Computing system for cognitive impairment detection

This computing system can detect cognitive impairment. It presents memory and verbal tasks, analyzes user input, and provides care recommendations.

Background

Digital healthcare solutions are being intensely studied for rapid identification and effective management of patients with Alzheimer’s disease and related disorders (ADRD).

Early detection of cognitive decline is critical to the health and well-being of patients and their family members. Primary care physicians (PCPs) are often the first to hear concerns about cognitive decline because they are the first line of medical care.

Under-diagnosis of ADRD in primary care settings is widely recognized, as are the many barriers to routine cognitive screening. PCPs may lack confidence in their abilities to know who to screen, which brief assessment tool to use, and next steps that should be taken. As a result, a significant percentage of individuals with cognitive impairment go undiagnosed, despite the benefits of early detection.

Technology description

This computing system is designed to help doctors identify cognitive impairment early on. It uses a series of tasks, including a timed symbol matching exercise to test processing speed and working memory, and a verbal task, such as recounting a personal event.

The system analyzes the user’s performance on these tasks, considering factors like speed and accuracy for the symbol matching, and linguistic and acoustic features extracted from the verbal task. Based on this analysis, it calculates a cognitive performance score and provides recommendations for the user’s care, which may include further evaluation or intervention.

What sets this technology apart is its comprehensive approach to cognitive assessment. It combines both verbal and non-verbal tasks, addressing a wider range of cognitive abilities often affected by neurodegenerative diseases.

The system utilizes Bayesian adaptive regression trees (BART) analysis, a machine learning technique, to incorporate and analyze various features from the user’s input, leading to a more accurate and nuanced assessment.

This approach, coupled with its automated nature, makes it a valuable tool for primary care settings where time and specialized expertise may be limited, ultimately facilitating early detection and intervention for cognitive decline.

Benefits

Early detection of cognitive decline, which is crucial for patient well-being and family support
Addresses the challenge of under-diagnosis of cognitive impairment in primary care settings
Provides PCPs with a tool to confidently screen for cognitive impairment, even with limited expertise
Offers a combination of verbal and nonverbal tasks, accommodating patients with various abilities
Delivers automated and efficient assessment, minimizing staff time and resources
Generates a cognitive performance score and provides specific recommendations for patient care