IC‐P‐128: Robust, Large‐Scale Intensity Standardization of ADNI MRI Dataset
Bibliographic record
Abstract
MRI taken from different scanners appear dissimilar from each other due to a variety of scanner-dependent variations. This situation is particularly acute in large, multi-centric settings such as ADNI. Intensity standardization must therefore be performed in order that similar intensities will have similar tissue meaning in the transformed images across scanners. We employed a novel technique for intensity standardization (STI), which makes use of available reference image tissue masks (background, grey and white matters). After global linear registration of the subject's original image to the standard (example shown in (A)), we computed a mapping function based on the intensity correspondences obtained for each tissue, thereby implicitly binding histogram matching to tissue correspondence. We compared STI to an histogram-matching technique (PCT - Nyul, Udupa et al. 2000), in its original (PCT-10: percentile landmarks spaced by 10%) and modified form (PCT-1: 1% landmark spacing). We based our comparison on four measures: 1) mean absolute error (MAE) w.r.t. standard, 2) Kullback-Leibler divergence (KLD) w.r.t. standard, 3) normalized mutual information (NMI) w.r.t. pre-standardization image, and 4) sum of the diagonal elements of the joint histogram of standard vs. post-standardization images (JHDS). We applied the technique on 757 baseline MRIs from controls, mild cognititive impairment and probable Alzheimer's disease subjects from the ADNI dataset, visually inspected all data, and calculated comparison measures. Unsurprisingly, on histogram-centric measures (MAE, KLD) the PCT-10 and PCT-1 technique fares better than STI (B), with better histogram match (C); yet, as can be appreciated even to the naked eye (A), it does so at the cost of introducing signal variability within the background and white matter. In contrast, over the entire ADNI dataset, STI showed a statistically significant improvement in NMI and JHDS scores (D), corroborating visual inspection (A). With this new technique we were able to standardize intensities in the large, multi-centric ADNI dataset. The major limitation of STI is that global linear registration to the standard image is necessary.
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 | 0.005 | 0.008 |
| Meta-epidemiology (narrow) | 0.003 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.004 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.007 |
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".