A Physiology-Based Seizure Detection System for Multichannel EEG

Abstract

BackgroundEpilepsy is a common chronic neurological disorder characterized by recurrent unprovoked seizures. Electroencephalogram (EEG) signals play a critical role in the diagnosis of epilepsy. Multichannel EEGs contain more information than do single-channel EEGs. Automatic detection algorithms for spikes or seizures have traditionally been implemented on single-channel EEG, and algorithms for multichannel EEG are unavailable.MethodologyThis study proposes a physiology-based detection system for epileptic seizures that uses multichannel EEG signals. The proposed technique was tested on two EEG data sets acquired from 18 patients. Both unipolar and bipolar EEG signals were analyzed. We employed sample entropy (SampEn), statistical values, and concepts used in clinical neurophysiology (e.g., phase reversals and potential fields of a bipolar EEG) to extract the features. We further tested the performance of a genetic algorithm cascaded with a support vector machine and post-classification spike matching.Principal FindingsWe obtained 86.69% spike detection and 99.77% seizure detection for Data Set I. The detection system was further validated using the model trained by Data Set I on Data Set II. The system again showed high performance, with 91.18% detection of spikes and 99.22% seizure detection.ConclusionWe report a de novo EEG classification system for seizure and spike detection on multichannel EEG that includes physiology-based knowledge to enhance the performance of this type of system.

The detection of epileptic seizure signals based on fuzzy entropy

“…Using ANN, Akareddy et al (2013) studied the EEG signals of epileptics based on ApEn, with a classification accuracy of 90%. With the calculated SampEn adopted as the index, Shen et al (2013) also conducted classifications of epilepsy, and their calculated accuracy was as high as 91.18%. As stated above, favorable classification results have been achieved with the adoption of multiple types of entropy, suggesting that in general, entropy-based methods are promising for the EEG analysis of epilepsy.…”

A retrospective glance at automatic detection of epileptic spike in electroencephalogram

“…The early automatic artifact rejection strategies are based on context features (1,48). Many recent studies implement an artifacts removal preprocess, usually pre-threshold on raw EEG (19,33,60). Wavelet transform has also been proposed due to its ability to separate the artifact subbands, in which case the artifacts will be removed easily (46).…”

“…Advanced SVM maps data vectors into higher dimension space to solve nonlinear problems. Due to these qualities, several recent researchers have made attempts in the utility of SVM (1,52,60).…”

“…Lower value of sample entropy indicates higher regularity. Sample entropy is robust to noisy physiological time series and thus could be useful for detecting seizures or spikes in EEG data (60).…”

Computer assisted interpretation of the human EEG

“…Recent studies have focused on either one or a combination of the following approaches: parametric methods; mimetic analysis; Fourier or wavelet analysis; artificial neural networks; template matching; context-based rules; and event clustering (Scherg et al, 2012;Nonclercq et al, 2012Nonclercq et al, , 2009Halford et al, 2012;Ji et al, 2011a;Wang et al, 2010;De Lucia et al, 2008;Indiradevi et al, 2008;Zhou et al, 2012;Lodder et al, 2013;Argoud et al, 2006;Subasi, 2006;Inan and Kuntalp, 2007;Van Hese et al, 2008;Kumar et al, 2012;Shen et al, 2013). Clearly, a lot of thought has been given to e ective ways of detecting inter-ictal activity.…”