AutoMAP – Smart Wearables for Depression: Toward Autonomous Mental Health Monitoring for the Elderly
- Publication Type:
- Thesis
- Issue Date:
- 2022
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
In the first year of the pandemic, anxiety, and depression increased by 25% in all populations. Many older-aged people believe that mental illness is not something that can be treated and is often stigmatized. More so, older-aged people often have an increased risk of other mental issues, and misdiagnosis as general age-related deterioration such as general cognitive impairment, Alzheimer’s disease, or dementia. This increased dependence onothers and loss of general socio-economic status increases the likelihood of depression. In older aged people, mental illness and other age-related conditions inevitably lead to caregiver burden and dependence. One approach to aid in potentially alleviating caregiver burden while allowing older-aged people to live alone is the use of consumer-grade smart wearables to monitor depressive tendencies. Increasingly, studies are being conducted on emotion recognition using contemporary technologies, particularly smart wearables. There is however a lack of focus on older-aged people (65+ years) as the target demographic. When this cohort has been investigated, study materials consist of obtrusive on-body sensors or are administered in lab-based settings with external emotion stimulation. As such, these are inapplicable or less accurate in real-world settings. This thesis aims to address the above issues by proposing AutoMAP - a framework that incorporates consumer-grade smart wearables for implicit user input and machine learning to infer depressive tendencies, a framework was proposed based on a comprehensive literature review in emotion recognition and machine learning. This review was conducted to develop a clearer overview of mental health issues around the world, particularly in older-aged people, and the current state of research. Through an in-depth analysis of existing research, strengths and limitations of existing approaches were highlighted. Assessment of these strengths and limitations allowed for the refinement of the current research project, aiming to reduce some of the identified limitations. A conceptual framework that provides autonomous mental health monitoring for older aged people was developed. An experimental study was performed with mixed-design methods to validate the proposed framework. Lastly, data outputs from the experimental study were analysed to validate the feasibility of the framework. Activity data were used to train and evaluate machine learning models to assess their predictive performance toward depressive tendency. Based on the above, a platform was created that consisted of (1) implicit user input, (2) machine learning, and (3) emotion reporting infrastructure. Various models will be explored for predictive modeling on device data.
Please use this identifier to cite or link to this item: