Innovative data mining using deep learning advanced photoplethysmography pulse wave analysis for the assessment of endothelial dysfunction in systemic sclerosis and related cardiovascular conditions.
This project focuses on exploring the potential of simple, low-cost and portable Photoplethysmography (PPG) signals in the diagnosis of Systemic Sclerosis. It aims to use signal processing knowledge to help find PPG based biomarkers of this disease. This could in future lead to the development of a cost-effective portable device to aid the doctors in diagnosing Systemic Sclerosis.
When the normal endothelial function is impaired it is called Endothelial Dysfunction (ED). ED is believed to cause several diseases like diabetes, insulin resistance, chronic kidney failure, tumor growth, stroke, heart (cardiovascular) disease, severe viral infections and connective tissue diseases like Systemic Sclerosis (SSc) and Raynaud’s phenomenon (RP).
In this study, a simpler method of identifying ED using Photoplethysmography (PPG) technique will be evaluated. PPG is a low cost method in which data readings can easily be collected from the fingers, toes and ears by placing simple electrodes. In this study data will be analysed using advanced data mining techniques specially Deep learning (DL). DL has already produced great results in the field of image, speech and language processing and is widely being used by companies like Google, Apple etc. to do advance computing. In this study the benefits of DL will be applied to the medical PPG signals in order to extract information that can link Endothelial Dysfunction (ED) to specific diseases. An attempt will be made to identify Systemic Sclerosis (SSc), Raynaud’s phenomenon (RP), and cardiovascular disease.
The information obtained from the advanced computing techniques in this research, could then help the doctors in better diagnosis of patients. Identifying diseases in early stage can help in preventing the disease or better treatment of patients. Since PPG is a portable technology, so in future easy to carry medical device can be developed which can benefit both the patients and the doctors.
- Review of the literature.
- Study state-of-the-art pulse wave analysis and deep learning data mining techniques
- Develop pulse wave analysis algorithms and deep learning data mining techniques to establish novel ED biomarkers
- Key statistical analyses
- PhD Thesis, and prepare papers for publication.