Abstract
Sleep is necessary for the proper functioning of our bodies and minds becauseit helps our bodies to recuperate and rebuild while our minds organise and evaluate
memories. Poor sleep, on the other hand, can have a severe impact on both physical
and emotional well-being, leading to a variety of health problems such as anxiety,
depression, heart disease, obesity, dementia, diabetes, and cancer. University students,
in particular, are at risk of poor sleep quality and mental health issues, which can
exacerbate the risk of anxiety and despair. It is critical to treat sleep disorders and
prioritise adequate sleep practises in order to ensure good health and well-being.
Many individuals may be unaware they suffer from conditions like insomnia,
depression, or anxiety, often because they fail to recognize the symptoms. These issues
are typically diagnosed by medical professionals during consultations. While experts
understand the connection between poor sleep patterns and mental health, this
knowledge gap among the general population highlights the potential for technological
intervention. Computer science, particularly through smartphone technology, can
bridge this gap by enabling early detection and intervention, helping to preserve mental
health. This thesis aims to explore and develop methods using current mobile
technology specifically, built-in sensors to diagnose sleep issues and provide
interventions for associated mental health problems, focusing on young adults.
Existing literature on using mobile sensors to track sleep and mental health reveals
several limitations, which this thesis seeks to address. While numerous mobile apps
track behavioural signals, such as sleep, and correlate them with mental health, many
rely on wearable or non-wearable sensors (e.g., accelerometers, microphones, GPS,
light, and screen on/off sensors), but each has limitations in terms of accuracy and
usability.
This thesis centers on the development of SleepTracker, an app designed to go
beyond basic sleep tracking to address mental health concerns like depression and
anxiety often associated with insomnia. The SleepTracker app was developed with the
goal of accurately detecting sleep duration and disturbances using mobile phone
sensors, specifically screen on/off events and accelerometer data. Its development was
II
structured in two distinct phases, each building on the last to progressively enhance
the app’s functionality, data accuracy, and user experience.
In the first phase, initial development efforts focused on refining the app’s core
algorithms through iterative testing. This phase was supported by a Patient and Public
Involvement (PPI) session, where users provided feedback on the app’s feasibility and
potential benefits. Following this, two field tests were conducted to validate the app’s
ability to detect sleep patterns accurately: the first test spanned six nights, while the
second test ran for seven nights. The results from these field tests, along with post-test
survey feedback, informed the subsequent steps in the app’s development.
As the study progressed to the second phase, an enhanced algorithm was
introduced to detect sleep disturbances based on screen on/off activity and
accelerometer data. Prior to launching this phase, an additional PPI session was held
to refine the study’s focus, gathering user insights specifically on the app's aim to
investigate links between sleep disturbances and mental health concerns such as
anxiety and depression. This 56-day phase involved ongoing data collection and
analysis to explore these relationships further, enabling the app to evolve into a tool
capable of identifying potential mental health risks associated with sleep issues.
To ensure data security and privacy throughout both phases, two separate
databases were created: one storing sensitive personal information (e.g., name, email,
gender, ethnicity) and another for sleep and sensor data. In the first phase, Firebase
was used for data storage; however, due to concerns raised by the ethics committee
regarding enhanced data protection, the app transitioned to Amazon Web Services
(AWS) in the second phase. This move, along with the separation of databases, was
critical in maintaining data integrity and anonymity, providing a strong foundation for
ethical compliance and secure data management.
At the conclusion of the study, the SleepTracker app’s acceptability was
evaluated using both quantitative and qualitative analyses. The quantitative analysis
focused on numerical data collected through participant surveys, applying a theoretical
framework based on the Technology Acceptance Model (TAM). The qualitative
analysis used thematic analysis to explore participants’ experiences with the app,
uncovering a range of insights regarding its usability and effectiveness. Together, these
analyses provided a holistic understanding of the SleepTracker app’s feasibility and
its potential for monitoring mental health.
III
In summary, the SleepTracker app shows promise as a tool for monitoring
sleep patterns and identifying potential mental health concerns among young adults.
Its integration of advanced technology and data-driven algorithms offers the potential
to yield significant insights into the intricate relationship between sleep disturbances
and mental health. This research aims to contribute to improved mental well-being and
address the challenges posed by sleep-related issues in future health interventions.
| Date of Award | 18 Mar 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Ian Nabney (Supervisor), Sarah E Bennett (Supervisor) & Esther Crawley (Supervisor) |