Long-Term Short-Term Memory Networks of Retinal OCT Images Predict the Incidence Trend of Alzheimer's Disease
Bibliographic record
Abstract
Alzheimer's disease is currently a neurodegenerative disease that is clinically difficult to cure, and if it can be prevented and screened as soon as possible, it will reduce the clinical diagnosis rate and alleviate the trend of younger age. For early screening, clock drawing test, mental state scale (MMSE), Montreal cognitive screening scale (MoCA) and so on are widely used, but the scale is highly subjective and has a limited scope of application. At present, time series analysis is mostly used to establish long-term monitoring of patients in order to accurately predict the development trend of AD. The retina is part of the central nervous system that provides information about the state of the brain and its changes, and the thickness of the retinal nerve fiber layer (RNFL) can be observed using OCT technology (optical coherence tomography). In this paper, the Kaggle open-source OCT dataset is used to establish a long short-term memory (LSTM) time series model. In the training model, the data is divided into a training set and a test set, and by continuously training the model, it has been proved that the data features of the training set can be learned and verified by the test set. In this paper, RelayNet convolutional blocks (encoders) are used to segment images after convolutional pooling. The ReLayNet algorithm uses the gradient-weighted class activation mapping method to generate heat maps to highlight the lesion area, increase the model interpretability, segment the retinal layered structure in the OCT image, and the inner and outer retinal Dice coefficients reach 0.9612 and 0.9501, which have good image segmentation effects, respectively. The RNFL of healthy controls is significantly thicker than that of AD patients, so the RNFL thickness of patients with mild cognitive impairment (MCI) can be tracked for a long time and the incidence trend can be predicted.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".