6.3.3 - Tiltmeter-Based Forecasting and Anomaly Detection for Proactive Structural Monitoring of Railway Infrastructure
- Event
- EUROSENSORS 2026
2026-09-06 - 2026-09-09
Zurich - Band
- Lectures
- Chapter
- Signal Analysis, Machine Learning and Artificial Intelligence for Sensing
- Author(s)
- P. Saha, F. Hernandez-Ramirez, J. D. Prades - Technische Universität Braunschweig,Braunschweig (Germany)
- Pages
- 213 - 214
- DOI
- 10.5162/eurosensors2026/6.3.3
- ISBN
- 978-3-910600-12-6
- Price
- free
Abstract
We present a Graph Deviation Network (GDN) framework to detect early alterations in the neuronal network of human Forebrain Organoids (hFOs) associated with schizophrenia risk. MEA-recorded signals from PCCB-knockdown hFOs are preprocessed into amplitude-modulated spike arrays and represented as time-varying graphs. Topological descriptors extracted from these graphs feed a linear discriminant analysis classifier, achieving 91% accuracy in distinguishing PCCB-knockdown from control hFOs at early time sessions. This approach provides a powerful tool for identifying disease-relevant electrophysiological signatures in neuronal organoid models.