TL;DR
This study explored the use of a smartphone-based cognitive assessment tool, DANA, to detect Mild Cognitive Impairment (MCI). They found that specific digital cognitive measures, particularly those related to executive function from the "Code Substitution" task, were significantly associated with MCI, suggesting mobile assessments could be a valuable tool for early detection.
What Problem Does This Paper Address?
Early and accurate detection of Mild Cognitive Impairment (MCI) is crucial for timely intervention and management of neurodegenerative diseases like Alzheimer's. Traditional cognitive assessments can be time-consuming, resource-intensive, and may not be easily accessible for frequent monitoring. The paper addresses the need for more efficient, accessible, and scalable methods for cognitive assessment, exploring the potential of self-administered mobile applications. Specifically, it investigates whether digital cognitive measures derived from smartphone tasks can effectively identify individuals with MCI.
Methods
The researchers utilized the Defense Automated Neurobehavioral Assessment (DANA) platform, a suite of six cognitive tasks administered via smartphone. For each task, five digital cognitive measures were extracted: average response time for all test trials (ART), standard deviation of response time for all test trials (SDRT), average response time for correct test trials (ART_correct), standard deviation of response time for correct test trials (SDRT_correct), and cognitive efficiency (CE). These digital measures were standardized to have a mean of zero and a standard deviation of one. Logistic regression analyses, adjusted for age, sex, and education, were employed to determine the association between these digital measures and cognitive status (cognitively intact vs. MCI).
Datasets
The study used data from participants enrolled in the Boston University Alzheimer's Disease Research Center (BU ADRC). Participants were categorized as either cognitively intact or diagnosed with MCI based on consensus diagnostic meetings at the BU ADRC, adhering to the National Alzheimer's Coordinating Center Uniform Data Set criteria. A total of 132 participants were included in the study. No new benchmarks were introduced; the study evaluated existing DANA tasks against established diagnostic criteria.
Key Insights and Ideas
- Executive Function is Key: The findings strongly suggest that digital measures reflecting executive function are the most sensitive indicators of MCI. This highlights the importance of tasks that challenge planning, inhibition, and working memory for early detection.
- Mobile Assessments Show Promise: The study provides evidence that self-administered smartphone applications can serve as a viable alternative for cognitive monitoring and early detection of cognitive impairment, offering a more accessible and potentially frequent assessment method.
- Task Specificity Matters: Not all digital cognitive measures from all tasks were equally effective. The "Code Substitution" task, in particular, demonstrated significant associations with MCI, while other tasks like "Go/No-Go" and "Spatial Processing" did not yield significant results for most measures.
Key Results
Out of 132 participants (mean age 71.9 ± 10.2 years, 57.6% female), 17 were diagnosed with MCI.
- Code Substitution Task: All five digital measures from this task were significantly associated with MCI. Notably, a one-standard deviation increase in cognitive efficiency (CE) for this task was associated with a 76% reduction in the odds of MCI (OR = 0.24, 95% CI = 0.09-0.51, P < 0.001).
- Procedural Response Time Task: Most measures from this task were associated with MCI, with ART showing an odds ratio of 2.29 (95% CI = 1.26-4.40, P = 0.008) and ART_correct showing an odds ratio of 2.52 (95% CI = 1.37-5.01, P = 0.005). CE also showed a significant association (OR = 0.31, 95% CI = 0.13-0.64, P = 0.004).
- Other Tasks: Digital measurements from the "Go/No-Go," "Match-to-Sample," "Spatial Processing," and "Simple Response Time" tasks were generally not associated with MCI.
Limitations and Future Work
The authors acknowledge that this was a pilot study with a relatively small sample size (n=132), particularly for the MCI group (n=17). This limits the generalizability of the findings. Future work should involve larger, more diverse cohorts to validate these results and further explore the utility of specific DANA tasks and measures for longitudinal monitoring and differentiating MCI subtypes. Investigating the predictive power of these digital measures for future cognitive decline or conversion to Alzheimer's disease would also be a valuable next step.
Summary updated