P3‐413: HETEROGENEOUS TAU‐PET SIGNAL IN THE HIPPOCAMPUS HELPS RESOLVE DISCREPANCIES BETWEEN IMAGING AND PATHOLOGY
Bibliographic record
Abstract
In Alzheimer's disease (AD), neurofibrillary tau tangles (NFTs) are believed to first aggregate in the entorhinal cortex (ERC), and eventually spread to the hippocampus, possibly via trans-synaptic mechanisms. However, while NFTs can now be measured in living humans using positron emission tomography (PET), PET signal in the hippocampus has not corroborated findings from pathology studies. In the current study, we examine data-driven patterns of PET signal in the hippocampus, to establish whether discrepancies in this region are due to heterogeneous signal sources. AV1451 images were downloaded from the Alzheimer's Disease Neuroimaging Initiative website (143 older controls, 88 mild cognitive impairment, 27 AD dementia). In a previous study, we used an advanced clustering algorithm in the Swedish BioFINDER cohort to segregate different spatial distribution patterns of AV1451 across the AD spectrum. Average AV1451 signal was extracted across the whole hippocampus, as well as separately across hippocampal voxels belonging to each AV1451 cluster identified in the previous analysis. We then tested whether different AV1451 signal patterns were differentially related to clinical diagnosis, presence of Aβ pathology and episodic memory (EM) scores. Finally, we used diffusion tractography to measure the number of connections between the ERC and the different hippocampal signal clusters, based on a connectome extracted from 114 young controls using ndmg. AV1451 signal in the hippocampus covaried either with other regions involved in early NFT aggregation ("Early NFT"), or with regions susceptible to off-target AV1451 binding ("Off-Target"; Figure 1). Relationships between AV1451 and diagnosis, Aβ status and EM scores were substantially enhanced when only using voxels in the Early NFT cluster, compared with voxels in the Off-Target cluster, or the whole hippocampus (Figure 2). Tractography analysis revealed a greater number of anatomical connections between the ERC and the Early NFT cluster of the hippocampus compared to the Off-Target cluster (p<0.001; Figure 2). Clustering across AV1451 images revealed five spatial covariance networks. Two of these networks, were represented inside the hippocampus, indicating signal heterogeneity in this structure. (Left) The spatial extent of the two clusters. (Above) Membership of each hippocampal voxel in the two clusters projected onto a hippocampal surface. Individuals with higher hippocampal AV1451 signal (gray) were more likely to be Aβ positive (top left), had worse episodic memory scores (top right) and were more frequently diagnosed with AD dementia (bottom left). These associations were enhanced when only looking in hippocampal voxels within the early NFT cluster (turquoise), and were diminished when using voxels in the off-target cluster (purple). The Early NFT cluseter demonstrated greater anatomical connections with the entorhinal cortex as measured using diffusion tractography imaging (bottom right). Using data-driven methods, we were able to enhance the associations expected from pathology studies between hippocampal AV1451 signal and other pathological and cognitive markers. These findings partially resolve discrepancies between previous PET and pathology studies and provide a template for future AV1451 studies.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| 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 | 0.001 | 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".