Classifying mild cognitive impairment from normal cognition: fMRI complexity matches tau PET performance
Bibliographic record
Abstract
Background: Tau protein accumulation is closely linked to synaptic and neuronal loss in Alzheimer's disease (AD), resulting in progressive cognitive decline. Although tau-PET imaging is a direct biomarker of tau pathology, it is costly, carries radiation risks, and is not widely accessible. Resting-state functional MRI (rs-fMRI) complexity-an entropy-based measure of BOLD signal variation-has been proposed as a non-invasive surrogate biomarker of early neuronal dysfunction associated with tau pathology. Objectives: To determine whether fMRI-based brain complexity (sample entropy and multiscale entropy) can match or exceed tau-PET in classifying cognitively normal (CN) versus cognitively impaired (MCI/AD) individuals. And to investigate and compare the most influential network regions-of-interest (ROIs) for classification between fMRI complexity and tau-PET, thereby identifying key neuroanatomical correlates of AD-related changes. Design: A cross-sectional study employing 3D convolutional neural network (CNN) classification with five-fold cross-validation and leave-one-network-out analysis. Setting: Data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Participants: One hundred forty-seven older adults (age 72.5 ± 7.5 years), including 95 CN, 45 MCI, and 7 AD. Measurements: We created whole-brain complexity maps from rs-fMRI and standardized uptake value ratio (SUVR) maps from tau-PET. Each modality was separately fed into CNN classifiers. Region-based analyses (leave-one-network-out) were performed to identify critical ROIs for classification. Results: fMRI complexity showed classification accuracy comparable to tau-PET yet surpassed it in F1-score (0.64 vs. 0.61) and area under the curve (AUC; 0.73 vs. 0.67). Salience and dorsal attention networks contributed most to fMRI-based classification, and a dorsal attention network contributed most to tau-PET-based classification. Conclusions: fMRI complexity performs similarly to tau-PET in detecting cognitive impairment related to AD and identifies partially distinct critical ROIs, suggesting an alternative, radiation-free imaging biomarker for earlier detection and broader clinical application.
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) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| 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".