Lightweight Deep Learning Models with Explainable AI for Early Alzheimer’s Detection from Standard MRI Scans
Bibliographic record
Abstract
Background: Dementia refers to a spectrum of clinical conditions characterized by impairments in memory, language, and cognitive function. Alzheimer’s Disease (AD) is the most common cause of dementia and it accounted for 60–70% of the estimated 57 million cases worldwide as of 2021. The exact pathology of this neurodegenerative condition is not fully understood. While it is currently incurable, progression to more critical stages can be slowed, and early diagnosis is crucial to alleviate and manage some of its symptoms. Contemporary diagnostic practices hinder early detection due to the high costs and inaccessibility of advanced neuroimaging tools and specialists, particularly for populations with resource-constrained clinical settings. Methods: This paper addresses this challenge by developing and evaluating computationally efficient lightweight deep learning models, MobileNetV2 and EfficientNetV2B0, for early AD detection from 2D slices sourced from standard structural magnetic resonance imaging (MRI). Results: For the challenging multi-class task of distinguishing between Cognitively Normal (CN), Early Mild Cognitive Impairment (EMCI), and Late Mild Cognitive Impairment (LMCI), our best model, EfficientNetV2B0, achieved 88.0% mean accuracy across a 5-fold stratified cross-validation (std = 1.0%). To enhance clinical interpretability and build trust, we integrated explainability methods, Grad-CAM++ and Guided Grad-CAM++, to visualize the anatomical basis for the models’ predictions. Conclusions: This work delivers an accessible and interpretable neuroimaging tool to support early AD diagnosis and extend expert-level capabilities to routine clinical practice.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 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".