Exploring the intricacies and pitfalls of the ATN framework: An assessment across cohorts and thresholding methodologies
Bibliographic record
Abstract
Abstract The amyloid/tau/neurodegeneration (ATN) framework has redefined Alzheimer’s disease (AD) toward a primarily biological entity. While it has found wide application in AD research, it was so far typically applied to single cohort studies using distinct data-driven thresholding methods. This poses the question of how concordant thresholds obtained using distinct methods are within the same dataset as well as whether thresholds derived by the same technique are interchangeable across cohorts. Given potential differences in cohort data-derived thresholds, it remains unclear whether individuals of one cohort are actually comparable with regard to their exhibited disease patterns to individuals of another cohort, even when they are assigned to the same ATN profile. If such comparability is not evident, the generalizability of results obtained using the ATN framework is at question. In this work, we evaluated the impact of selecting specific thresholding methods on ATN profiles by applying five commonly-used methodologies across eleven AD cohort studies. Our findings revealed high variability among the obtained thresholds, both across methods and datasets, linking the choice of thresholding method directly to the type I and type II error rate of ATN profiling. Moreover, we assessed the generalizability of primarily Magnetic Resonance Imaging (MRI) derived biological patterns discovered within ATN profiles by simultaneously clustering participants of different cohorts who were assigned to the same ATN profile. In only two out of seven investigated ATN profiles, we observed a significant association between individuals’ assigned clusters and cohort origin for thresholds defined using Gaussian Mixture Models, while no significant associations were found for K-means-derived thresholds. Consequently, in the majority of profiles, biological signals governed the clustering rather than systematic cohort differences resulting from distinct biomarker thresholds. Our work revealed that: 1) the thresholding method selection is a decision of statistical relevance that will inevitably bias the resulting profiling, 2) obtained thresholds are most likely not directly interchangeably across two independent cohorts, and 3) MRI-based biological patterns derived from distinctly thresholded ATN profiles can generalize across cohort datasets. Conclusively, in order to appropriately apply the ATN framework as an actionable and robust biological profiling scheme, a comprehensive understanding of the impact of used thresholding methods, their statistical implications, and the validation of achieved results is crucial.
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 | 0.273 | 0.365 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.003 | 0.004 |
| Scholarly communication | 0.006 | 0.002 |
| Open science | 0.004 | 0.006 |
| Research integrity | 0.002 | 0.003 |
| 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".