Development of Markerless Systems for Automatic Analysis of Movements and Facial Expressions: Applications in Neurophysiology

This project is focused on the development of markerless methods for studying facial expressions and movements in neurology, focusing on Parkinson’s disease (PD) and disorders of consciousness (DOC). PD is a neurodegenerative illness that affects around 2% of the population over 65 years old. Impairments of voice/speech are among the main signs of PD. This set of impairments is called hypokinetic dysarthria, because of the reduced range of movements involved in speech. This reduction can be visible also in other facial muscles, leading to a hypomimia. Despite the high percentage of patients that suffer from dysarthria and hypomimia, only a few of them undergo speech therapy with the aim to improve the dynamic of articulatory/facial movements. The main reason is the lack of low cost methodologies that could be implemented at home. DOC after coma are Vegetative State (VS), characterized by the absence of self-awareness and awareness of the environment, and Minimally Conscious State (MCS), in which certain behaviors are sufficiently reproducible to be distinguished from reflex responses. The differential diagnosis between VS and MCS can be hard and prone to a high rate of misdiagnosis (~40%). This differential diagnosis is mainly based on neuro-behavioral scales. A key role to plan the rehabilitation in DOC patients is played by the first diagnosis after coma. In fact, MCS patients are more prone to a consciousness recovery than VS patients. Concerning PD the aim is the development of contactless systems that could be used to study symptoms related to speech and facial movements/expressions. The methods proposed here, based on acoustical analysis and video processing techniques could support patients during speech therapy also at home. Concerning DOC patients the project is focused on the assessment of reflex and cognitive responses to standardized stimuli. This would allow objectifying the perceptual analysis performed by clinicians.