Automatic analysis of Alzheimer’s disease: Evaluation of eye movements in natural conversations
Bibliographic record
Abstract
Abstract Background Among the studies concerned with the description of changes in nonverbal communications, some authors have established a link between eye movement and AD [1]. In this presentation, we introduce a method to automatically analyze and evaluate the performance of AD patients during natural conversations through eye movements. This project was conducted as part of a research program funded by Natural Sciences and Engineering Research Council of Canada ( "Pattern recognition for the detection and monitoring of verbal and non‐verbal alterations in Alzheimer's disease", RGPIN‐2018‐05714). Method We analyzed 22 videos in the Carolinas Conversations Collection, a dataset where conversations with elderly people were recorded. These videos have 370 minutes approximately, with 17 subjects, 9 healthy controls (HC), and 8 AD patients. We tracked the eye movements through time in two ways, as a vector [Figure 1] and as landmarks on the eyes [Figure 2]. Then we trained classifiers to distinguish HCs and AD subjects, and performed the experiments using gaze vectors and eye landmarks as features (individually, combined). Result We obtain the performance of the classifiers measuring the accuracy, precision, recall, and ROC curve. Our results show that the combination of eye landmarks and gaze vectors features is strongly related with the AD condition. The classification sensitivity (ROC) is 77% performing with gaze vector features, 85% using the eyes landmarks in 3D, and combining both, we obtained 88% sensitivity (HC vs. AD) [Table 1]. Conclusion The automatic analysis of the eye movements could be a useful tool for clinicians and researchers studying early signs of AD. In future work, we will analyze facial expressions in order to study the link between verbal features (related to speech and content) and non‐verbal features (facial expressions, eye movements), and their correlations with AD. [1] Fernández G, Mandolesi P, Rotstein NP, Colombo O, Agamennoni O, Politi LE. Eye movement alterations during reading in patients with early Alzheimer disease. Invest Ophthalmol Vis Sci 2013;54:8345–52. https://doi.org/10.1167/iovs.13‐12877.
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 | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.000 |
| 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) | 0.001 | 0.001 |
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".