Using a Digital Neuro Signature to measure longitudinal individual-level change in Alzheimer’s disease: the Altoida large cohort study

Abstract

Conventional neuropsychological assessments for Alzheimer’s disease are burdensome and inaccurate at detecting mild cognitive impairment and predicting Alzheimer’s disease risk. Altoida’s Digital Neuro Signature (DNS), a longitudinal cognitive test consisting of two active digital biomarker metrics, alleviates these limitations. By comparison to conventional neuropsychological assessments, DNS results in faster evaluations (10 min vs 45–120 min), and generates higher test-retest in intraindividual assessment, as well as higher accuracy at detecting abnormal cognition. This study comparatively evaluates the performance of Altoida’s DNS and conventional neuropsychological assessments in intraindividual assessments of cognition and function by means of two semi-naturalistic observational experiments with 525 participants in laboratory and clinical settings. The results show that DNS is consistently more sensitive than conventional neuropsychological assessments at capturing longitudinal individual-level change, both with respect to intraindividual variability and dispersion (intraindividual variability across multiple tests), across three participant groups: healthy controls, mild cognitive impairment, and Alzheimer’s disease. Dispersion differences between DNS and conventional neuropsychological assessments were more pronounced with more advanced disease stages, and DNS-intraindividual variability was able to predict conversion from mild cognitive impairment to Alzheimer’s disease. These findings are instrumental for patient monitoring and management, remote clinical trial assessment, and timely interventions, and will hopefully contribute to a better understanding of Alzheimer’s disease.

Automated Early Detection of Alzheimer’s Disease by Capturing Impairments in Multiple Cognitive Domains with Multiple Drawing Tasks

“…Our results also support the viability of automated screening for AD in non-specialist settings. Previous studies have proposed automated screening tools [72], including both digital versions of conventional tests [15,18,73] and novel digital tests [74][75][76], and acceptance of those tools has been reported [44]. Our findings may improve the reliability of those tools by facilitating improved accuracy and a greater potential interpretability.…”

Associations of digital neuro-signatures with molecular and neuroimaging measures of brain resilience: The altoida large cohort study

“…Altoida has demonstrated excellent intra-individual longitudinal changes in cognition, it can also be used as an instrument that predicts and detects transitions in cognitive status. Specifically, a DHT, like Altoida, can be used to detect whether someone will transition from normal cognitive function to mild cognitive impairment and from mild cognitive impairment to dementia/AD ( 14 ).…”

“…However, when focusing on brain resilience being linked to patterns of digital monitoring biomarkers, it is crucial to accurately define both brain health and its determinants through a dynamic trajectory model incorporating antecedent risk factors and also investigate a complex (composite) digital biomarker as a proxy for brain health outcomes ( 13 ). With such requirements in mind, two avenues for the validation of digital monitoring biomarkers for brain resilience can be explored: (a) creating a digital biomarker platform based on “digital footprints” for brain function, physical function, social function, protective or risk factors, such as systemic vascular risk and mental health, and/or (b) identifying a unique complex Digital Neuro Signature ( 14 ) (DNS)™ that exists as a proxy for brain resilience and disease prevention, such as LOAD ( 15 ). To satisfy the first option, a digital brain health platform can be created to assess overall physical health, nutrition, sleep, physical activity, cognitive activity, socialization, and diet recorded via smartphones, wearables or other sources of the Internet of Things (IoT) and on the other side inferred casualty with various biological variables ( 16 ).…”

“…The ability to measure off-target effects on molecular signatures such as brain metabolism will increasingly come into play, after a recent draft guidance document from the FDA to provide recommendations to sponsors, investigators, and other stakeholders on the use of digital health technologies (DHTs) to acquire data remotely from participants in clinical investigations evaluating medical products ( 19 ). Furthermore, to capture the full picture of ongoing brain processes, we describe the longitudinal validation of this monitoring biomarker, based on continuous dynamical interpretation of neuroimaging measurements from the Altoida large cohort study ( 14 ). This would enable meaningful monitoring of cognitive change, to collect more precise predictions of therapeutic response through personalized measurements of the pharmacodynamic drug effects coupled with automated date/time stamps.…”

Validation of a remote and fully automated story recall task to assess for early cognitive impairment in older adults: a longitudinal case-control observational study

“…More importantly, test performance is measurably influenced by a range of state factors, such as sleep (4), exercise (5), mood (6) and stress (7). This variation can give the inaccurate impression of improvement or decline over time (8). Higher frequency sampling can generate more stable and reliable estimates of constructs of interest by controlling for state effects (9) and delineating short-term cognitive fluctuations from longer term changes associated with treatment response and disease progression (8).…”