Parkinson’s disease diagnosis using Convolutional Neural Networks and figure-copying tasks

Marta Vellejo

Friday 14 June 2019
12:00 - 13:00
Room 3.05 , Post Graduate Centre

Abstract

Parkinson’s disease (PD) is a chronic neurodegenerative disorder, which leads to a range of motor and cognitive impairments. An accurate PD diagnosis is a challenging task since its signs and symptoms are very similar to other related diseases such as normal ageing and essential tremor. This work aims to automate the PD diagnosis process by using a Convolutional Neural Network, a type of deep neural network architecture. To differentiate between healthy and PD patients, our approach focuses on discovering deviations in patient’s motor capabilities with the use of drawing tasks. Besides that, since different data representations may capture various aspects of this disease, this work explores which drawing task, cube or pentagon, is more effective in the discrimination process. The goal is to evaluate which data format is more effective in diagnosing PD.

Hosts: Katya Komendantskaya, LAIV group