Multimodal HYPR-based Denoising for Improving Brain Pattern Analysis
Bibliographic record
Abstract
Brain pattern analysis is an exploratory yet powerful method for deriving disease-related patterns that function as imaging biomarkers. Given that multiple functional systems may be affected in disease, pattern analysis applied to multimodal imaging datasets provides an increasingly detailed view of pathophysiology. Voxelwise pattern analysis is a purely data-driven method for identifying which regions of the brain are most affected. Given the high voxel noise in these datasets, aggressive spatial filtering is usually employed before analysis. However, it has yet to be investigated whether more sophisticated, resolution-preserving denoising operators, such as HYPR, can improve the precision, interpretability, and robustness of identified patterns. As a foray into these questions, we apply HYPR-based denoising to a small pilot FDG-PET/fMRI dataset of Parkinson’s disease subjects before applying pattern analysis. FDG data are reconstructed using PSF-HYPR4D-K-TOFOSEM and fMRI data are processed with a proposed novel 4D HYPR operator. PET denoising reduces noise-induced variation by up to 40% and HYPR-based fMRI denoising enables voxelwise joint analysis of the PET and fMRI data, allowing one to analyze patterns of accordance and discordance from the functional metrics of each modality. Applying voxelwise pattern analysis to the denoised dataset versus traditional aggressive spatial filtering minimizes the amount of resolution loss, allowing for subregional pattern structure to be more precisely characterized. Therefore, more sophisticated denoising provides truly voxelwise, more interpretable patterns, which may have downstream effects on disease understanding and treatment.
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 | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 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; 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".