be used, called NeuCube. NeuCube is the latest neuroscience software tool developed at KEDRI, AUT, for spatio- and spectro -temporal pattern recognition of brain data, for the creation of concrete models to map, learn and understand STBD. A NeuCube model is based on a 3D evolving SNN that is an approximate map of structural and functional areas of interest of the brain related to the modeling STBD. An evaluation of feasibility of NeuCube for classification of Spatio-Temporal EEG brain perception data is performed in this study. An authentication methodology is proposed and illustrated on several small-scale examples of classification of EEG human brain perception data collected on audio and visual stimuli pairs. A methodology for person identification is proposed that uses a certain audio and/or video stimulus as a “security key” for the authentication process. The stimuli pairs used for experiments in this study were created the following way; an audio pair is highly structural classical music versus disordered / chaotic noise, a visual pair is a set of opposing structural repetitive archetypical video patterns from an abstract modern art video, and also mixed audio/video pairs are used. The term of “brainprints” is offered by the analogy with fingerprints and prospectively having the ability to supply similar functionality but with an even higher level of security. In addition the following hypotheses on the nature of human creativity are proposed in this study: human creativity might be defined as naturally inherited human ability and necessity to decipher, digest and transfer the universe global programme (human-independent) into patterned structures, expressed in some unique distinctive way. Therefore the concept of a genius might be defined as human ability to decipher and translate “global universe postulates” into human-readable patterns, performed the best possible way for the majority of a certain population.

Description

Keywords

Pattern, Spatio-temporal EEG data classification, Perception of art

Source

DOI

Publisher's version

Rights statement

Permanent link

https://hdl.handle.net/10292/7702

Collections

Masters Theses