Approaching multimodal Deep Learning diagnosis of standard clinical assessments in Parkinson’s Disease

Parkinson’s Disease (PD) is a neurodegenerative disease of high incidence in the ageing population. This project aims at the application of deep learning technologies to a clinical dataset that contains information of patients with prodromal or early-stage PD. By analysing and processing digitalised movement data, captured by three standard clinical assessments, the classifier will be expected to characterise bradykinesia, a slowing of movement, which is the fundamental motor feature of PD. The complex nature of bradykinesia makes it difficult to reliably identify it, particularly at the early stages of the disease (Ahlrichs and Lawo, 2013).

The three clinical assessments used in this study are the following:
1. Finger tapping
2. Hand pronation-supination
3. Hand opening-closing

Previous research done in this area was conducted using a type of evolutionary algorithm called Cartesian Genetic Programming (Muhamed et al., 2018), achieving an effective classification performance capable of reaching an accuracy of 84%. The idea is to create a system that substitutes or complements these previous technologies using deep learning techniques, by means of accomplishing a better multimodal accuracy characterisation of early-stage PD with the combination of the data from these three clinical assessments.

References<\b>:
Muhamed, S.A., Newby, R., Smith, S.L., Alty, J.E., Jamieson, S. and Kempster, P., 2018. Objective Evaluation of Bradykinesia in Parkinson's Disease using Evolutionary Algorithms. In BIOSIGNALS (pp. 63-69).
Ahlrichs, C. and Lawo, M., 2013. Parkinson's disease motor symptoms in machine learning: A review. arXiv preprint arXiv:1312.3825.

Supervisor name: 
Marta Vallejo
Supervisor and Deputy email addresses: 
m.vallejo@hw.ac.uk