Early epileptic seizure prediction (ESP) has informative challenges due to the complexity of electroencephalogram (EEG) signals, patient variability, privacy, security issues regarding consumer health data, and the on-time alarm triggers before an upcoming seizure to provide sufficient time for the patients and caregivers to take appropriate action. Therefore, the proposed study presents a novel patient-specific seizure prediction framework with the Consumer Internet of Things (CIoT) for the smart healthcare system to anticipate the onset of upcoming seizures and improve the quality of healthcare and early treatment. Initially, the multi-handicraft and deep features are extracted in feature engineering modules and then concatenated in the fusion module. The fused feature fed into the Bidirectional Long Short-term Memory (BiLSTM) network to present the temporal dependency of the EEG signals. The Attention Mechanism is applied to reduce the dimension of the feature. Moreover, the CIoT module is integrated for real-time seizure prediction and sending alerts to doctors and emergency units through the cloud platform. Testing via the Leave-One-Out cross-validation method revealed the model’s consistent performance across various seizure types, emphasizing real-time clinical applications. The model achieved 91.39± 3.34% accuracy, 91.30± 2.80% sensitivity, 90.06± 4.84% specificity, and a false positive rate (FPR) of 0.12± 0.04 h-1 in the case of seizure prediction Horizon (SPH) of 10 minutes. The CIoT remotely monitoring the patients ensures timely treatment, maintains data security and privacy, and improves the performance of real-time applications in smart consumer healthcare systems.

Epilepsy is a neurological disorder characterized by abnormal neuronal discharges that manifest in life-threatening seizures. These are often monitored via EEG signals, a key aspect of biomedical signal processing (BSP). Accurate epileptic seizure (ES) detection significantly depends on the precise identification of key EEG features, which requires a deep understanding of the data's intrinsic domain. Therefore, this study presents an Advanced Multi-View Deep Feature Learning (AMV-DFL) framework based on machine learning (ML) technology to enhance the detection of relevant EEG signal features for ES. Our method initially applies a fast Fourier transform (FFT) on EEG data for traditional frequency domain feature (TFD-F) extraction and directly incorporates time domain (TD) features from the raw EEG signals, establishing a comprehensive traditional multi-view feature (TMV-F). Deep features are subsequently extracted autonomously from optimal layers of one-dimensional convolutional neural networks (1D CNN), resulting in multi-view deep features (MV-DF) integrating both time and frequency domains. A multi-view forest (MV-F) is an interpretable rule-based advanced ML classifier used to construct a robust, generalized classification. Tree-based SHAP explainable artificial intelligence (T-XAI) is incorporated for interpreting and explaining the underlying rules. Experimental results confirm our method's superiority, surpassing models using TMV-FL and single-view deep features (SV-DF) by 4% and outperforming other state-of-the-art methods by an average of 3% in classification accuracy. The AMV-DFL approach aids clinicians in identifying EEG features indicative of ES, potentially discovering novel biomarkers, and improving diagnostic capabilities in epilepsy management.

Epilepsy is a prevalent neurological disorder that poses life-threatening emergencies. Early electroencephalogram (EEG) seizure detection can mitigate the risks and aid in the treatment of patients with epilepsy. EEG based epileptic seizure (ES) detection has significant applications in epilepsy treatment and medical diagnosis. Therefore, this paper presents an innovative framework for efficient ES detection, providing coefficient and distance correlation feature selection algorithms, a Bagged Tree-based classifer (BTBC), and Explainable Artificial Intelligence (XAI). Initially, the Butterworth filter is employed to eliminate various artifacts, and the discrete wavelet transform (DWT) is used to decompose the EEG signals and extract various eigenvalue features of the statistical time domain (STD) as linear and Fractal dimension-based non-linear (FD-NL). The optimal features are then identified through correlation coefficients with P-value and distance correlation analysis.These features are subsequently utilized by the Bagged Tree-based classifer (BTBC). The proposed model provides best performance in mitigating overfitting issues and improves the average accuracy by 2% using (CD, E), (AB, CD, E), and (A, B) experimental types as compared to other machine learning (ML) models using well-known Bonn and UCI-EEG benchmark datasets. Finally, SHapley additive exPlanation (SHAP) was used to interpret and explain the decision-making process of the proposed model. The results highlight the framework’s capability to accurately classify ES, thereby improving the diagnosis process in patients with brain dysfunctions.

Introduction: Stereoelectroencephalography (SEEG) refers to a commonly used diagnostic procedure to localise and define the epileptogenic zone of refractory focal epilepsies, by means of minimally invasive operation techniques without large craniotomies. Objective: This study aimed to investigate the influence of different registration methods on the accuracy of SEEG electrode implantation under neuronavigation for paediatric patients with refractory epilepsy. Methods: The clinical data of 18 paediatric patients with refractory epilepsy were retrospectively analysed. The SEEG electrodes were implanted under optical neuronavigation while the patients were in the prone position. Patients were divided into two groups on the basis of the surface-based registration of MR scan method and refined anatomy-based registration of CT scan. Registration time, accuracy, and the differences between electrode placement and preoperative planned position were analysed. Results: Thirty-six electrodes in 7 patients were placed under surface-based registration of MR scan, and 45 electrodes in 11 patients were placed under refined anatomy-based registration of CT scan. The registration time of surface-based registration of MR scan and refined anatomy-based registration of CT scan was 45 ± 12 min and 10 ± 4 min. In addition, the mean registration error, the error of insertion point, and target error were 3.6 ± 0.7 mm, 2.7 ± 0.7 mm, and 3.1 ± 0.5 mm in the surface-based registration of MR scan group, and 1.1 ± 0.3 mm, 1.5 ± 0.5 mm, and 2.2 ± 0.6 mm in the refined anatomy-based registration of CT scan group. The differences between the two registration methods were statistically significant. Conclusions: The refined anatomy-based registration of CT scan method can improve the registration efficiency and electrode placement accuracy, and thereby can be considered as the preferred registration method in the application of SEEG electrode implantation under neuronavigation for treatment of paediatric intractable epilepsy.