Wearable Seizure Detection and Alert System for Neurodegenerative Patients

Abstract

Neurodegenerative disorders, such as epilepsy, pose a significant threat to patients quality of life. Seizures can occur unexpectedly, leading to injuries, emotional trauma, and social stigma. The study presents a wearable seizure detection and alert system designed for neurodegenerative patients. Our system utilizes a multimodal sensor approach, incorporating electrodermal activity (EDA) sensors to measure skin conductance, gyroscope and accelerometer sensors for motion detection, and pulse oximetry (SpO2) and heart rate sensors for vital sign monitoring. Machine learning algorithms are employed to analyze the sensor data and detect seizure onset. Upon detection, the system alerts caregivers and emergency services through a mobile application, ensuring timely intervention and support. The results demonstrate the effectiveness of the proposed system in detecting seizures with high accuracy, sensitivity, and specificity. The wearable system has the potential to improve the lives of neurodegenerative patients, enhancing their safety, independence, and overall well-being.

Introduction

Overview

Neurodegenerative disorders like epilepsy cause seizures that significantly impact patient safety and quality of life. Traditional detection methods like EEG and video monitoring are invasive, costly, and limited to clinical settings. This study proposes a wearable, real-time seizure detection system using multimodal sensors and machine learning, offering a non-invasive, affordable, and mobile solution to monitor and manage seizures.

Problem Statement

• Seizure detection remains difficult due to unpredictable onset and reliance on patient or caregiver input.

• Existing systems often use single sensors, lack real-time alerts, and fail to detect the full range of seizure symptoms.

Proposed Solution

The system integrates multiple sensors:

• Electrodermal Activity (EDA)

• Heart Rate and SpO2 (MAX30100)

• Accelerometer and Gyroscope (MPU6050)

These sensors feed into an Arduino Nano microcontroller. Data is analyzed using:

• Signal preprocessing (e.g., filtering noise)

• Feature extraction (e.g., BPM, motion spikes, oxygen dips)

• Threshold-based detection

• Lightweight decision tree machine learning model (for improved accuracy with low hardware requirements)

Alerts are sent to caregivers via Bluetooth (HC-05) and displayed on a mobile app developed using MIT App Inventor.

Key Features

• Real-time monitoring

• Multimodal detection for higher accuracy

• Machine learning for improved classification

• Mobile alert system

• Long battery life (~10+ hours)

• Low-cost, compact, wearable design

Related Research

1. Parkinson’s patients – Combined EEG and accelerometer data improves seizure detection.

2. Alzheimer’s patients – Accelerometer + heart rate sensor system provides quick alerts.

3. Machine Learning Models – Used to predict seizures from EEG and motion data.

4. Non-invasive wearable systems – Combine motion and autonomic signals for real-time alerts.

5. EEG + Accelerometer Fusion – Enhanced accuracy for neurodegenerative seizure monitoring.

These works support the importance of multi-sensor integration and ML in wearable seizure detection.

System Methodology

1. Sensor Data Acquisition: Real-time capture of EDA, heart rate, SpO2, motion.

2. Signal Preprocessing: Filtering noise and normalizing data for consistent interpretation.

3. Feature Extraction: Identifying key metrics (e.g., BPM spikes, motion peaks).

4. Threshold-Based Detection: Initial layer of logic based on experimental thresholds.

5. ML Integration: Decision tree model trained offline, implemented in microcontroller firmware.

6. Alert System: Bluetooth alerts to app, including patient info and sensor data.

7. Simulation & Testing: Lab-based trials with simulated seizures showed >90% accuracy and alerts within 1.5 seconds.

Hardware Components

• Arduino Nano: Main controller for sensor data processing.

• Heart Rate Sensor (MAX30100/AD8232): Tracks pulse and oxygen levels.

• EDA Sensor (GSR): Detects stress and autonomic response changes.

• SpO2 Sensor: Monitors oxygen saturation, indicating respiratory issues.

• Accelerometer: Captures seizure-like movement patterns.

Conclusion

This study presents a portable, cost-effective wearable system designed for real-time seizure detection, particularly in individuals with neurodegenerative disorders. It employs a suite of physiological and motion sensorsincluding electrodermal activity (EDA), blood oxygen saturation (SpO?), heart rate, accelerometer, and gyroscopeto continuously track biomarkers commonly affected during seizure onset.Built around an Arduino-based microcontroller, the system offers low power consumption, affordability, and ease of use, making it suitable for clinical and home-care settings. During prototype testing using simulated conditions, the device effectively differentiated seizure-like episodes from normal activities, showing promising reliability.The system features Bluetooth connectivity, allowing real-time alerts to be transmitted to a caregiver\'s mobile device, enabling faster emergency responses and improved patient safety. Its open-source hardware and software design promotes customization, scalability, and adoption by researchers and developers.Detection is currently handled through a hybrid approach combining threshold logic with lightweight decision tree classifiers. Planned improvements include the integration of advanced machine learning and deep learning algorithms to enhance detection accuracy and reduce false positives. Personalization based on individual health data will also be explored.Future work will focus on enhancing connectivity via IoT modules for cloud-based data sharing, allowing remote health monitoring by professionals. Clinical trials with real patients will be essential for validating the system’s performance across diverse seizure types. The wearable device represents a step forward in continuous health monitoring and timely intervention.

References

[1] Onorati, F., Regalia, G., Caborni, C., Migliorini, M., Bender, D., Poh, M. Z., ... & Picard, R. W. (2017, November). Multicenter clinical assessment of improved wearable multimodal convulsive seizure detectors. Epilepsia, 58(11), 1870–1879. Retrieved May 8, 2025. [2] Poh, M. Z., Loddenkemper, T., Reinsberger, C., Swenson, N. C., Goyal, S., Sabtala, M. C., & Picard, R. W. (2012). Convulsive seizure detection using a wrist-worn electrodermal activity and accelerometry biosensor. Epilepsia, 53(5), e93–e97. Retrieved May 8, 2025. [3] Meisel, C., Schulze-Bonhage, A., & Meisel, A. (2019, June). Deep learning from wristband sensor data: towards wearable, non-invasive seizure forecasting. arXiv preprint, arXiv:1906.00511. Retrieved May 8, 2025. [4] Aziz, A., Ur Rehman, S., & Khan, M. A. (2024, January). An Unobtrusive and Lightweight Ear-worn System for Continuous Epileptic Seizure Detection. arXiv preprint, arXiv:2401.05425. Retrieved May 8, 2025. [5] Hügle, M., Rösch, C., Gollwitzer, S., & Schulze-Bonhage, A. (2018). Early Seizure Detection with an Energy-Efficient Convolutional Neural Network on an Implantable Microcontroller. arXiv preprint, arXiv:1806.04549. Retrieved May 8, 2025. [6] Um, T. T., Pfister, F. M., Pichler, D., Endo, S., Lang, M., Hirche, S., & Fietzek, U. (2018). Parkinson’s Disease Assessment from a Wrist-Worn Wearable Sensor in Free-Living Conditions: Deep Ensemble Learning and Visualization. arXiv preprint, arXiv:1808.02870. Retrieved May 8, 2025. [7] Billeci, L., Vettori, S., Morettini, M., & Valenza, G. (2019). Biosensors for Epilepsy Management: State-of-Art and Future Aspects. Sensors, 19(7), 1525. doi:10.3390/s19071525. Retrieved May 8, 2025. [8] Van de Vel, A., Cuppens, K., Bonroy, B., Milosevic, M., Jansen, K., Van Huffel, S., & Lagae, L. (2019). Evaluation of Wearable Electronics for Epilepsy: A Systematic Review. Electronics, 9(6), 968. doi:10.3390/electronics9060968. Retrieved May 8, 2025. [9] Beniczky, S., Ryvlin, P., & Jeppesen, J. (2019). Tonic-clonic seizure detection using accelerometry-based wearable sensors: A prospective, video-EEG controlled study. Epilepsy & Behavior, 98, 106–110. doi:10.1016/j.yebeh.2019.06.008. Retrieved May 8, 2025. [10] Vergara, J., Quintana, M., & Ventura, S. (2017). An IoT Platform for Epilepsy Monitoring and Supervising. Journal of Sensors, 2017, 1–12. doi:10.1155/2017/6043069. Retrieved May 8, 2025.

Copyright

Copyright © 2025 Mr. J.A. Kevin Paul, Dharun Hari D, Navaneethan P. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.