Data mining for electroencephalogram signal processing and analysis

Abstract

Electroencephalography (EEG) is a complex signal that requires advanced signal processing and feature extraction methodologies to be interpreted correctly. EEG, is usually utilized to estimate the trace and the electrical brain activity. It is employed in the discovery and forecast of epileptic and non-epileptic seizures and neurodegenerative pathologies. In this article, we give an overview of the various computational techniques used in the past, in the present and the future to preprocess and analyze EEG signals. In particular, this work aims to briefly review the state of research in this field, trying to understand the needs of EEG analysis in the medical field, with special focus on neurodegenerative pathologies, and epileptic and not-epileptic diseases. After presenting the main pre-processing, feature selection and extraction phases, we focus on classification processes and on Data Mining techniques applied to classify EEGs. Then, through the EEG analysis a discussion of the implementation is provided to investigate, predict and diagnose some cognitive diseases and epilepsy.