P1‐313: Longitudinal Brain Structure Changes in Healthy/MCI Patients: A Deep Learning Approach for The Diagnosis and Prognosis of Alzheimer’s Disease
Bibliographic record
Abstract
The initial AD pathology develops in situ while the patient is cognitively normal. At some point, sufficient brain damage accumulates to cause cognitive impairment. Clinical experience with AD mainly rests in diagnosis, which heavily depends on brain imaging and a number of biomarkers. In this work, we propose a deep learning based framework for AD diagnosis and prognosis, which formulates a time series prediction problem as multiclass classification. The Alzheimer’s Disease Neuroimaging Initiative (ADNI) database was utilized in our study. The neuroimaging and biological data from 634 ADNI participants (229 Healthy and 405 MCI patients) are used for model construction. Of these, 25 ADNI patients with clear pathology progression are used for risk factor identification. The neuroimaging data are processed by FreeSurfer to measure cortical thickness and volume of neuroanatomical structures. The deep learning autoencoder is adopted for prognosis. The importance of various medical variables is defined as how much the system performance changes if the variable is ruled out. We study two different scenarios, i.e. Healthy-to-MCI progression and MCI-to-AD progression. For both tasks, both right and left sides of the brain show comparable contribution, with the right side slightly stronger. For comparison, we also show the diagnosis results, the brain volumes derived from the left side of the brain show relatively larger importance scores. This indicates that the brain atrophy information of the left brain gives stronger impact on the diagnosis of AD, compared with the equal impact of both sides on disease progression. The proposed system shows an overall prediction accuracy of 81.67%. Despite the stronger impact of left brain on diagnosis, both sides of the brain play an important role in prognosis. In particular, the hippocampus volume is highly correlated with AD diagnosis but not directly associated with progression. The identified high risk regions distribute around the back of the brain (e.g. Occipital and Parietal), which is consistent with the clinical knowledge about AD pathology. Moreover, deep learning shows inspiring results in AD prognosis, which provides evidence for the effectiveness of artificial intelligence in AD study.
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 | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| 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".