Background:
Sleep is a fundamental physiological process, critical for the restoration of cognitive functions, memory consolidation, and overall physical recovery. It consists of several stages, each with its own significance in promoting well-being and health. Unfortunately, conditions like attention deficit hyperactivity disorder (ADHD) have the potential to interfere with regular brain functions, often resulting in disturbed sleep patterns. Traditionally, the assessment of sleep stages and disturbances has been carried out using polysomnography. This multi-instrumentation technique records various physiological signals including brain activity, oxygen levels, heart rate, and muscle movements. However, due to its invasive nature, coupled with the associated expenses and intricate setup requirements, polysomnography is primarily suited for in-hospital sleep studies, thus making such evaluations less accessible for many individuals.
Objective and Methodology:
In partnership with the Institute of Clinical Psychology at the University of Bergen, this project will develop machine learning methods to monitor and categorize sleep stages utilizing a user-friendly wearable, the BioPoint. By capturing data such as ECG, EMG, PPG, EDA, Skin Temperature, BioImpedance, and Actigraphy, the project seeks to delve deeper into understanding sleep patterns. Such insights have the potential to refine treatments and enhance ongoing condition monitoring for individuals with sleep-related challenges.
Data Validation and Significance:
Maintaining precision and data reliability is crucial. For this project, the models will be trained and evaluated using polysomnography data, recorded at the University of Bergen. This method ensures that the candidate can effectively compare and validate the results from the wearable against the established polysomnography standards, thereby providing a reliable reference for refining the predictive algorithms.
Expected Outcomes:
By the project's conclusion, the candidate aims to present a machine learning model adept at real-time, non-invasive monitoring and classification of sleep stages, with a tailored focus on medical applications, for individuals living with ADHD. This advancement could be pivotal in enhancing treatments, optimizing condition tracking, and ultimately elevating the quality of life for those living with ADHD.