Qualitative and quantitative ultrasound analysis of alveolar liquids production in brazilian portuguese children speakers

Author(s)

Cássio Eduardo Esperandino

• Document type: Master's Dissertation
• Press: Marília. 2020-06-01.
• Institution: Universidade Estadual Paulista (Unesp). Faculdade de Filosofia e Ciências. Marília
• Defense date: 2020-05-11
• Advisor: Larissa Cristina Berti

Abstract

Introduction: Although liquid consonants have been described as complex segments involved in the concomitant production of double articulatory gestures, tip and dorsum, and are commonly involved in phonological simplification processes, there are opposite arguments to this respect (Recasens, 2016), without a consensus in the literature. Objective: to describe qualitatively and quantitatively the typical production and simplification processes (substitution of non-lateral to lateral liquid, as well as substitution of non-lateral liquid by a semivowel) of alveolar liquids consonants from Brazilian Portuguese in children aged 4 to 9 years, based on ultrasound analysis. Based on evidence from previous studies, the hypotheses defined are: children’s typical productions would resemble adults’ productions, characterized by the presence of double gestures; the atypical productions, such as those involving the substitution processes of non-lateral liquids for a lateral one (/ɾ/ [l]) and semivocalization (/ɾ/[j]) are gestural changes of magnitude reduction and/or omission of one of the gestures and a gestural fusion, respectively. Methods: Sixty productions (US images and audios) from 10 children with typical production and 20 children with speech sound disorders (SSD) involving the aforementioned processes were selected from a database. Among the typical productions, 10 corresponded to the target production of [l] and 10 corresponded to the target production of [ɾ]. Among the atypical productions, 20 were related to the target productions of [l], 10 substitution errors of /ɾ/[l] and 10 substitution errors of /ɾ/[j]. Using the Sound Forge Pro 10 software, the frame related to the point of maximum tongue constriction of all samples was selected from the ultrasound images. Then, the images were arranged in pairs (ex: image of the point of maximum constriction for [l] (target) vs image of the point of maximum constriction for [ɾ] target and/or simplified). The 30 pairs of images were judged by 20 judges, using Fleiss' Kappa Coefficient to analyze the agreement between the judges. In the quantitative analysis, the parameters considered for analysis were: average height of the anterior portion (AP) of the tongue, average height of posterior portion (PP) of the tongue, ratio between anterior and posterior portion and normalized ratio between anterior and posterior portion of the tongue. Results: The typical productions were described as having double gestures (tip and dorsum) for both phonemes, with a higher tip and a more anterior dorsum to [ɾ] compared to [l]. In productions with substitution of /ɾ/[l], there was a very close pattern to typical productions for both phonemes, but with subtle magnitude differences in the tip elevation. For the substitution of /ɾ/ [j], there was only an elevation of medial portion of the tongue (body of the tongue), without tip elevation (gestural fusion), confirming all hypotheses. In the quantitative analysis, it was observed that the typical productions differ from the parameter of ratio between anterior portion (AP) and posterior portion (PP); productions with semivocalization differ by height of PP, ratio between AP and PP and normalized ratio between AP and PP; and productions with substitution of /ɾ/[l] did not show significance in the adopted parameters. Conclusion: The alveolar liquid consonants of BP are complex segments, since two simultaneous gestures were observed. The posterior gesture of the tongue always was present in the productions.. The tongue ultrasound was sensitive to characterize the differences qualitatively (for all productions) and quantitatively (for typical productions and with semivocalization). It is suggested to combine the US analysis in the characterization of errors involving liquids in children with speech sound disorders. (AU)

• FAPESP's process: 18/01519-3 - Gradient productions of alveolar liquids: perceptual-auditory and ultrasonographic analysis
• Grantee: Cássio Eduardo Esperandino da Silva
• Support Opportunities: Scholarships in Brazil - Master