Publication

Abstract : Focusing on rapid and reliable discrimination of EEG patterns associated with motor imagery and to evaluate the cognitive and memory performance of human user authentication in image-based password systems. This study presents a methodology which uses a nonlinear pattern recognition to study the spatial distribution of EEG patterns accompanying higher cortical functions. The multivariate decision rules associate EEG patterns differentiating performance of motor and mental tasks. These patterns discriminate between the tasks are consistent with, and extend the results of, univariate analysis of spectral intensities. Commercial EEG setup (Lievesley et al., 2011) was used to extract EEG patterns from 14 electrodes. Raw EEG signals were pre-processed using Surface Laplacian filter, Band pass filter (Babiloni et al., 2000; Blanchard and Blankertz, 2004; Cincotti et al., 2001; Dornhege et al., 2003). Fast Fourier Transform, power spectrum density and independent component analysis to extract features (Hosni et al., 2007; Olesen, 2012). The main focus have been discrimination of α and β rhythms for mental and motor imagery tasks. In this study we used EEG recordings of motor task and image based password authentication system to evaluate cognitive and memory performance of human user authentication in image-based password systems. Also we characterized the EEG signals of two different motor imagery tasks for applying as brain-based control of a robotic articulator. Time courses of two different imagery and mental tasks were investigated by the calculation of instantaneous band power changes and patterns and compared whether the quantification of these features could be classified using machine learning classifiers. We observed the variance change of 4.28 and 3.05 % allowing discrimination α - β components. We also found machine learning was not significantly reliable to discriminate both movement imagery data and image-based authentication tasks unlike α -β categorisation