ICAM: A Non-Invasive Low-Cost Wearable Device for Resource-Efficient Monitoring of Non-Convulsive Epileptic Seizures
CSEF · 2026 Medicine & Physiology (Senior Division)
Overview
Epilepsy is a neurological disorder affecting approximately 50 million people worldwide and involves repeated seizures from abnormal neuron firing. Around 60% of patients have non-convulsive seizures (NCS), which include focal aware and impaired awareness seizures. Because NCS lack visible motor symptoms, they require continuous scalp or invasive stereoelectroencephalography (EEG) monitoring, which is costly and mandates specialist oversight. Given these challenges, current methods remain inaccessible to 75% of patients in under-resourced countries. An alternative biomarker for NCS detection is ictal central apnea (ICA), the concurrent autonomic changes of respiratory arrest and heart-rate variability. The goal of ICAM is to develop a low-cost wearable device that detects ICA non-invasively through photoplethysmography (PPG) and respiratory surface electromyography (sEMG). To support low-compute environments, ICAM proposes a novel architecture called adaptive multi-scale Echo State Network (AMES-ESN), which was deployed on low-cost hardware for real-time ICA predictions from sEMG and PPG. ICA annotations from synthetic training data were clinically validated, exhibiting strong similarity to real EMG and PPG signals (RMSSD: 3.3 ms, RRD: 1.06 bpm). AMES-ESN exceeds engineering goals in efficiency on ICAM hardware (548% less memory than state-of-the-art) and performance across EEG vs EMG/PPG modalities by ±5% for precision, recall, and F1-Score (94.10% vs. 84.52% 1D-CNN-LSTM; 77.40% SVM; p<0.001), all while being ~210x cheaper than state-of-the-art ($111.72). This preventive technology will pave the path to accessible seizure monitoring and care, particularly in rural and underfunded communities, potentially saving the lives of millions who struggle with NCS.
Competition history
- CSEF 2026
Related projects
ISEF · 2020
Real-Time Seizure Forecasting for Epileptics on a Consumer Product
JSHS · 2025
Design and Testing of an Adjustable Functional Near-Infrared Spectroscopy (fNIRS)
AJAS · 2020
An Intelligent System for the Prediction of Epileptic Seizures
ISEF · 2021
Design and Development of an Alert Prototype to Assist People with Simple Motor Seizures Events Using Triaxial Micro-sensors
ISEF · 2020
A Biologically-inspired, Biomarker-driven, Rapid Early Warning System for Epileptic Onset Prediction and Seizure Detection Using Machine Learning
ISEF · 2026
A Multimodal Approach to Real-Time Seizure Prediction Using Discreet Wearables
ISEF · 2026
NeuroShield: Personalized Adaptive Filtration and Clinical Trigger Mapping for Photosensitive Epilepsy
ISEF · 2019
Engineering a Novel Wearable Biosensing Mechanism through the Implementation of Microelectromechanical Systems and Machine Learning to Realize Anomalies Hinting towards Future Cardiac Episodes
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: California Science & Engineering Fair public projects