Identification of Epileptic Seizures by Characterizing Instantaneous Energy Behavior of EEG

Abstract

Automatic seizure detection has been often treated as a classification problem that aims at determining the label of electroencephalogram (EEG) signals by computer science, as the EEG monitoring is a helpful adjunct to the diagnosis of epilepsy. In most existing work, the traditional signal energy of the EEG has been applied for classification, since the energy pattern of epileptic seizures differs from that of non-seizures. Although they are effective, the accuracy either heavily depends on additional information besides energy or is limited by the shortcoming of energy-based features. To address this issue, the proposed approach achieves the classification based on the instantaneous energy of the EEG signals instead. The proposed approach first measures the instantaneous energy related to changes in the EEG signals. Then, energy behavior over time is characterized by instantaneous energy-based features from different aspects. Finally, the classification is carried out on the features to produce output labels. By processing instantaneous energy, the information of energy evolution is involved. As such, the accuracy is improved without bringing in extra information besides energy, or complicated transformation. In multi-class problems, the proposed approach has obtained promising results for identifying the ictal EEG, which indicates the tremendous potential of the proposed approach for epileptic seizure detection. INDEX TERMSSeizure identification, electroencephalogram (EEG) signal, instantaneous energy, energy behavior.

Seizure Detection: A Low Computational Effective Approach without Classification Methods

“…As the EEG signals contain normal physiological rhythms, the difficulty in identifying critical biomarkers prevents the exclusive use of rhythms for seizure detection. Hence, other discriminatory features need to be used to enhance the correctness of seizure identification [ 38 , 39 , 40 ]. Production of the “feature signal” requires four low-computational steps: signal preprocessing, EEG signals averaging, application of the feature filter, candidate seizure epochs detection, and production of the signals mask.…”

Classification of P300 based brain computer interface systems using long short-term memory (LSTM) neural networks with feature fusion

“…is the Fourier transform of the analytic similarity matrix z[n] derived from the raw data x(n). M is the length of the z x [k] signal obtained by Fourier transform [30,31]. The signal-power spectrum output of 1 participant at 15 Hz frequency resolution via feature fusion is shown in Figure 9 (a).…”

Robust Optimization of electroencephalograph (EEG) Signals for Epilepsy Seizure Prediction by utilizing VSPO Genetic Algorithms with SVM and Machine Learning Methods

“…In order to provide output labels, the functions are measured and graded. The knowledge of energy evolution is used in the production of instantaneous energy (5). The authors explore to differentiate between natural and pathological signals immediately and to improve the accuracy of epileptic seizures while reducing the expense of computing.…”