Neurorobotics Model of Parkinson’s Disease

This project will use humanoid robots to gain further insights into the mechanisms of the disease based on the first computational model of Parkinson’s Disease (PD) created from recent gathered data from marmoset monkeys.
Description of the Project: 

This project is part of a wider project and aims to contribute to the development and understanding of a Neurorobotics model of Parkinson’s Disease (PD). Parkinson’s disease (PD) affects more than 3% of people over 65 years old, with figures set to double in the next 15 years. It is a neurodegenerative disease, whose symptoms include cognitive and motor disorders; in late stages, it possibly leads to depression and dementia. In search for a cure, scientists use computational and animal models of the disease. Both approaches are incomplete: the former neglects that PD symptoms emerge out of the animal brain-body-environment interactions; the latter only partially mimics the mechanisms of the disease, since a complete understanding of the biological causes of PD is still missing. A promising alternative is neurorobotics, which is at the intersection of robotics and neuroscience and focus on embedding neurobiological structures underlying animal behaviour in robots. This project will explore modern computational methods, such as machine learning, to implement neurobiological structures on humanoid robots based on current available data gathered from healthy and PD animal models (i.e., marmoset monkeys, considered the biomedical supermodel). The goal is to replicate the animal features of the disease into the neurorobotics model. In this way, the resulting neurorobotics model will facilitate the investigation of the mechanisms of the disease and support the development of techniques that can inform new PD therapies. Importantly, experiments with robots are cheaper and faster than the animal counterpart, thus another important outcome of this project is to contribute to reduce the number of animals used in research. 

Resources required: 
High performance computing equipment, NVIDIA GPUs, iCub robot and NAO robot.
Project number: 
First Supervisor: 
Heriot-Watt University
Second Supervisor(s): 
First supervisor university: 
Heriot-Watt University
Essential skills and knowledge: 
Good programming skills.
Desirable skills and knowledge: 
Programming skills in Python, Matlab, robotics, computer vision, GPU programming, numerical methods, machine learning.

1] Neurology 68(5):384-6, 2007
[2] Ann. N.Y. Acad. Sci. 991:1-14, 2003
[3] J. Theor. Biol. 257:642–63, 2009
[4] Neuron 93(3): 480-90, 2017
[5] Nature 459:483, 2009
[6] PNAS 105:3593-98, 2008
[7] J. Neurosci. Methods 205:283–94, 2012

[8] Front. Comp. Neurosci. 7(39), 2013

[9] J. Comput. Neurosci. 40(2):207-29, 2016

[10] Prog. Brain Res. 165:425-45, 2007
[11] Neuron 84(4):716-22, 2014