Abstract
Language is a complex cognitive function which relies upon widely distributed brain networks (including, but not limited to, the left-dominant perisylvian regions). Consequently, even minor brain changes can result in subtle language alterations. A representative example is Dementia syndromes, where language symptoms due to brain atrophy are reported at the earliest stages and in full-blown pathology, playing a crucial role in the diagnosis. AKiD will build on past findings of the project partners and their experience in analyzing pathological and non-pathological speech to investigate the phonetic profile of patients with Dementia. Reviewing the scientific literature, this component of linguistic competence emerges as highly reliable to automatically detect preclinical and clinical cognitive decline due to protein misfolding Ministero dell'Università e della Ricerca MUR - BANDO 2022 PNRR diseases (e.g., Alzheimer's dementia and Frontotemporal dementia). However, the impact of Dementia on speech has been considered only for durational and prosodic cues, without including possible deficits at a segmental level. An innovative experimental method that combines in-depth acoustic and articulatory analysis will be implemented. We will enroll a balanced cohort of adult/elderly subjects living in Emilia Romagna: the sample will include patients aged 65-85 years diagnosed with preclinical and clinical cognitive decline due to dementia, an equal number of healthy aging peers, and a matching group of neurotypical adults. After the neuropsychological assessment, controlled and semi-spontaneous speech tasks will be administered to collect acoustic data and midsagittal ultrasound images of tongue shape and location through Ultra Tongue Imaging. In line with previous literature, acoustic and articulatory cues will be used to highlight the degree of gestural cohesiveness during speech. They will include vowel formants, plosives’ VOT in singleton and geminate contexts, and articulatory degree of anticipation of consonant gesture. The acoustic and articulatory data will be included – with different degrees of accessibility, in line with FAIR data principles – in the AKiD corpus, together with experimental protocols, result tables and graphs. The effectiveness of the project can be measured in terms of the quantity and quality of collected corpora (Milestone2), delivered post-processed data (M3), and published scientific results in high-impact journals (M4). In keeping with the curiosity-driven nature of PRIN 2022 PNRR, AKiD will explore a cutting-edge and mostly uncharted topic. Moreover, the project could pave the way, in the near-future, for the investigation of innovative research subjects, such as the neurobiological underpinnings of phonetic disturbances in Dementia, to date largely unexplored. Given the limited number of similar resources, the creation of a multilevel acoustic and articulatory corpus of Italian speech in Dementia will constitute an international reference point for clinical speech analysis.
Dettagli del progetto
Responsabile scientifico: Gloria Gagliardi
Strutture Unibo coinvolte:
Dipartimento di Filologia Classica e Italianistica
Coordinatore:
ALMA MATER STUDIORUM - Università di Bologna(Italy)
Contributo totale di progetto: Euro (EUR) 242.871,00
Contributo totale Unibo: Euro (EUR) 160.221,00
Durata del progetto in mesi: 24
Data di inizio
30/11/2023
Data di fine:
28/02/2026
