Language Modelling for the Clinical Semantic Verbal Fluency Task

Nicklas Linz (1) , Johannes Tröger (1) , Hali Lindsay , Alexandra Konig (2, 3) , Philippe Robert (4) , Jessica Peter , Jan Alexandersson (1)

1 DFKI - Deutsches Forschungszentrum für Künstliche Intelligenz GmbH = German Research Center for Artificial Intelligence
2 CobTek - Cognition Behaviour Technology
3 STARS - Spatio-Temporal Activity Recognition Systems
4 Technicolor Imaging Science Lab

Semantic Verbal Fluency (SVF) tests are common neuropsychological tasks, in which patients are asked to name as many words belonging to a semantic category as they can in 60 seconds. These tests are sensitive to even early forms of dementia caused by e.g. Alzheimer's disease. Performance is usually measured as the total number of correct responses. Clinical research has shown that not only the raw count, but also production strategy is a relevant clinical marker. We employed language modelling (LM) as a natural technique to model production in this task. Comparing different LMs, we show that perplexity of a persons SVF production predicts dementia well (F1 = 0.83). Demented patients show significantly lower perplexity, thus are more predictable. Persons in advanced stages of de-mentia differ in predictability of word choice and production strategy-people in early stages only in predictability of production strategy.

Mots clés: Dementia; Alzheimer's Disease; Semantic Verbal Fluency; Language Modelling; Machine Learning

Domaines: Sciences cognitives; Linguistique; Neurosciences; Psychologie

Nicklas Linz, Johannes Tröger, Hali Lindsay, Alexandra Konig, Philippe Robert, et al.. Language Modelling for the Clinical Semantic Verbal Fluency Task. LREC 2018 Workshop RaPID-2: Resources and ProcessIng of Linguistic, Para-Linguistic and Extra-Linguistic Data from People with Various Forms of Cognitive/Psychiatric Impairments, May 2018, Miyazaki, Japan. ⟨hal-01851411⟩