Automatic detection of interictal spikes using data mining models

A prospective candidate for epilepsy surgery is studied both the ictal and interictal spikes (IS) to determine the localization of the epileptogenic zone. In this work, data mining (DM) classification techniques were utilized to build an automatic detection model. The selected DM algorithms are: Decision Trees (J 4.8), and Statistical Bayesian Classifier (naïve model). The main objective was the detection of IS, isolating them from the EEG's base activity. On the other hand, DM has an attractive advantage in such applications, in that the recognition of epileptic discharges does not need a clear definition of spike morphology. Furthermore, previously 'unseen' patterns could be recognized by the DM with proper 'training'. The results obtained showed that the efficacy of the selected DM algorithms is comparable to the current visual analysis used by the experts. Moreover, DM is faster than the time required for the visual analysis of the EEG. So this tool can assist the experts by facilitating the analysis of a patient's information, and reducing the time and effort required in the process.

[1]  Hsiao-Wen Chung,et al.  Automatic spike detection via an artificial neural network using raw EEG data: effects of data preparation and implications in the limitations of online recognition , 2000, Clinical Neurophysiology.

[2]  Thomas G. Dietterich.  What is machine learning? , 2020, Archives of Disease in Childhood.

[3]  J. Ross Quinlan,et al.  C4.5: Programs for Machine Learning , 1992 .

[4]  Silvia Kochen,et al.  Chapter 67 Wavelet analysis preceding seizures , 2002 .

[5]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques with Java implementations , 2002, SGMD.

[6]  Michael J. Pazzani,et al.  Searching for Dependencies in Bayesian Classifiers , 1995, AISTATS.

[7]  J R Smith,et al.  Automatic analysis and detection of EEG spikes. , 1974, IEEE transactions on bio-medical engineering.

[8]  C. Binnie,et al.  A glossary of terms most commonly used by clinical electroencephalographers. , 1974, Electroencephalography and clinical neurophysiology.

[9]  Alberto Maria Segre,et al.  Programs for Machine Learning , 1994 .

[10]  G. Dietsch,et al.  Fourier-Analyse von Elektrencephalogrammen des Menschen , 1932, Pflüger's Archiv für die gesamte Physiologie des Menschen und der Tiere.

[11]  A Flexer,et al.  Statistical Methods in Medical Research Data Mining and Electroencephalography , 2022 .

[12]  Pedro M. Domingos,et al.  On the Optimality of the Simple Bayesian Classifier under Zero-One Loss , 1997, Machine Learning.

[13]  J Bancaud,et al.  [The "epileptogenic zone" in humans: representation of intercritical events by spatio-temporal maps]. , 1987, Revue neurologique.

[14]  L Diambra.  Detecting Epileptic Spikes , 2002, Epilepsia.

[15]  Fabrice Wendling,et al.  Mining reproducible activation patterns in epileptic intracerebral EEG signals: application to interictal activity , 2004, IEEE Transactions on Biomedical Engineering.

[16]  P. Chauvel,et al.  Spatio-temporal characteristics of paroxysmal interictal events in human temporal lobe epilepsy , 1995, Journal of Physiology-Paris.

[17]  Lotfi Senhadji,et al.  Epileptic transient detection: wavelets and time-frequency approaches , 2002, Neurophysiologie Clinique/Clinical Neurophysiology.

[18]  P J Bones,et al.  Wavelet Analysis of Transient Biomedical Signals and its Application to Detection of Epileptiform Activity in the EEG , 2000, Clinical EEG.

[19]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques, 3rd Edition , 1999 .

[20]  J R Carrie,et al.  A hybrid computer technique for detecting sharp EEG transients. , 1972, Electroencephalography and clinical neurophysiology.