Lexical features for repetition task: Turkish language perspective

Abstract

AbstractBackgroundPrimary Progressive Aphasia (PPA) is a neurodegenerative syndrome that presents with speech and language impairments and is commonly classified into three variants based on varying symptomatology and pathology: semantic (svPPA), nonfluent/ agrammatic (nfvPPA), and logopenic variants (lvPPA). While more than 85% of the world population don’t speak English language as their first or second language, PPA research has largely focused on English speakers. Thus, we may lack adequate understanding to PPA symptoms of typologically different languages. Turkish language is an agglutinative language that commonly combines multiple morphemes in a word. Therefore, Turkish words are frequently high in syllables, morphemes, and words. Repetition test is a frequently used tool to diagnose lvPPA patients by assessing auditory verbal short‐term memory function. However, the capacity for auditory verbal short‐term memory has been known to be influenced by time and morphological structure, which are generally more complex in Turkish. Our study investigates the lexical features critical for repetition task for Turkish language speakers.MethodIn this study, Turkish‐speaking participants are tasked to repeat six lists of words. List 1 and 2 consist of real and pseudo‐words with one morpheme and one word but increasing in syllables; List 3 and 4 comprised of real and pseudo‐words with increasing number of morphemes and syllables; and List 5 and 6 contained words with increasing morphemes, syllables and words. We then studied the performance of cognitively normal and PPA participants when repeating these lists of words and investigated the effect of number of morphemes, syllables and words.ResultsA total of 15 Turkish‐speaking controls and 12 PPA individuals (7 lvPPA, 1 nfvPPA, 4 svPPA) completed the repetition task. Individuals with svPPA and lvPPA scored significantly lower than controls when repeating List 5 and 6 (p = 0.05 and p = 0.03 respectively). Using general linear model, when covarying for number of words, morphemes, and syllable, we found that repetition performance in Turkish‐speaking participants is predicted by the number of words, morphemes, and syllables in both real and pseudo‐word lists.ConclusionWhen designing repetition task for speakers of agglutinative or polysynthetic languages, lexical features such as number of morphemes and syllables might need to be accounted.