V Congreso Latinoamericano de Ingeniería Biomédica (CLAIB2011)

EEG pattern recognition by multidimensional space-time-frequency analysis
Martínez Montes Eduardo, Márquez Bocalandro Yuniel, Lage Castellanos Agustín

Última modificación: 2011-03-26 11:41


Multidimensional space-time-frequency analysis of the EEG is possible via the Parallel Factor Analysis (PARAFAC). In this work we show an application of this methodology to the characterization of epileptic EEG data. Then, we illustrate with real and simulated data its potential as a feature extraction method, which can be combined with classification methods to develop a technique for automatic detection of epileptic seizures, based on the recognition of pattern spatial and spectral profiles.