A Machine Learning‐Based MRI Marker Predicts Incident Hypertension and Mediates the Relationship Between Hypertension and Cognition
Bibliographic record
Abstract
BACKGROUND: Hypertension (HTN) is an established risk factor for neurodegeneration and dementia, supported by epidemiological and neuroimaging studies, although there is large variation in individual outcomes. We developed a machine learning (ML)-based model, termed SPARE-HTN, to quantify the spatial pattern of HTN-related neurodegeneration observable in individual structural magnetic resonance images (sMRI). SPARE-HTN demonstrates superior sensitivity compared to the most widely-used measure of HTN-related brain changes, correlates with cognitive performance, and detects early changes in mid-life years. This study investigated the predictive capacity of SPARE-HTN for incident HTN and its mediating role in the relationship between HTN and cognition. METHODS: SPARE-HTN, derived from N = 37,098 cognitively unimpaired individuals from diverse cohorts, was evaluated in N = 968 (59% female, mean age 63.0 ± 11 years) individuals with longitudinal clinical data from six studies (Table 1). Baseline SPARE-HTN values were compared across participants categorized by longitudinal HTN status: persistently normotensive, persistently hypertensive, or incident HTN. The risk of incident HTN was evaluated among baseline normotensive participants using Cox regression model across the baseline SPARE-HTN quartiles, adjusted for age and sex. The average causal mediated effect (ACME) of SPARE-HTN and total white matter hyperintensity (WMH) volume on the relationship between hypertension and cognitive test scores were assessed using mediation models with 5000 simulated permutations. RESULTS: Table 1 presents the participant characteristics at baseline, stratified by longitudinal HTN status. SPARE-HTN was significantly elevated (Figure 1A) in participants who were normotensive at baseline but developed HTN within 3-7 years (+0.44, p = 0.01), but not in those who developed HTN in 8-12 years (p = 0.65). Normotensive participants with elevated baseline SPARE-HTN scores exhibited significantly higher Cox proportional hazard ratios for HTN incidence (Figure 1B). Mediation analysis demonstrated that SPARE-HTN mediated up to 26% of the effect of HTN on cognitive measures, whilst the conventional WMH volumes showed little mediation effect (Figure 2). CONCLUSION: Our ML-based sMRI marker predicted incident HTN prior to formal clinical diagnosis, suggesting the presence of subclinical cerebrovascular changes possibly associated with blood pressure variations. These markers, particularly relevant in midlife, offer potential for informing dementia prevention trials by enabling individualized risk stratification and potentially more sensitive measurement of therapeutic efficacy.
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 | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 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".