MR-EEGWaveNet: Multiresolutional EEGWaveNet for Seizure Detection from Long EEG Recordings

2025-05-23

AI Summary

Epileptic seizure detection in long-term EEG is challenged by severe artifacts, difficulty identifying low signal-to-noise ratio (SNR) segments, and poor model generalizability. To address these issues, we propose MR-EEGWaveNet, an end-to-end multi-resolution model that innovatively integrates depthwise separable convolutions with spatiotemporal convolutions to jointly capture temporal dynamics and inter-channel spatial dependencies. A hierarchical feature dimensionality reduction and concatenation mechanism enables robust discrimination among seizure, background, and artifact classes. Additionally, an anomaly-score-based post-processing module significantly suppresses false positives. Evaluated on the Siena and Juntendo datasets, MR-EEGWaveNet achieves F1 scores of 0.336 and 0.488—improving over baselines by 89.8% and 49.2%, respectively—and accuracy gains of 15.9% and 20.62%. To our knowledge, this is the first single-model framework enabling adaptive multi-scale feature fusion and artifact-aware classification.

Technology Category

• Machine Learning: Multimodal Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Computer Vision: Multi-modal Vision

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Feature engineering for generalized seizure detection models remains a significant challenge. Recently proposed models show variable performance depending on the training data and remain ineffective at accurately distinguishing artifacts from seizure data. In this study, we propose a novel end-to-end model, ''Multiresolutional EEGWaveNet (MR-EEGWaveNet),'' which efficiently distinguishes seizure events from background electroencephalogram (EEG) and artifacts/noise by capturing both temporal dependencies across different time frames and spatial relationships between channels. The model has three modules: convolution, feature extraction, and predictor. The convolution module extracts features through depth-wise and spatio-temporal convolution. The feature extraction module individually reduces the feature dimension extracted from EEG segments and their sub-segments. Subsequently, the extracted features are concatenated into a single vector for classification using a fully connected classifier called the predictor module. In addition, an anomaly score-based post-classification processing technique was introduced to reduce the false-positive rates of the model. Experimental results were reported and analyzed using different parameter settings and datasets (Siena (public) and Juntendo (private)). The proposed MR-EEGWaveNet significantly outperformed the conventional non-multiresolution approach, improving the F1 scores from 0.177 to 0.336 on Siena and 0.327 to 0.488 on Juntendo, with precision gains of 15.9% and 20.62%, respectively.

Problem

• Detecting seizures accurately in long EEG recordings
• Distinguishing seizures from artifacts and noise effectively
• Improving F1 scores and precision in seizure detection

Innovation

• Multiresolutional EEGWaveNet for seizure detection
• Depth-wise and spatio-temporal convolution features
• Anomaly score-based post-classification processing