ABSTRACT
During the past several decades, an enormous effort has been devoted by the scienti–c and clinical communities to seizure prediction in medically refractory epilepsy. Several papers have been published in journals and conference proceedings proposing different predictors to meet this challenge. These studies have evolved from straightforward descriptions of seizure precursors to controlled
30.1 Introduction .......................................................................................................................... 417 30.2 EPILAB Software Architecture ........................................................................................... 418
30.2.1 GUI Structure ........................................................................................................... 419 30.2.2 Data Structures ......................................................................................................... 419
30.3 Univariate and Multivariate Features .................................................................................. 420 30.3.1 Data Preprocessing ................................................................................................... 420 30.3.2 Feature Extraction and Feature Reduction ............................................................... 420
30.3.2.1 EEG Features ............................................................................................. 421 30.3.2.2 ECG............................................................................................................ 422 30.3.2.3 Features Selection and Reduction .............................................................. 422
30.4 Seizure Prediction Algorithms ............................................................................................. 422 30.4.1 Computational Intelligence Algorithms ................................................................... 422
30.4.1.1 Arti–cial Neural Networks ........................................................................ 423 30.4.1.2 Support Vector Machines ..........................................................................424
