Design and implementation of an embedded system for real time analysis of speech from people with Parkinson's disease

Parkinsons disease (PD) is a neurodegenerative disorder that affects the coordination and regulation of muscles and limbs, including those responsible of speech production, resulting alterations in voice intelligibility. It is already demonstrated that the constant speech therapy can improve the communication skills of the patients; however, it is required the development of flexible tools able to assess and guide the speech therapy of the patients without the intervention of the experts. In this paper one embedded system for the real time analysis of speech from people with Parkinson's disease is presented. Two platforms are developed. First, a Matlab Graphical User Interface is presented; second, the application is deployed in a minicomputer that has an audio codec, storage capacity and a efficient processing unit. The device has been equipped with an LCD monitor to display the information on real time, and a mini keyboard for the interface with the user. Different measurements that are commonly used in the assessment of speech from people with PD are evaluated. In agreement with the state of the art, an increment of the pitch variation is showed for patients with PD, additionally, lower values of the vowel space area (VSA) are also shown for speech recordings uttered by people with PD.