The deep learning-based automatic detection of epilepsy electroencephalogram (EEG), which can avoid the artificial influence, has attracted much attention, and its effectiveness mainly depends on the deep neural network model. In this paper, an attention-based multi-scale residual network (AMSRN) was proposed in consideration of the multiscale, spatio-temporal characteristics of epilepsy EEG and the information flow among channels, and it was combined with multiscale principal component analysis (MSPCA) to realize the automatic epilepsy detection. Firstly, MSPCA was used for noise reduction and feature enhancement of original epilepsy EEG. Then, we designed the structure and parameters of AMSRN. Among them, the attention module (AM), multiscale convolutional module (MCM), spatio-temporal feature extraction module (STFEM) and classification module (CM) were applied successively to signal reexpression with attention weighted mechanism as well as extraction, fusion and classification for multiscale and spatio-temporal features. Based on the Children’s Hospital Boston-Massachusetts Institute of Technology (CHB-MIT) public dataset, the AMSRN model achieved good results in sensitivity (98.56%), F1 score (98.35%), accuracy (98.41%) and precision (98.43%). The results show that AMSRN can make good use of brain network information flow caused by seizures to enhance the difference among channels, and effectively capture the multiscale and spatio-temporal features of EEG to improve the performance of epilepsy detection.
Epilepsy is a complex and widespread neurological disorder that has become a global public health issue. In recent years, significant progress has been made in the use of wearable devices for seizure monitoring, prediction, and treatment. This paper reviewed the applications of invasive and non-invasive wearable devices in seizure monitoring, such as subcutaneous EEG, ear-EEG, and multimodal sensors, highlighting their advantages in improving the accuracy of seizure recording. It also discussed the latest advances in the prediction and treatment of seizure using wearable devices.
The existing epilepsy seizure detection algorithms have problems such as overfitting and poor generalization ability due to high reliance on manual labeling of electroencephalogram’s data and data imbalance between seizure and interictal periods. An unsupervised learning detection method for epileptic seizure that jointed graph attention network (GAT) and Transformer framework (GAT-T) was proposed. In this method, channel correlations were adaptively learned by GAT encoder. Temporal information was captured by one-dimensional convolution decoder. Combining outputs of the two mentioned above, predicted values for electroencephalogram were generated. The collective anomaly score was calculated and the detection threshold was determined. The results demonstrated that GAT-T achieved the average performance exceeding 90% (or 99%) with a 0.25 s (or 2 s) time segment length, which could effectively detect epileptic seizures. Moreover, the channel association probability matrix was expected to assist clinicians in the initial screening of the epileptogenic zone, and ablation experiments also reflected the significance of each module in GAT-T. This study may assist clinicians in making more accurate diagnostic and therapeutic decisions for epilepsy patients.