TL;DR:
- Altoida’s RADAR-AD study reveals AR and ML-based digital biomarkers’ potential in early Alzheimer’s detection.
- The study involves 121 participants over 50 years old, using remote monitoring technologies (RMTs).
- Altoida’s AR app distinguishes healthy controls from preclinical and prodromal AD with high accuracy.
- No learning effects were observed, indicating the app’s reliability.
- Augmented reality-based assessments offer promise for early AD diagnosis.
- Altoida aims to support AD treatment development through swift participant identification and continuous monitoring.
Main AI News:
In a groundbreaking development, Altoida, a trailblazer in the realm of digital biomarkers for neurological conditions using augmented reality (AR) and machine learning (ML), has unveiled the results of the RADAR-AD consortium study, featured in Nature Digital Medicine. This study sheds light on the potential of Altoida’s AR and ML-based digital cognitive assessment in the early detection of Alzheimer’s disease (AD).
The ‘Remote Assessment of Disease and Relapse – Alzheimer’s Disease’ (RADAR-AD) study, an independent validation endeavor encompassing 121 participants over the age of 50, delved into cognitive and functional decline in AD by harnessing various remote monitoring technologies (RMTs), including the Altoida AR digital cognitive and functional assessment, known as the “Altoida AR app.” Diverging from conventional pen-and-paper clinical assessments, RMTs are engineered to identify early impairments through frequent and objective monitoring of functional performance during tasks associated with instrumental activities of daily living (IADL), all without the need for clinician intervention.
The Altoida AR app was administered in the study to three distinct groups: amyloid beta-negative healthy control participants (HC, N=57), amyloid beta-positive cognitively normal preclinical AD participants (preAD, N=27), and prodromal AD participants (proAD, N=37). This investigational test comprises both motor and AR tasks. The motor tasks encompass a series of brief exercises assessing fine motor skills and reaction times, allowing for the establishment of personalized reference values against population statistics for motor skills, visual capabilities, and reaction times. On the other hand, the AR tasks aim to replicate a complex IADL-like activity referred to as a “place-and-find” task, where participants engage in a virtual hide-and-seek with virtual objects following specific instructions. The Altoida AR app’s output is generated by an ML model, fine-tuned to differentiate between cognitively normal and impaired participants. Leveraging data from internal device sensors, the Altoida AR app identifies digital biomarkers, meticulously trained from cohort data. The study also involved the administration of the Altoida AR app at home, on personal devices, by participants themselves, for a period of up to 8 weeks, in addition to in-clinic assessments alongside a standard neuropsychological assessment battery.
The results of this study unveiled the Altoida AR app’s remarkable ability to distinguish healthy controls from individuals with preclinical AD and prodromal AD, both within the clinic and through at-home tests. Notably, it showcased a significant breakthrough: the Altoida AR app achieved a level of performance in preclinical Alzheimer’s disease that surpasses the capabilities of standard cognitive tests. Furthermore, the study revealed that there were no learning effects associated with the Altoida AR app’s usage.
Professor Dag Aarsland, Chair of Old Age Psychiatry at King’s College London and academic leader of RADAR-AD, expressed his optimism, stating, “There is immense potential in digital devices and sensor technologies for objective and continuous monitoring of Alzheimer’s disease symptoms. Altoida’s AR cognitive assessment appears to be one of the most promising emerging technologies in this field, and the RADAR-AD study results indicate that Altoida’s test can detect subtle cognitive and functional changes in individuals with AD even before they exhibit manifest deficits. With the advent of new disease-modifying Alzheimer’s drugs, early diagnosis is now more crucial than ever.“
Marc Jones, CEO of Altoida, emphasized the company’s commitment to innovation and its mission to create responsible, evidence-based digital assessments for the early identification of neurological diseases, including AD. “Our feature in Nature Digital Medicine represents a pivotal milestone in our journey,” he stated. “Altoida’s primary objective is to facilitate the swift identification of suitable study participants for AD treatment developers, continuously monitor their cognitive and functional responses throughout the study, and collect pertinent data to support regulatory filings, both for our partners and ourselves. We remain dedicated to ongoing innovation and enhancement of every facet of the Altoida AR app to achieve this crucial goal.”
Augmented reality-based assessments, designed to simulate IADL, offer promising avenues for measuring cognition and function essential for IADL in early AD, both within clinical settings and the comfort of one’s home. RMTs, encompassing smartphone apps and smartwatches, are revolutionizing the assessment of functional and cognitive performance in AD. Their sensitivity, objectivity, and capability for long-term high-frequency measurements hold the potential to detect subtle cognitive decline occurring in the earliest stages of AD.
RADAR-AD (Remote Assessment of Disease And Relapse – Alzheimer’s Disease) represents a European initiative funded by the Innovative Medicines Initiative (IMI). The project encompasses two pivotal areas in AD research, namely, functioning and technology. “Functioning” refers to the daily activities individuals engage in, and RADAR-AD explored the impact of AD on these activities, as well as how individuals’ functioning may evolve. The project also investigated the utility of widely-used technologies, such as smartphones, smart wristbands/fitness trackers, and home-based sensors, in measuring changes in functioning.
Altoida’s investigational algorithm, the digital neuro signature (DNS) for mild cognitive impairment, amalgamates motor and AR tasks. The digital score, ranging from 0 to 100, with higher scores indicating better performance and lower scores suggesting a higher likelihood of cognitive impairment, is rooted in a machine learning model. This model leverages data from the touch screen, accelerometer, and gyroscope, along with ground truth data from clinically validated cohorts, to distinguish between cognitively normal and impaired participants.
Conclusion:
Altoida’s innovative AR and ML-based approach to Alzheimer’s detection could revolutionize the market by providing a reliable and early assessment tool. With the potential for continuous monitoring and sensitivity in detecting cognitive decline, it aligns with the growing need for timely intervention in Alzheimer’s disease management. This technology holds significant promise for both research and clinical applications, potentially shaping the future of Alzheimer’s diagnostics and treatment development.
