Keywords
Epilepsy; CNN; Deep Learning; Data Science; Seizure; EEG
Abstract
Epilepsy is a common brain disorder where neurons in the brain rapidly fire, causing recurring seizures. The brain activity during a seizure can be detected by electroencephalogram (EEG) signals; however, this process is not only labor-intensive and time-consuming but is also subject to inter-rater variability, with a study showing only moderate agreement when diagnosing patients, even among experts. Convolutional Neural Networks (CNNs) are often proposed to detect seizures automatically, achieving high performance. The focus on performance comes at a cost of losing interpretability, leaving the model as effective but seen as a ’black box’. This thesis confronts the interpretability knowledge gap by conducting a theoretical analysis of a 1D CNN trained on the Bonn EEG Dataset. The analysis reveals how exactly the model learns, showing that the first convolutional layer develops specific filters. Across most classification tasks, the model learned to focus on the approximately 22 Hz beta-wave band as a key neurophysiological feature. Furthermore, the model’s stability was quantified using the Lipschitz bound. This work successfully bridges the gap between high-performance metrics and theoretical understanding, providing a framework for interpreting CNN based seizure detection.
Thesis Completion Year
2025
Thesis Completion Semester
Fall
Thesis Chair
Chudamani Poudyal
College
College of Sciences
Department
School of Data, Mathematical, and Statistical Sciences (SDMSS)
Thesis Discipline
Data Science
Language
English
Access Status
Open Access
Length of Campus Access
None
Campus Location
Orlando (Main) Campus
STARS Citation
Small, Jackson T., "Theoretical Analysis of CNNs for Automatic Seizure Detection in EEG Signals" (2025). Honors Undergraduate Theses. 462.
https://stars.library.ucf.edu/hut2024/462
Rights Statement
