This project investigates the use of sleep behavior data to predict changes in the diagnosis of Alzheimer's Disease. By evaluating the sleep patterns and behaviors in healthy individuals and those with mild cognitive impairment (MCI), we seek to understand how sleep influences the likelihood of a diagnosis change to Alzheimer's Disease.
Explorative Data Analysis, Machine Learning, Python, scikit-learn
The primary goal of this study is to examine sleep patterns and behaviors in healthy individuals and those with mild cognitive impairment (MCI) before their diagnoses change to Alzheimer's Disease (AD). By training machine learning models on sleep behavior data and analyzing their performance, we aim to identify potential connections between sleep patterns and disease progression, as well as help establish the importance of sleep as a predictive factor for the development of MCI and AD.
Previous research has established a link between sleep quality and cognitive well-being, suggesting that sleep disturbances might be an indicator for the development of mild cognitive impairment (MCI) and Alzheimer's Disease (AD). Hence, analyzing sleep behaviors of healthy and MCI patients can provide insights into the progression of the diseases. The Alzheimer's Disease Neuroimaging Initiative (ADNI) provides an extensive dataset of patients information collected over an extended period of time.
The results suggest that sleep-related measurements might contain valuable information for predicting the onset of cognitive decline and neurodegenerative diseases, emphasizing the potential of leveraging sleep patterns for early detection and intervention strategies. The analysis found that certain machine learning models, particularly decision trees, K-nearest neighbors (KNN), and support vector machines (SVM) performed well in predicting the transition from a healthy state or mild cognitive impairment (MCI) to Alzheimer's Disease using sleep features. These models demonstrated their capability to capture complex relationships in the sleep behavior data.
Duration: 1 Semester, Effort in hours: 360, Team size: 2 People
Prof. Dr. Arzu Cöltekin and Dr Leticia Fernández Moguel
Institute for Interactive Technologies
University of Applied Science and Arts Northwestern Switzerland
Navjot Zubler (BSc Data Science), Lukas Reber (BSc Data Science)
Arzu Cöltekin (arzu.coltekin@fhnw.ch), Leticia Fernández Moguel (leticia.fernandezmoguel@fhnw.ch)
