A Vision Transformer Approach for Fully Automated and Scalable Dementia Screening using Clock Drawing Test Images
Bibliographic record
Abstract
BACKGROUND: The clock drawing test (CDT) has been used as a cognitive screening tool for Alzheimer's disease (AD) and other forms of dementia. However, its clinical utility is constrained by requirements for trained scorers and non-standardized diagnostic criteria. METHOD: We developed a fully-automated vision transformer (ViT)-based dementia diagnostic pipeline incorporating convolutional neural network (CNN) preprocessing of hand drawn CDT images. The architecture implements fine-tuned ViT feature extraction followed by linear classification for dementia prediction. The method was trained using the National Health and Aging Trends Study (NHATS) dataset (n = 54027) and tested on an independent clinical cohort from the Toronto Dementia Research Alliance (TDRA) (n(dementia diagnosis) = 522, n(normal cognition) = 340). RESULTS: The ViT-based approach demonstrated superior predictive performance (balanced accuracy = 76.5%) compared to both traditional human-scored CDT features (balanced accuracy = 74.3%) and three published deep learning architectures when evaluated on the TDRA dataset (balanced accuracy: MiniVGG = 73.3%, MNv2 = 72.3%, RF-VAE = 69.1%). CONCLUSION: This pen-and-paper compatible, transformer-based diagnostic system enables scalable remote cognitive screening through automated CDT image analysis that is competitive with human-scored features, potentially increasing diagnostic accessibility and comfort for elderly populations across diverse socioeconomic contexts.
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) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
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".