Description

Narcolepsy is a sleep disorder that involves sudden sleep attacks, even if the patient feels fully rested from the night before. Due to one of its symptoms, cataplexy, appearing characteristic of an epileptic seizure, narcolepsy can get misdiagnosed as epilepsy. This leads to a longer time for the patient to acquire proper diagnosis and treatment, if at all. Machine learning models were trained and tested on sets with patients who had narcolepsy, epilepsy, or neither. This study implemented the use of convolutional neural networks (CNNs) as well as a random forest model for classification. Accuracy, precision, and F1 score were found, showing a model that could provide a foundation for future algorithms with more access to proper datasets.

Grant Proposal

Note: The following grant proposal was written for my previous project idea, which explored the effect of chronotype (a person's biological preferences for sleep and wakefulness) on different parameters of sleep stages. Since then, I have changed my project idea to what it is now.