Seizure Classification From EEG Signals Using Transfer Learning, Semi-Supervised Learning and TSK Fuzzy System

Abstract

Recognition of epileptic seizures from offline EEG signals is very important in clinical diagnosis of epilepsy. Compared with manual labeling of EEG signals by doctors, machine learning approaches can be faster and more consistent. However, the classification accuracy is usually not satisfactory for two main reasons: the distributions of the data used for training and testing may be different, and the amount of training data may not be enough. In addition, most machine learning approaches generate black-box models that are difficult to interpret. In this paper, we integrate transductive transfer learning, semi-supervised learning and TSK fuzzy system to tackle these three problems. More specifically, we use transfer learning to reduce the discrepancy in data distribution between the training and testing data, employ semi-supervised learning to use the unlabeled testing data to remedy the shortage of training data, and adopt TSK fuzzy system to increase model interpretability. Two learning algorithms are proposed to train the system. Our experimental results show that the proposed approaches can achieve better performance than many state-of-the-art seizure classification algorithms.

The future of sleep health: a data-driven revolution in sleep science and medicine

“…It has been suggested that personalisation may be especially useful when using data from noisy modalities, such as wearable devices 116 . Model personalising has been successfully applied in other fields, such as mood recognition 117 and seizure detection 118 . However, it remains relatively untapped in sleep science.…”

Cepstrum Coefficient Analysis from Low-Frequency to High-Frequency Applied to Automatic Epileptic Seizure Detection with Bio-Electrical Signals

“…The epileptic seizure-free EEG signals in datasets C and D were also produced, accordingly. Dataset E describes the epileptic seizure signals, which were collected by placing the electrodes in the epileptogenic zone, as shown in Table 1 [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][25][26][27][28][29][30][31][33][34][35][36][37][38][39][40][41][42][43]. The sample segment of each dataset is shown in Figure 1.…”

“…A confusion matrix generally involves the following parameters: The following aggregate metrics can then be calculated based on the above parameters. The criteria for performance evaluation are usually employed in biomedical studies, which include three parts: sensitivity (the proportion of the total number of labeled ictal EEGs that are correctly classified), specificity (the proportion of the total number of labeled inter-ictal EEGs that are correctly classified), and classification accuracy (the proportion of the total number of EEG signals that are correctly classified) [2,[6][7][8][9][10][13][14][15][16][25][26][27][28][29][30][31][32]34,35,[38][39][40][41]49,53,59].…”

“…All of these evaluations are costly, which places a heavy financial burden on patients, especially those living in areas with frequent epileptic seizure breakout and poverty. Unfortunately, these methods have many limitations: for example, they are time-consuming, tedious, subjective, [9,10,[13][14][15][16]25], and face many challenges in terms of inevitable factors on reducing misjudgments [13]. Computer-aided detection with EEG signals can therefore be considered a promising technology.…”

“…It is not only regarded as a linear classifier, but also as a nonlinear classifier. Usually, SVM can be employed to minimize the operational error and maximize the margin hyperplane to maximize the classification performance [5,7,9,10,14,25,27,35].…”