The diagnostic utility of picture description task in english, chinese and italian speaking individuals with primary progressive aphasia

Abstract Background Primary progressive aphasia (PPA) is a language‐predominant dementia syndrome (Mesulam et al.., 201) that impacts speakers of all languages. Despite there are more than 7,000 living languages, the diagnostic model is mainly tailored for English‐language speakers. Cross‐linguistic comparisons are called for, especially considering the heterogeneities of linguistic features in the different languages (Blasi et al., 2022). We intend to study the speech and language markers of English‐, Mandarin‐, Cantonese‐ and Italian‐speaking individuals with PPA when completing verbal picture description test. The goal is to (1) characterize the speech profiles of semantic(sv), logopenic(lv), nonfluent/agrammatic(nfv) PPA for each specific language, identifying key linguistic markers for the diagnosis; (2) identify similarities/differences in terms of markers across languages. Method We collected the oral picture description, based on the PicNic Scene from WAB battery (Schewan & Kertesz, 1980), of 30 English‐, 30 Chinese‐ (Mandarin and Cantonese), and 30 Italian‐ speaking PPA patients (equally distributed across the three variants) with 45 matched healthy controls (HC). Collected speech samples underwent CLAN and Whisper software analysis to characterize informativeness, lexico‐semantic, morpho‐syntactic, phonological and prosody features. We standardized PPA data relatively to their respective. Differences between PPA variants and with HC for the different features were only carried out for the Italian group, using non parametric tests. Performances across different language groups will be investigated as well. Result Analysis on Italian sample revealed that mean length of utterances, number of empty, filled pauses, abandoned words, and words repetition significantly distinguished PPA from HC. svPPA produced more irrelevant words and adverbs than other PPA groups. nfvPPA produced more sentences without verbs compared to HC and PPA. lvPPA produced more abandoned words than other PPA. svPPA and lvPPA produced less nouns and noun‐verbs ratio compared to HC and nfvPPA, meanwhile lvPPA and nfvPPA produced less words, open class words and more determiners‐noun ratio compared to HC. svPPA and nfvPPA showed reduced idea density compared to HC. Conclusion Those results are consistent with previous literature on PPA oral production. However, future comparisons across languages are extremely important and needed to ensure equitable assessment and minimize disparities globally.