Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SeizNet: An AI-enabled Implantable Sensor Network System for Seizure Prediction

Ali Saeizadeh, Douglas Schonholtz, Daniel Uvaydov, Raffaele Guida, Emrecan Demirors, Pedram Johari, Jorge M. Jimenez, Joseph S. Neimat, Tommaso Melodia

TL;DR

SeizNet addresses the challenge of predicting epileptic seizures in drug-resistant patients by fusing multi-modal iEEG and ECG data within an edge-enabled implantable/wearable sensor network. It deploys patient-specific deep learning models with a focal loss to handle severe class imbalance and employs time-channel voting to stabilize predictions, all executed at the edge via ultrasonic intra-body communication. Evaluated on the EPILEPSIAE dataset, SeizNet achieves exceptionally high performance, reporting near-100% sensitivity, specificity, and accuracy, with very low false-positive rates when combining modalities, demonstrating the practicality of a closed-loop, low-power seizure prediction and intervention system. The work highlights significant implications for real-time, privacy-conscious seizure management and responsive neuromodulation in refractory epilepsy.

Abstract

In this paper, we introduce SeizNet, a closed-loop system for predicting epileptic seizures through the use of Deep Learning (DL) method and implantable sensor networks. While pharmacological treatment is effective for some epilepsy patients (with ~65M people affected worldwide), one out of three suffer from drug-resistant epilepsy. To alleviate the impact of seizure, predictive systems have been developed that can notify such patients of an impending seizure, allowing them to take precautionary measures. SeizNet leverages DL techniques and combines data from multiple recordings, specifically intracranial electroencephalogram (iEEG) and electrocardiogram (ECG) sensors, that can significantly improve the specificity of seizure prediction while preserving very high levels of sensitivity. SeizNet DL algorithms are designed for efficient real-time execution at the edge, minimizing data privacy concerns, data transmission overhead, and power inefficiencies associated with cloud-based solutions. Our results indicate that SeizNet outperforms traditional single-modality and non-personalized prediction systems in all metrics, achieving up to 99% accuracy in predicting seizure, offering a promising new avenue in refractory epilepsy treatment.

SeizNet: An AI-enabled Implantable Sensor Network System for Seizure Prediction

TL;DR

SeizNet addresses the challenge of predicting epileptic seizures in drug-resistant patients by fusing multi-modal iEEG and ECG data within an edge-enabled implantable/wearable sensor network. It deploys patient-specific deep learning models with a focal loss to handle severe class imbalance and employs time-channel voting to stabilize predictions, all executed at the edge via ultrasonic intra-body communication. Evaluated on the EPILEPSIAE dataset, SeizNet achieves exceptionally high performance, reporting near-100% sensitivity, specificity, and accuracy, with very low false-positive rates when combining modalities, demonstrating the practicality of a closed-loop, low-power seizure prediction and intervention system. The work highlights significant implications for real-time, privacy-conscious seizure management and responsive neuromodulation in refractory epilepsy.

Abstract

In this paper, we introduce SeizNet, a closed-loop system for predicting epileptic seizures through the use of Deep Learning (DL) method and implantable sensor networks. While pharmacological treatment is effective for some epilepsy patients (with ~65M people affected worldwide), one out of three suffer from drug-resistant epilepsy. To alleviate the impact of seizure, predictive systems have been developed that can notify such patients of an impending seizure, allowing them to take precautionary measures. SeizNet leverages DL techniques and combines data from multiple recordings, specifically intracranial electroencephalogram (iEEG) and electrocardiogram (ECG) sensors, that can significantly improve the specificity of seizure prediction while preserving very high levels of sensitivity. SeizNet DL algorithms are designed for efficient real-time execution at the edge, minimizing data privacy concerns, data transmission overhead, and power inefficiencies associated with cloud-based solutions. Our results indicate that SeizNet outperforms traditional single-modality and non-personalized prediction systems in all metrics, achieving up to 99% accuracy in predicting seizure, offering a promising new avenue in refractory epilepsy treatment.
Paper Structure (12 sections, 4 equations, 4 figures, 1 table)

This paper contains 12 sections, 4 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Sensor Network
  3. Dataset
  4. Combined -Based Seizure Prediction
  5. Deep Learning Model
  6. Focal Loss Function
  7. Time and Channel Voting
  8. Experimental Results
  9. Performance Metrics
  10. -based perdicter with new Focal Loss Function
  11. , , and Combined Seizure Predictors
  12. Conclusion

Figures (4)

  • Figure 1: System architecture; Gateway receives the classification results from the and nodes that execute algorithms.
  • Figure 2: Deep Learning Model Structure.
  • Figure 3: SeizNet improvements using the new loss function, compared to the baseline model (AiEEG).
  • Figure 4: Average Sensitivity, Specificity, and Accuracy among all the patients with , , and combined model on test dataset.