Models and Approaches for Comprehension of Dysarthric Speech Using Natural Language Processing: Systematic Review (Preprint)

Abstract

UNSTRUCTURED In the past decade, Natural Language Processing has attracted significant attention in medical sciences. This includes speech intelligibility and partially speech comprehension, particularly in Dysarthric speech. Dysarthria is characterized by irregularities in the speed, strength, pitch, breath control, range, steadiness, and accuracy of muscle movements required for articulatory aspects of speech production. This study examined the contributions made by other studies involved in Dysarthric speech comprehension; in relation to the modes of meaning extraction used, applied method types, speech representations used, and databases sourced from. This study followed a systematic review approach to review and map 27 related studies. The findings of this study indicated a significant gap in the inclusion of listener and speech independent features, majorly attributed to the non-robust speech representations used. An index was formulated to illustrate the mappings of the reviewed literature and as such, further research is proposed regarding to the formulation of semantic ontologies that will be useful in the inclusion key features of listener and speech independent features for meaning extraction of dysarthric speech.