Quantitative assessment of dysarthric speech in neurological disorders

Dysarthria is a speech motor disorder characterized by muscle weakness and poor
muscle coordination. It is a neurological symptom arising from cerebral
dysfunction at the level of brainstem nuclei, supra nuclear brain dysfunction
or neuromuscular impairment. It occurs in neurological disorders such as
Parkinson*s Disease, Amyotrophic Lateral Sclerosis, Multiple Sclerosis and
Myasthenia Gravis. It can also occur after brain damage due to e.g.
cerebrovascular disease. The speech patterns of dysarthric patients include
constant disturbances in prosody, phonation and articulation, causing symptoms
such as reduced pitch range, loudness invariation and irregular articulation
rate. We hypothesize that cues in dysarthric speech can be used to
automatically and remotely recognize dysarthria enabling a clinical
classification. Moreover, given the association between dysarthria and
dysphagia, characterizing the severity of dysarthria provides an index to
assess the necessity of whether an individual may require medical intervention,
particularly to prevent choking due to dysphagia.

Our primary objective is to identify the disturbances in prosody, phonation and
articulation with which dysarthria can be automatically recognized in speakers
of Dutch, and German. To this end, we will first perform an experiment that
involves the ranking and sorting of validated pathological and control,
previously obtained audio recordings by UMCG neurologists. Results will provide
insight into which strategies are most reliably used by neurologists to
classify dysarthria by type/pathology and will be used in the second, main part
of this study. This will entail a multiple case study with individuals who have
dysarthria as well as healthy controls. This case study, which is the core of
this research, is geared towards the acquisition of phonetic, phonological, and
acoustic data which can be used in refining the method with which the primary
objective can be achieved.

Multiple case study. Participants will perform four speech tasks, which will be
recorded with audio and video recording equipment. These tasks include: (1) a
prosody task which consists of four exercises, including sentence completion,
repetition and the production of negative/affirmative questions and statements;
(2) a picture description task; (3) a short interview and (4) a reading task.
Additional relevant information which may influence the outcome of the
aforementioned tasks, will also be collected about the participant. This
information includes the presence of speech and language disorders, vision or
hearing problems, medication use and therapy, frequency of communication and
reading, educational attainment and the last profession, all of which may have
an influence on speech production. The study is a collaboration between the
University Medical Center Groningen, the Netherlands (leading center), Klinikum
Emden, Germany and INCAS3 in Assen, the Netherlands.

This study does not involve any risks for the participants and the burden is
limited to the time invested in the study. No personal benefit to the
participants is derived from this study, however, participation in this study
may improve assessment of dysarthria in the future.