IntelliSpeechApp 1.0 (ISA 1.0)
Luigi Marotta, Laura Barbagli, Debora Colangelo, Chiara Falchi, Elisa Matozzi, Chiara Santacroce, Giada Viglietti, Benedetto Del Greco
Communicative effectiveness is based on the principles of quantity, quality, relevance and modality (Grice, 1975). In verbal communication, quality and modality also depend on intelligibility, i.e. the degree to which a speech signal is understood. Intelligibility increases significantly between the ages of four and six years, so monitoring it is crucial for identifying early communication difficulties. However, the tools currently available are often affected by operator error. This study aims to develop an automated instrument that can quickly, objectively and simply measure intelligibility in preschool children using an innovative automatic speech recognition method (Baevski et al., 2020). 178 typically developing children aged between 4.1 and 6.1 years from different Italian regions took part in the study. Each child was given a battery of standardised tests (BVL 4_12; Marini et al., 2015) and an app (B-VI Test), which was developed specifically to measure intelligibility. Word and non-word repetition tasks, selected based on frequency, length, and articulatory complexity, were then incorporated into the app. The examiners worked in blinded mode, unaware of the results obtained by the children in the two different modes of analysis. Judgments of intelligibility were compared. The ‘BVI-Test’ app, by allowing an organised and systematic recording of evidence, will be able to support the assessment of language impairment and its impact on the child’s social functioning in clinical and forensic settings.
Keywords
Test, Articulation, Language models, App.
