Biomarkers.

Area of Science:

• Neuroscience
• Gerontology
• Public Health

Background:

• Dementia diagnosis in Africa is challenged by multilingualism and low literacy.
• Existing cognitive tools lack validation in African populations.
• Semantic verbal fluency (SVF) is adaptable to diverse linguistic and literacy backgrounds.

Purpose of the Study:

• To evaluate a digital, automated semantic verbal fluency (SVF) test for dementia detection in older Kenyan adults.
• To assess the diagnostic value of practice effects in a multilingual SVF task.
• To determine the accuracy of an automated scoring system for SVF in a low-resource setting.

Main Methods:

• Recruited 24 cognitively unimpaired and 24 dementia-diagnosed older Kenyan adults.
• Administered two consecutive 60-second digital SVF tasks (animal naming) with language switching.
• Utilized automated speech analysis (ki:elements' SIGMA) for scoring and feature extraction.

Main Results:

• Significant group differences in word count were observed on both SVF trials, with a larger gap on the second trial.
• Performance change across trials indicated practice effects in controls and reduced improvement in the dementia group.
• Logistic regression demonstrated high discriminative power (ROC-AUC = .85 in trial 1, .89 in trial 2).

Conclusions:

• An automatically scored, multilingual digital SVF test reliably detects dementia in older Kenyan adults.
• Digital SVF is a promising, culturally adaptable tool for rapid cognitive screening in low-resource, multilingual settings.
• This approach can improve dementia diagnosis accessibility and promote inclusive clinical research in Africa.