Multilingual Cognitive Impairment Detection in the Era of Foundation Models

Damar Hoogland, Boshko Koloski, Jaya Caporusso, Tine Kolenik, Senja Pollak, Christina Manouilidou, Matthew Purver


Abstract
We evaluate cognitive impairment (CI) classification from transcripts of speech in English, Slovene, and Korean. We compare zero-shot large language models (LLMs) used as direct classifiers under three input settings—transcript-only, linguistic-features-only, and combined—with supervised tabular approaches trained under a leave-one-out protocol. The tabular models operate on engineered linguistic features, transcript embeddings, and early or late fusion of both modalities. Across languages, zero-shot LLMs provide competitive no-training baselines, but supervised tabular models generally perform better, particularly when engineered linguistic features are included and combined with embeddings. Few-shot experiments focusing on embeddings indicate that the value of limited supervision is language-dependent, with some languages benefiting substantially from additional labelled examples while others remain constrained without richer feature representations. Overall, the results suggest that, in small-data CI detection, structured linguistic signals and simple fusion-based classifiers remain strong and reliable signals.
Anthology ID:
2026.rapid-1.1
Volume:
Proceedings of the Sixth Resources and ProcessIng of linguistic, para-linguistic and extra-linguistic Data from people with various forms of cognitive/psychiatric/developmental impairments in cooperation with the MENTAL.ai consortium
Month:
May
Year:
2026
Address:
Palma, Mallorca, Spain
Editors:
Dimitrios Kokkinakis, Charalambos Themistocleous, Gaël Dias, Kathleen C. Fraser, Fredrik Öhman, Sebastião Pais
Venues:
RaPID | WS
SIG:
Publisher:
European Language Resources Association (ELRA)
Note:
Pages:
1–12
Language:
External URL:
https://lrec.elra.info/lrec2026-ws-rapid6mentalai-01
DOI:
10.63317/487aco6yfwyv
Bibkey:
Cite (ACL):
Damar Hoogland, Boshko Koloski, Jaya Caporusso, Tine Kolenik, Senja Pollak, Christina Manouilidou, and Matthew Purver. 2026. Multilingual Cognitive Impairment Detection in the Era of Foundation Models. In Proceedings of the Sixth Resources and ProcessIng of linguistic, para-linguistic and extra-linguistic Data from people with various forms of cognitive/psychiatric/developmental impairments in cooperation with the MENTAL.ai consortium, pages 1–12, Palma, Mallorca, Spain. European Language Resources Association (ELRA).
Cite (Informal):
Multilingual Cognitive Impairment Detection in the Era of Foundation Models (Hoogland et al., RaPID 2026)
Copy Citation: