An optimized cognitive biomarker that accurately predicts future cognitive decline and improves the detection of MCI and early dementia
Bibliographic record
Abstract
BACKGROUND: Mild Cognitive Impairment (MCI) affects over 12 million individuals in the US, 50% of whom will progress to Alzheimer's disease or another form of dementia in 3-5 years. But 90% of individuals with MCI remain undiagnosed due to challenges in screening. With the advent of disease-modifying therapies (DMTs) that can slow progression, it is now critical that new and improved tools become available to quickly and accurately screen for MCI, and especially for the subpopulation of individuals at highest risk for developing future dementia. METHODS: Here we use data-driven design to create a novel tool for cognitive assessment that combines information-efficient test-items identified by analysis of the National Alzheimer's Coordinating Center (NACC) NIH Uniform Data Set version 3 (UDSv3) study items in > 10,000 CN, MCI, and early dementia individuals. We assessed the ability to detect MCI measured by the receiver operating characteristic (ROC) area under the curve (AUC) values, selected the best-performing items, and determined the optimal weighting between them. RESULTS: Using both data-driven item selection and optimal item weighting allowed for the creation of a 4-item brief optimized cognitive composite (BOCC) test with a 4-5 minute administration time that can dramatically outperform the 10-14 minute MoCA at detecting both MCI and mild dementia. We find AUC values for the BOCC that are 35% closer to the ideal 1.00 value for BOCC compared to MoCA scores for detecting MCI (0.87 vs 0.80, p <0.0001) and 80% closer to ideal for BOCC vs MoCA scores for detecting mild dementia (0.99 vs 0.95, p <0.0001). Remarkably, the BOCC score also provides more information for predicting 6-year future conversion to dementia for MCI and CN individuals than the dementia specialists' clinical diagnoses (AUCs: 0.95 vs 0.88, p <0.0001). CONCLUSION: A 4-5 minute brief optimized cognitive test can dramatically outperform the 10-14 minute MoCA in detecting MCI, and it can predict future decline to dementia comparably to, or better than, a clinical diagnosis made by clinicians with access to extensive cognitive and clinical information. This test could improve MCI detection in the more than 52 million people in the US age 65 or over.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.015 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".