Identifying preclinical Alzheimer disease from driving patterns: A machine learning approach
Bibliographic record
Abstract
Abstract Background Alzheimer disease (AD) is the most prevalent form of age‐related dementia. The clinical manifestation of AD is generally preceded by a silent preclinical phase during which early AD brain changes are present but dementia symptoms have not yet appeared. Molecular biomarkers have been used to ascertain the presence of AD brain changes, which are obtained via imaging and lumbar puncture. However, the widespread use of these methods is limited by cost and availability. Therefore, there is a need for a non‐invasive and low‐cost solution for identifying individuals who are likely to have preclinical AD. Since the preclinical phase of AD has been shown to impact driving, daily driving behaviours captured using Global Positioning System (GPS) devices can serve as a digital biomarker to detect preclinical AD. The objective of the present study is to use machine learning methods to evaluate the ability of in‐vehicle GPS devices to distinguish cognitively normal older drivers with preclinical AD from those without preclinical AD. Method We used commercial in‐vehicle GPS devices to study the naturalistic driving behaviours of 144 cognitively normal older drivers (aged 65+) over one year. The cohort included 69 individuals with and 75 without preclinical AD, as determined by cerebrospinal fluid (CSF) biomarkers. Four Random Forest (RF) models were trained with three sets of variables: (1) driving features only, (2) driving features and age, and (3) driving features, age and APOE ε4 status. Finally, the strongest predictors of preclinical AD were identified using an RF‐based Recursive Feature Elimination technique. Result The F1 score of the RF models for identifying preclinical AD was 82% using GPS‐based driving indicators, 88% using age and driving indicators, and 91% using age, APOE ε4 status and driving. The area under the receiver operating curve for the final model was 0.96. APOE ε4 status and age were the two most important features for predicting preclinical AD, and the most important driving feature was the vehicle’s jerk, which is a measure of driving smoothness. Conclusion Driving behaviours captured with GPS can accurately distinguish cognitively normal older drivers with preclinical AD from those without preclinical AD.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.002 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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; both teacher heads agree on what is shown here.
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".