ResNet-50 Transfer Learning for Early Detection of Alzheimer’s Disease on OASIS-1 MRI with Explainable AI Interpretability Analysis Using Grad-CAM
Abstract
Abstract-Early detection of Alzheimer's disease (AD) is es-sential for timely clinical intervention. This study proposes an interpretable deep learning framework for binary AD classifi-cation using two-dimensional coronal slices from the OASIS 1 structural MRI dataset. A pretrained ResNet-50 model was fine tuned under five training scenarios to evaluate the roles of class weighting, data augmentation, and optimization strategies. Evaluation was conducted at both slice and subject levels using multiple metrics, with the F2 score prioritized to emphasize re-call-oriented performance. Multiple-run experiments using five independent random seeds demonstrated that the final configu-ration was stable across different initializations. The best model achieved a subject level AUC of 0.924 and an F2 score of 0.860 at the optimal threshold, with a recall of 0.987 across runs. Gra-dient weighted Class Activation Mapping showed that the model's attention aligned with established AD biomarkers in the medial temporal lobe, particularly the hippocampus and para-hippocampal gyrus. These findings indicate that a ResNet-50-based approach can provide accurate and interpretable predic-tions for early AD screening using structural MRI.
Keywords-Alzheimer, early detection, deep learning, transfer learning, ResNet-50, MRI, Grad-CAM, explainable artificial in-telligence
References
S. Safiri, A. Gha, and A. Fazlollahi, “Alzheimer ’ s disease : a
comprehensive review of epidemiology , risk factors , symptoms
diagnosis , management , caregiving , advanced treatments and
associated challenges”.
S. Dardouri, “An efficient method for early Alzheimer’s disease
detection based on MRI images using deep convolutional neural
networks,” Front. Artif. Intell., vol. 8, no. April, pp. 1–11, 2025, doi:
3389/frai.2025.1563016.
F. Momeni, D. Shahbazi-Gahrouei, T. Mahmoudi, and A. Mehdizadeh,
“Transfer Learning and Neural Network-Based Approach on Structural
MRI Data for Prediction and Classification of Alzheimer’s Disease,”
Diagnostics, vol. 15, no. 3, 2025, doi: 10.3390/diagnostics15030360.
D. Coluzzi et al., “Biomarker Investigation Using Multiple Brain
Measures from MRI Through Explainable Artificial Intelligence in
Alzheimer’s Disease Classification,” Bioengineering, vol. 12, no. 1,
pp. 1–28, 2025, doi: 10.3390/bioengineering12010082.
Z. Knapińska and J. Mulawka, “Patient-Tailored Dementia Diagnosis
with CNN-Based Brain MRI Classification,” Appl. Sci., vol. 15, no. 9,
, doi: 10.3390/app15094652.
L. Arya, “Comparative Evaluation of Deep Learning Models in
Alzheimer ’s,” Procedia Comput. Sci., vol. 258, pp. 2352–2361, 2025,
doi: 10.1016/j.procs.2025.04.498.
J. ur Rahman, M. Hanif, O. ur Rehman, U. Haider, S. Mian Qaisar, and
P. Pławiak, “Stages prediction of Alzheimer’s disease with shallow 2D
and 3D CNNs from intelligently selected neuroimaging data,” Sci.
Rep., vol. 15, no. 1, pp. 1–17, 2025, doi: 10.1038/s41598-025-93560-
x.
R. Turrisi, A. Verri, and A. Barla, “Deep learning-based Alzheimer’s
disease detection: reproducibility and the effect of modeling choices,”
Front. Comput. Neurosci., vol. 18, 2024, doi:
3389/fncom.2024.1360095.
M. Taiyeb Khosroshahi et al., “Explainable Artificial Intelligence in
Neuroimaging of Alzheimer’s Disease,” Diagnostics, vol. 15, no. 5,
, doi: 10.3390/diagnostics15050612.
B. A. C. Ramalho et al., “The impact of the orientation of MRI slices
on the accuracy of Alzheimer’s disease classification using
convolutional neural networks (CNNs),” J. Med. Artif. Intell., vol. 7,
pp. 0–3, 2024, doi: 10.21037/jmai-24-51.
V. Viswan, N. Shaffi, M. Mahmud, K. Subramanian, and F.
Hajamohideen, Explainable Artificial Intelligence in Alzheimer’s
Disease Classification: A Systematic Review, vol. 16, no. 1. Springer
US, 2024. doi: 10.1007/s12559-023-10192-x.
B. H. M. van der Velden, H. J. Kuijf, K. G. A. Gilhuijs, and M. A.
Viergever, “Explainable artificial intelligence (XAI) in deep learningbased medical image analysis,” Med. Image Anal., vol. 79, p. 102470,
, doi: 10.1016/j.media.2022.102470.
T. Mahmud, K. Barua, S. U. Habiba, N. Sharmen, M. S. Hossain, and
K. Andersson, “An Explainable AI Paradigm for Alzheimer’s
Diagnosis Using Deep Transfer Learning,” Diagnostics, vol. 14, no. 3,
pp. 1–24, 2024, doi: 10.3390/diagnostics14030345.
D. Wang et al., “Deep neural network heatmaps capture Alzheimer’s
disease patterns reported in a large meta-analysis of neuroimaging
studies,” Neuroimage, vol. 269, no. January, 2023, doi:
1016/j.neuroimage.2023.119929.
S. S. Vinayak E, S. A, and N. Khan A, “Dementia Prediction on OASIS
Dataset using Supervised and Ensemble Learning Techniques,” Int. J.
Eng. Adv. Technol., vol. 10, no. 1, pp. 244–254, 2020, doi:
35940/ijeat.a1827.1010120.
D. S. Marcus, T. H. Wang, J. Parker, J. G. Csernansky, J. C. Morris,
and R. L. Buckner, “Open Access Series of Imaging Studies (OASIS):
Cross-sectional MRI Data in Young, Middle Aged, Nondemented, and
Demented Older Adults,” J. Cogn. Neurosci., vol. 19, no. 9, pp. 1498–
, Sep. , doi: 10.1162/jocn.2007.19.9.1498.
A. Kebaili, J. Lapuyade-Lahorgue, and S. Ruan, “Deep Learning
Approaches for Data Augmentation in Medical Imaging: A Review,”
J. Imaging, vol. 9, no. 4, 2023, doi: 10.3390/jimaging9040081.
V. P. Nithya, N. Mohanasundaram, and R. Santhosh, “An Early
Detection and Classification of Alzheimer’s Disease Framework Based
on ResNet-50,” Curr. Med. Imaging Former. Curr. Med. Imaging Rev.,
vol. 20, pp. 1–20, 2023, doi: 10.2174/1573405620666230825113344.
F. Hasan Saif, M. N. Al-Andoli, and W. M. Y. W. Bejuri, “Explainable
AI for Alzheimer Detection: A Review of Current Methods and
Applications,” Appl. Sci., vol. 14, no. 22, 2024, doi:
3390/app142210121.
D. P. Veitch et al., “The Alzheimer’s Disease Neuroimaging Initiative
in the era of Alzheimer’s disease treatment: A review of ADNI studies
from 2021 to 2022,” Alzheimer’s Dement., vol. 20, no. 1, pp. 652–694,
, doi: 10.1002/alz.13449.
J. Zhou et al., “A deep learning model for early diagnosis of
alzheimer’s disease combined with 3D CNN and video Swin
transformer,” Sci. Rep., vol. 15, no. 1, pp. 1–17, 2025, doi:
1038/s41598-025-05568-y.
