Poster Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions

LiMoN-2 : Bridging languages and biomarkers: harmonizing multilingual speech and EEG markers of neurodegeneration across Switzerland and Peru

Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
This poster is part of the Sandbox Series.

Alessa Hausmann¹, Claudia Rivera-Fernandez^2,3, Agustina Birba⁵, Marcio Soto-Añari⁴, Adolfo M. García^5,6,7, Valentina Borghesani¹; ¹University of Geneva, ²Universidad Tecnológica del Peru, ³Universidad Complutense de Madrid, ⁴Universidad Catolica San Pablo, ⁵Universidad de San Andrés, ⁶Global Brain Health Institute, University of California San Francisco, USA; and Trinity College Dublin, Ireland, ⁷Facultad de Humanidades

Automated speech and language analysis (ASLA) is gaining momentum in the quest for scalable markers of Alzheimer’s disease (AD) and behavioral variant frontotemporal dementia (bvFTD). Yet, growing evidence is strongly biased toward monolingual English-speaking populations (although approximately one in every two patients is likely to be bilingual) and devoid of brain-behavior mappings (thereby limiting biological validity). This is particularly problematic in contexts where patients are assessed in a second language (L2), especially if it proves underrepresented and structurally different from English, let alone the massive inter-individual heterogeneity in L2 age of acquisition, proficiency level, degree of exposure, and contexts of use. The challenge is even greater in settings where MRI and fMRI are unavailable or unaffordable. New ASLA frameworks are thus needed to target multilingual populations with scalable neuroscientific methods. The LIMON-2 project (Linguistic Markers of Neurodegeneration in Bilinguals) aims to establish a harmonized, cross-culturally valid, EEG-supported framework for detecting neurodegenerative changes in bilingual populations from Switzerland and Peru. Our project combines expertise in neurolinguistics, neuropsychology, neuroimaging, and computational neuroscience through an international collaboration involving multilingual cohorts across Peru and Switzerland. The Peruvian cohort evaluates Quechua–Spanish bilinguals, while the Swiss cohort includes speakers of French, German, Swiss German, Catalan, Italian, English, Portuguese, and Spanish. Strategically, we leverage the Toolkit to Examine Lifelike Language (TELL), a web-based, multi-language ASLA platform, and EEG, a portable, low-cost neuroscientific tool capturing syndrome-specific anomalies in dementia. Our first aim is to identify ASLA markers of neurodegeneration in multilinguals with dementia. Across sites, using TELL, participants complete harmonized speech tasks sampling different levels of language production, including verbal fluency, picture description, spontaneous autobiographical narratives, articulation tasks, and video narration. Current efforts are focused on protocol translation, cultural adaptation, and harmonization across languages. Importantly, we are currently collecting connected speech and audio recordings in Quechua from healthy adults aged 30–80 years through a Peruvian initiative supported by national funding. This effort aims to create one of the first large-scale Quechua speech datasets for the study of preclinical cognitive decline and will directly contribute to LIMON-2’s objective of developing inclusive and globally relevant speech biomarkers. Our second aim is to characterize neural markers of neurodegeneration using multimodal EEG approaches. In addition to resting-state EEG, we are currently refining and harmonizing a Fast Periodic Visual Stimulation (FPVS) paradigm across sites. FPVS provides a rapid, objective, and language-independant measure of neural processing with minimal task demands, making it particularly suitable for multilingual and clinically vulnerable populations. Ongoing work focuses on harmonizing acquisition parameters, preprocessing pipelines, and culture-independent stimulus selection to ensure comparability across countries and languages. Resting-state EEG data will also allow us to capture functional connectivity signatures of the most sensitive ASLA features. As a Sandbox submission, this poster will focus on the rationale, methodological framework, and harmonization pipeline underlying LIMON-2. We will present pilot work, protocol development, multilingual adaptations, and discuss how future linguistic and EEG findings may contribute to scalable, language-inclusive, and culturally equitable approaches for the early detection of neurodegenerative diseases.

Topic Areas: Multilingualism, Development of Resources, Software, Educational Materials, etc.