Neuroimaging Feature Fusion and Multimodal Classification of Alzheimer’s Disease

Authors

Mehdi Shojaie, Florida International UniversityFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Electrical and Computer Engineering

First Advisor's Name

Malek Adjouadi

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Mercedes Cabrerizo

Second Advisor's Committee Title

Committee Member

Third Advisor's Name

Armando Barreto

Third Advisor's Committee Title

Committee Member

Fourth Advisor's Name

Naphtali Rishe

Fourth Advisor's Committee Title

Committee Member

Fifth Advisor's Name

David A. Loewenstein

Fifth Advisor's Committee Title

Committee Member

Keywords

Machine Learning, Alzheimer's Disease, Early Diagnosis, Classification, PET, MRI, Transfer Learning, LSTM

Date of Defense

11-2-2022

Abstract

Alzheimer’s Disease (AD) is a neurodegenerative disorder and the most prevalent cause of dementia. Early detection of AD is critical for enabling early intervention and for slowing its progression. This dissertation aims to present effective machine learning frameworks using multimodal biomarkers for the diagnosis of AD, specifically in the earliest stage. Moreover, a transfer-learning framework is proposed to transfer model-learning knowledge from a source domain with a large amount of labeled data to a target domain with insufficient data for creating an ML model from scratch.

Accordingly, a feature ranking metric is formulated based on the mutual information index to assess the relevance and redundancy of regional biomarkers and improve AD classification accuracy. The multivariate mutual information metric was utilized to capture the redundancy and complementarity of the predictors and develop a feature ranking approach. This was followed by evaluating the capability of single-modal and multimodal biomarkers in predicting the cognitive stage. Moreover, the Amyloid-Tau-Neurodegeneration (AT(N)) biomarker framework was used to explore the misclassified cases. The F1-score show that although amyloid-β deposition is an earlier event in the disease trajectory, tau PET with feature selection yielded a higher early-stage classification score (65.4%) compared to the modalities of amyloid-β PET (63.3%) and MRI (63.2%). The support vector classifier (SVC) multimodal scenario with feature selection improved the F1-score to 70.0% and 71.8% for the early and late-stage, respectively.

As another research endeavor, an instance-based transfer-learning framework is presented based on the gradient boosting machine (GBM) to transfer knowledge from a source to a target domain. In our transfer learning version of GBM (TrGB) a weighting mechanism based on the residuals of the base learners is defined for the source instances. Consequently, instances with different distribution than the target data will have a lower impact on the target learner. Using the Mount-Sinai dataset as target domain, and the ADNI dataset as source domain, TrGB improved the classification scores by 4.5% for MCI diagnosis compared to the baseline. Also, for the early MCI vs. late MCI, using knowledge transfer from the NC vs. AD of the source domain, scores improved by 5%.

Identifier

FIDC010956

Previously Published In

M. Shojaie, M. Cabrerizo, S. T. DeKosky, D. E. Vaillancourt, D. Loewenstein, R. Duara, and M. Adjouadi, “A transfer learning approach based on gradient boosting machine for diagnosis of Alzheimer’s disease,” Frontiers in Aging Neuroscience, vol. 14, 2022.

M. Shojaie, S. Tabarestani, M. Cabrerizo, ST. DeKosky, DE. Vaillancourt, D. Loewenstein, R. Duara and M. Adjouadi, "PET Imaging of Tau Pathology and Amyloid-β, and MRI for Alzheimer’s Disease Feature Fusion and Multimodal Classification", Journal of Alzheimer's Disease, vol. 84, no. 4, pp. 1497-1514, 2021.

S. Tabarestani, M. Aghili, M. Shojaie, C. Freytes, M. Cabrerizo, A. Barreto, N. Rishe, R. E. Curiel, D. Loewenstein, R. Duara, and M. Adjouadi, “Longitudinal prediction modeling of alzheimer disease using recurrent neural networks,” 2019 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI), 2019.

Recommended Citation

Shojaie, Mehdi, "Neuroimaging Feature Fusion and Multimodal Classification of Alzheimer’s Disease" (2022). FIU Electronic Theses and Dissertations. 5158.
https://digitalcommons.fiu.edu/etd/5158