Generalizable epileptic seizures prediction based on deep transfer learning.

Predicting seizures before they happen can help prevent them through medication. In this research, first, a total of 22 features were extracted from 5-s segmented EEG signals. Second, tensors were developed as inputs for different deep transfer learning models to find the best model for predicting epileptic seizures. The effect of Pre-ictal state duration was also investigated by selecting four different intervals of 10, 20, 30, and 40 min. Then, nine models were created by combining three ImageNet convolutional networks with three classifiers and were examined for predicting seizures patient-dependently. The Xception convolutional network with a Fully Connected (FC) classifier achieved an average sensitivity of 98.47% and a False Prediction Rate (FPR) of 0.031 h-1 in a 40-min Pre-ictal state for ten patients from the European database. The most promising result of this study was the patient-independent prediction of epileptic seizures; the MobileNet-V2 model with an FC classifier was trained with one patient's data and tested on six other patients, achieving a sensitivity rate of 98.39% and an FPR of 0.029 h-1 for a 40-min Pre-ictal scheme.

A review of critical challenges in MI-BCI: From conventional to deep learning methods.

Brain-computer interfaces (BCIs) have achieved significant success in controlling external devices through the Electroencephalogram (EEG) signal processing. BCI-based Motor Imagery (MI) system bridges brain and external devices as communication tools to control, for example, wheelchair for people with disabilities, robotic control, and exoskeleton control. This success largely depends on the machine learning (ML) approaches like deep learning (DL) models. DL algorithms provide effective and powerful models to analyze compact and complex EEG data optimally for MI-BCI applications. DL models with CNN network have revolutionized computer vision through end-to-end learning from raw data. Meanwhile, RNN networks have been able to decode EEG signals by processing sequences of time series data. However, many challenges in the MI-BCI field have affected the performance of DL models. A major challenge is the individual differences in the EEG signal of different subjects. Therefore, the model must be retrained from the scratch for each new subject, which leads to computational costs. Analyzing the EEG signals is challenging due to its low signal to noise ratio and non-stationary nature. Additionally, limited size of existence datasets can lead to overfitting which can be prevented by using transfer learning (TF) approaches. The main contributions of this study are discovering major challenges in the MI-BCI field by reviewing the state of art machine learning models and then suggesting solutions to address these challenges by focusing on feature selection, feature extraction and classification methods.

A transfer learning-based CNN and LSTM hybrid deep learning model to classify motor imagery EEG signals.

In the Motor Imagery (MI)-based Brain Computer Interface (BCI), users' intention is converted into a control signal through processing a specific pattern in brain signals reflecting motor characteristics. There are such restrictions as the limited size of the existing datasets and low signal to noise ratio in the classification of MI Electroencephalogram (EEG) signals. Machine learning (ML) methods, particularly Deep Learning (DL), have overcome these limitations relatively. In this study, three hybrid models were proposed to classify the EEG signal in the MI-based BCI. The proposed hybrid models consist of the convolutional neural networks (CNN) and the Long-Short Term Memory (LSTM). In the first model, the CNN with different number of convolutional-pooling blocks (from shallow to deep CNN) was examined; a two-block CNN model not affected by the vanishing gradient descent and yet able to extract desirable features employed; the second and third models contained pre-trained CNNs conducing to the exploration of more complex features. The transfer learning strategy and data augmentation methods were applied to overcome the limited size of the datasets by transferring learning from one model to another. This was achieved by employing two powerful pre-trained convolutional neural networks namely ResNet-50 and Inception-v3. The continuous wavelet transform (CWT) was used to generate images for the CNN. The performance of the proposed models was evaluated on the BCI Competition IV dataset 2a. The mean accuracy vlaues of 86%, 90%, and 92%, and mean Kappa values of 81%, 86%, and 88% were obtained for the hybrid neural network with the customized CNN, the hybrid neural network with ResNet-50 and the hybrid neural network with Inception-v3, respectively. Despite the promising performance of the three proposed models, the hybrid neural network with Inception-v3 outperformed the two other models. The best obtained result in the present study improved the previous best result in the literature by 7% in terms of classification accuracy. From the findings, it can be concluded that transfer learning based on a pre-trained CNN in combination with LSTM is a novel method in MI-based BCI. The study also has implications for the discrimination of motor imagery tasks in each EEG recording channel and in different brain regions which can reduce computational time in future works by only selecting the most effective channels.

Automatic sleep staging by cardiorespiratory signals: a systematic review.

BACKGROUND: Because of problems with the recording and analysis of the EEG signal, automatic sleep staging using cardiorespiratory signals has been employed as an alternative. This study reports on certain critical points which hold considerable promise for the improvement of the results of the automatic sleep staging using cardiorespiratory signals. METHODS: A systematic review. RESULTS: The review and analysis of the literature in this area revealed four outstanding points: (1) the feature extraction epoch length, denoting that the standard 30-s segments of cardiorespiratory signals do not carry enough information for automatic sleep staging and that a 4.5-min length segment centering on each 30-s segment is proper for staging, (2) the time delay between the EEG signal extracted from the central nervous system activity and the cardiorespiratory signals extracted from the autonomic nervous system activity should be considered in the automatic sleep staging using cardiorespiratory signals, (3) the information in the morphology of ECG signals can contribute to the improvement of sleep staging, and (4) applying convolutional neural network (CNN) and long short-term memory network (LSTM) deep structures simultaneously to a large PSG recording database can lead to more reliable automatic sleep staging results. CONCLUSIONS: Considering the above-mentioned points simultaneously can improve automatic sleep staging by cardiorespiratory signals. It is hoped that by considering the points, staging sleep automatically using cardiorespiratory signals, which does not have problems with the recording and analysis of EEG signals, yields results acceptably close to the results of automatic sleep staging by EEG signals.

Automatic sleep staging using empirical mode decomposition, discrete wavelet transform, time-domain, and nonlinear dynamics features of heart rate variability signals.

The conventional method for sleep staging is to analyze polysomnograms (PSGs) recorded in a sleep lab. The electroencephalogram (EEG) is one of the most important signals in PSGs but recording and analysis of this signal presents a number of technical challenges, especially at home. Instead, electrocardiograms (ECGs) are much easier to record and may offer an attractive alternative for home sleep monitoring. The heart rate variability (HRV) signal proves suitable for automatic sleep staging. Thirty PSGs from the Sleep Heart Health Study (SHHS) database were used. Three feature sets were extracted from 5- and 0.5-min HRV segments: time-domain features, nonlinear-dynamics features and time-frequency features. The latter was achieved by using empirical mode decomposition (EMD) and discrete wavelet transform (DWT) methods. Normalized energies in important frequency bands of HRV signals were computed using time-frequency methods. ANOVA and t-test were used for statistical evaluations. Automatic sleep staging was based on HRV signal features. The ANOVA followed by a post hoc Bonferroni was used for individual feature assessment. Most features were beneficial for sleep staging. A t-test was used to compare the means of extracted features in 5- and 0.5-min HRV segments. The results showed that the extracted features means were statistically similar for a small number of features. A separability measure showed that time-frequency features, especially EMD features, had larger separation than others. There was not a sizable difference in separability of linear features between 5- and 0.5-min HRV segments but separability of nonlinear features, especially EMD features, decreased in 0.5-min HRV segments. HRV signal features were classified by linear discriminant (LD) and quadratic discriminant (QD) methods. Classification results based on features from 5-min segments surpassed those obtained from 0.5-min segments. The best result was obtained from features using 5-min HRV segments classified by the LD classifier. A combination of linear/nonlinear features from HRV signals is effective in automatic sleep staging. Moreover, time-frequency features are more informative than others. In addition, a separability measure and classification results showed that HRV signal features, especially nonlinear features, extracted from 5-min segments are more discriminative than those from 0.5-min segments in automatic sleep staging.

Automatic sleep stage classification based on EEG signals by using neural networks and wavelet packet coefficients.

Currently in the world there is an alarming number of people who suffer from sleep disorders. A number of biomedical signals, such as EEG, EMG, ECG and EOG are used in sleep labs among others for diagnosis and treatment of sleep related disorders. The usual method for sleep stage classification is visual inspection by a sleep specialist. This is a very time consuming and laborious exercise. Automatic sleep stage classification can facilitate this process. The definition of sleep stages and the sleep literature show that EEG signals are similar in Stage 1 of non-rapid eye movement (NREM) sleep and rapid eye movement (REM) sleep. Therefore, in this work an attempt was made to classify four sleep stages consisting of Awake, Stage 1 + REM, Stage 2 and Slow Wave Stage based on the EEG signal alone. Wavelet packet coefficients and artificial neural networks were deployed for this purpose. Seven all night recordings from Physionet database were used in the study. The results demonstrated that these four sleep stages could be automatically discriminated from each other with a specificity of 94.4 +/- 4.5%, a of sensitivity 84.2+3.9% and an accuracy of 93.0 +/- 4.0%.