Digital cognitive assessment in preclinical Alzheimer’s disease: Detecting subtle cognitive changes through speech

This thesis aimed to integrate digital cognitive assessment (DCA) and speech analysis to evaluate cognition in the context of preclinical Alzheimer’s disease (AD). To this end, the first part focused on the current state of digital cognitive assessments for use in preclinical AD and mild cognitive impairment (MCI) stages (Chapter 2, 3). The second part focused on language production in the preclinical stage of AD (Chapter 4, 5, 6). The main findings of chapter 2 indicate that thorough validation of DCAs is limited, and highlight the need for more transparent and consistent reporting on practical and psychometric properties. Chapter 3 indicates that older adults with unimpaired cognition, subjective cognitive decline and MCI, and caregivers recognize both the value and challenges of DCAs. Barriers and facilitators towards using DCAs are highly personal, as indicated by factors such as early disease recognition vs. fear of dementia diagnosis, and self-administration at home vs. lack of personal contact. Other important factors included user-friendliness and transparency on data usage, which may minimize potential barriers, such as digital incompetence or privacy concerns. When considering conventional language tests, chapter 4 showed that although performance on conventional verbal fluency tests is not affected in individuals with preclinical AD, they do show amyloid-related decline in semantic fluency over time. Considering a remote speech-based DCA in chapters 5 and 6, results indicated that remote speech-based DCAs hold promise as a feasible and reliable measure of acoustic and linguistic characteristics of language production, while only trends of subtle speech differences between amyloid-positive and -negative groups were observed. Taken together, these findings indicate that DCAs for early AD stages warrant extensive further validation, but that end-users recognize their potential. Speech-based DCAs comprise a feasible and reliable method to measure speech production remotely, and results from this method may point towards very subtle patterns of speech deficits in preclinical AD, although further steps are required to establish optimal, sensitive, valid and clinically meaningful outcome measures.