An Empirical Investigation of Reconstruction-Based Models for Seizure Prediction from ECG Signals

2025-04-11

AI Summary

To address the high cost, noise sensitivity, and clinical deployment challenges of EEG-based seizure prediction, this paper proposes a novel non-invasive paradigm leveraging ECG signals. The core idea is to model normal heart rate dynamics and detect deviations in their reconstruction to capture preictal autonomic nervous system abnormalities. We first systematically demonstrate a strong correlation between ECG reconstruction error and impending seizures. A hybrid end-to-end time-frequency feature-driven reconstruction model is introduced, integrating STFT and CWT for spectral representation, deep autoencoders for nonlinear dimensionality reduction, and LSTM-Transformer architectures for temporal dependency modeling. Anomaly scoring is performed via sliding-window reconstruction error aggregation. Evaluated on the Siena dataset, the method achieves a mean prediction horizon of 45 minutes, with specificity of 99.16%, accuracy of 76.05%, and a false positive rate of 0.01/h—substantially enhancing practicality and robustness for clinical deployment.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Intelligent Robots: State Estimation
• Humans and AI: Brain-Sensing and Analysis

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Large-scale security measurements
• Economics, Online Markets and Human Computation: Sustainability of Web economics

Abstract

Epileptic seizures are sudden neurological disorders characterized by abnormal, excessive neuronal activity in the brain, which is often associated with changes in cardiovascular activity. These disruptions can pose significant physical and psychological challenges for patients. Therefore, accurate seizure prediction can help mitigate these risks by enabling timely interventions, ultimately improving patients' quality of life. Traditionally, EEG signals have been the primary standard for seizure prediction due to their precision in capturing brain activity. However, their high cost, susceptibility to noise, and logistical constraints limit their practicality, restricting their use to clinical settings. In order to overcome these limitations, this study focuses on leveraging ECG signals as an alternative for seizure prediction. In this paper, we present a novel method for predicting seizures based on detecting anomalies in ECG signals during their reconstruction. By extracting time-frequency features and leveraging various advanced deep learning architectures, the proposed method identifies deviations in heart rate dynamics associated with seizure onset. The proposed approach was evaluated using the Siena database and could achieve specificity of 99.16%, accuracy of 76.05%, and false positive rate (FPR) of 0.01/h, with an average prediction time of 45 minutes before seizure onset. These results highlight the potential of ECG-based seizure prediction as a patient-friendly alternative to traditional EEG-based methods.

Problem

• Predicting seizures using ECG signals instead of EEG
• Detecting anomalies in ECG signal reconstruction for seizure prediction
• Improving patient quality of life with timely seizure interventions

Innovation

• Uses ECG signals for seizure prediction
• Employs deep learning for anomaly detection
• Leverages time-frequency feature extraction