Machine learning analysis of speech and eye tracking data to distinguish Alzheimer's clinic patients from healthy controls
Bibliographic record
Abstract
Abstract Background Clinical trials of disease‐modifying therapies for Alzheimer’s disease (AD) are increasingly focused on recruiting individuals with preclinical or early‐stage disease. Artificial intelligence may help in enriching clinical trial populations with high‐risk individuals. We analyzed prospectively‐collected speech and eye‐tracking data to distinguish individuals with mild‐moderate AD, mild cognitive impairment (MCI), and subjective memory complaints (SMC) from age and sex‐matched healthy volunteers. Method Individuals with known clinical diagnoses of AD, MCI, and SMC from a specialty clinic, and healthy controls from the community, were prospectively recruited. Participants described the “Cookie Theft Picture” from the Boston Aphasia Battery. Their speech was recorded and manually transcribed and eye movements were assessed using an infrared eye‐tracker. Data underwent feature extraction for language‐related features including lexical and acoustic parameters (from transcripts and speech), and fixation and saccades (from eye‐tracking). Additionally, for language and eye‐tracking, features capturing spatial neglect are explored following the approach in [1]. Separate predictive models were examined using logistic regression (LR), K‐Nearest Neighbours (KNN), and random forests (RF). Result Recruitment and analysis are ongoing. Data from 34 clinic patients (9% SMC, 35% MCI, 56% mild‐moderate AD, mean age 7310, 44% female) and 39 controls (mean age 6811, 79% female) is reported. Best speech‐based models distinguished patients from controls with an Area Under the ROC Curve (AUC) of 79% (95% CI 0.70‐0.87). Eye‐tracking‐based models reached distinguished patients from controls with an AUC of 71% (95% CI 0.59‐0.82). Conclusion Machine‐learning mediated analysis of speech and eye‐tracking data achieved promising classification performance with the current cohort. Both increasing the size of the corpus with ongoing recruitment, and combining demographic and clinical data alongside multimodal feature fusion, may help to improve classification.
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 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| 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".