Harnessing AI-Driven Mobile Applications for Cognitive Impairment Screening
Charisse Litchman1, Larry Rubin1, Sharon Stoll2
1BeCare Link, 2Stoll Medical Group
Objective:
This study evaluates the efficacy of the BeCareLink mobile app in detecting cognitive impairment through quantification of neurologic function.
Background:
The diagnosis of Alzheimer’s disease often relies on a “rule out” approach, yielding only a 60% accuracy rate. With advancements in diagnostic tools for early cognitive impairment, there is a critical need for an accessible screening method to detect pre-clinical changes and improve patient outcomes. Digital technologies combined with artificial intelligence (AI) offer a novel strategy for early detection.
Design/Methods:
The BeCare Neuro App collects user-reported symptoms and measures aspects of neurologic function via gamified activities. It assesses cognitive skills across four domains using tasks similar to those in comprehensive neuropsychological exams, reporting performance against normative scores. Key tasks include: the “Cognitive Test” (decoding messages by pairing symbols with letters), the “Stroop Test” (naming the color of displayed words), the “Memory Test” (recalling a disappearing animal), and the “Tap Test” (tapping randomly appearing coins). The app has over 13,000 users and has recorded significant completions across all tests. It computes population percentiles based on combined time and accuracy metrics.
Results:
Sixty-five users who completed all tests and reported “cognitive” or “memory loss” were analyzed. Users consistently scored in the bottom third of the population. Specifically, 90.8% (59/65) had at least one activity in the bottom third, 66.2%(43/65) had two or more, and 30.8% had three or more activities in that range.
Conclusions:
The BeCare Neuro App demonstrates potential as a screening tool for pre-clinical and early cognitive impairment, assisting clinicians in identifying patients who may require further evaluation.
10.1212/WNL.0000000000209056
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.