Ponente
Descripción
Abstract: The signals associated to eye movements are influenced by different noises affecting important magnitudes as saccadic peak velocity and latency. These biomarkers serve as indicators of neurological disorders in the human and therefore constitute an important tool for the diagnosis of neurodegenerative diseases such as Spinocerebellar Ataxia type 2 (SCA2). The denoising process involves several methods as Median filter (MF) well known in its performance for removing impulsive noise and preserving the edges. In this paper the Stationary Wavelet Transform (SWT) and MF, are compared taking into account Magnitude-Squared Coherence (MSC) and Signal-to-Noise Ratio (SNR) metrics to determine the performance in the denoising process of the electrooculographic signals recorded by our OpenEOG technology. With this new methodology, the results show that SWT generates saccades with less levels of noise and better preservation of the waveform. In conclusion, SWT filter proved to be the most efficient, attenuating artifacts from other bioelectrical signals and successfully removing power line noise.
Keywords: denoising; electrooculograms (EOGs); filter; OpenBCI; saccades
