Título: Detecção de Espículas em Registros de Eletroencefalograma utilizando um Mapa Auto-organizável
Autores: Guedes, Jorge Roberto; Azevedo, Fernando Mendes
Resumo: This work describes a technique for spike detection on EEG records by using self-organizing maps. The EEG database was obtained from 7 patients and every record is 30 minutes long. These records own 32 channels at a sampling rate of 100 Hz. The analysis is made at intervals one second long. For each interval, an algorithm reckons the following parameters: maximum and minimum amplitudes, difference between the amplitudes, event duration, event angle, average, standard deviation, variance, crest factor and entropy. After that, the results are supplied as inputs for a self-organizing map, which was trained with parameters of signals containing spikes, alpha waves, background activity, blinks and noise. In the simulation, the self-organizing map distributes the parameters in clusters. By analyzing these clusters, an algorithm verifies the existence of spikes.
Código DOI: 10.21528/CBRN2003-086
Artigo em PDF: 6CBRN_086.PDF
Arquivo BibTex: 6CBRN_086.bib