Augmented reality to detect subtle alterations in the early stages of Alzheimer’s disease: Preliminary results of the RADAR‐AD study

Abstract

Background The ‘Remote Assessment of Disease and Relapse – Alzheimer’s Disease’ (RADAR‐AD) study is assessing functional decline in Alzheimer’s disease (AD) using remote monitoring techniques (RMT’s). Compared to traditional pen‐and‐paper clinical assessments, RMT’s can continuously and objectively monitor function during activities of daily living (ADL), which are arguably more sensitive to the earliest stages of AD. The aim of this abstract is to compare the results of the augmented reality task ‘Altoida’, that recreates an ADL requiring spatial navigation and memory, implemented as a tablet application, between 1) healthy controls, preclinical AD and prodromal AD, and with 2) standard clinical tests for cognitive and functional decline. Method We included amyloid negative cognitively normal (healthy controls, n=10), amyloid positive cognitively normal (preclinical AD, n=7) and amyloid positive mild cognitive impaired (prodromal AD, n=4) participants (Table 1) from the RADAR‐AD study. The outcome of the Altoida test, consisting of a motor task and two tasks in which participants have to hide‐and‐seek virtual objects, is the validated Neuromotor Index (NMI), with higher scores reflecting normative scoring, according to age, sex and education. Cognition was measured using a word‐list‐learning test, digit symbol substitution test (DSST), Rey complex figure, verbal fluency and Boston naming test. Functional decline was assessed using the Amsterdam Instrumented Activities of Daily Living (AIADL) questionnaire. Result In our preliminary sub‐sample, healthy controls showed higher NMI scores compared to the preclinical AD and prodromal AD participants (p=0.02) (Figure 1). The NMI was related to the DSST only (Figure 2). Conclusion NMI scores differed between cognitively normal healthy controls and cognitively normal preclinical AD participants, while no differences could be found in cognitive and functional tests between these groups. The sample size will increase in the coming months, but despite the currently small sample, the preliminary results are promising in evidencing that digital biomarkers are potentially more sensitive than standard clinical tests in detecting the early stages of AD, which could be helpful in developing new endpoints in clinical trials. This work has received support from the EU/EFPIA Innovative Medicines Initiative Joint Undertaking RADAR‐AD (grant No 806999) and their associated partners.

Artificial intelligence for dementia—Applied models and digital health

“…Measuring patterns such as day‐to‐day changes and mean bout length, or variability of physical activity, allows researchers to predict changes in habitual routines associated with dementia progression 46 . In addition to monitoring disease progression, digital markers such as gait and physical activity are likely to be key for predicting important post‐diagnostic outcomes, such as the risk of falls, 65 and functional decline 66 . There is additional scope for the role of physical activity (or general exposure) in green spaces 67 …”

“…46 In addition to monitoring disease progression, digital markers such as gait and physical activity are likely to be key for predicting important post-diagnostic outcomes, such as the risk of falls, 65 and functional decline. 66 There is additional scope for the role of physical activity (or general exposure) in green spaces. 67 Sleep disturbances can become increasingly common during the course of dementia; sleep markers can reflect this decline and allow clinicians to adapt care provision for an individual's need.…”

“… 46 In addition to monitoring disease progression, digital markers such as gait and physical activity are likely to be key for predicting important post‐diagnostic outcomes, such as the risk of falls, 65 and functional decline. 66 There is additional scope for the role of physical activity (or general exposure) in green spaces. 67 …”