Analysing data from patient's with Parkinson's disease.

The research will be in exploring and developing novel statistical techniques to analyse data from patients with Parkinson's disease, using data provided by the "What's driving Parkinson's Disease?" research group. The first stage of research is in the quantification of tremor, as this is a key feature of Parkinson's disease, and understanding tremor may help develop understanding of the disease from a longitudinal perspective whilst also could be a path towards earlier diagnosis. This work draws on topics from signal processing and time series analysis, comparing a selection of published methods that convert an acceleration signal to a displacement signal, conducting a controlled experiment to determine which of these present the best results for our data, modification of the method to fit our exact specification, and implementing this to obtain displacement. Further to this, I will be assessing and developing methods of extracting key data points from images and videos of patients allowing for objective quantification of key symptoms of Parkinson's disease. Following this, other sources of data will be analysed and used in conjunction the aforementioned data, and analysed with the objective of making inference about the data sources. Ultimately, I would aim to assess the effectiveness of the current UPDRS method of diagnosing Parkinson's, as this process not only permits the practitioner's subjectivity, but might also ignore subtle indicators of Parkinson's disease.