Abstract
A Rapid increase of older persons and improving public health system may affect life expectancy as well as a higher number of older adults with dementia. Nonetheless, a promising approach is to adopt a strategy of early detection and intervention for providing treatments or delaying the symptoms. By considering speech and language deficits has been evident of early decline of the brain in relative to other functions, the current study aimed to develop a tele-neuropsychological test battery using speech and language-based parameters and validate it in 403 community-dwelling older adults. All participants were divided into 252 healthy controls (HC) and 151 older adults with mild cognitive impairment (MCI). Also, all participants could be divided into two phases, that is, lab- and community-based testing. The lab-based testing composed of 300 community-dwelling older adults (i.e., 196 HCs and 104 MCIs) and the community-based testing consisted of 103 community-dwelling older adults (i.e., 56 HCs and 47 MCIs). Standardized neuropsychological tests were used to classify older adults with MCI. The given tests composed of the subjective memory complaint questionnaire, the Mattis dementia rating scale-2, the Wechsler’s memory scale –Logical memory I & II, the Montreal cognitive assessment, the activities of daily living and the Instrumental activities of daily living checklists. The tele-neuropsychological test battery consists of 10 subtests and a voice recorder for phase 1 and smartphones/tablet PCs for phase 2. The main findings revealed that the developed test battery produced good-to-very good reliability, small-to-moderate criterion validity, and acceptable predictive validity by using machine learning algorithms for both phases. Furthermore, language-based parameters outperformed speech-based parameters across settings. Whilst the 1st phase demonstrated a better classification than those of the 2nd phase. Thus, the automated scoring program was developed based on the unique findings from both phases.
