Exploring hippocampal asymmetrical features from magnetic resonance images for the classification of Alzheimer's disease

Author(s)

de Oliveira, Italo A. D. ; Poloni, Katia M. ; Ferrari, Ricardo J. ; DeHerrera, AGS ; Gonzalez, AR ; Santosh, KC ; Temesgen, Z ; Kane, B ; Soda, P

• Document type: Journal article
• Source: 2020 IEEE 33RD INTERNATIONAL SYMPOSIUM ON COMPUTER-BASED MEDICAL SYSTEMS(CBMS 2020); v. N/A, p. 6-pg., 2020-01-01.

Abstract

Alzheimer's disease (AD) is the most common cause of dementia, accounting for 60 to 80% of all cases. Because of population aging, this disease has become one of the most relevant global public health problems. Several studies have shown the hippocampal structures present significant asymmetry in AD, and that difference, measured from the volumes between left and right hippocampus, varies with the disease progression. Although imaging biomarkers have been proposed to investigate whether the asymmetry of hippocampus subfields changes through the disease progression, little attention has been paid to explore asymmetrical hippocampal image features to aid for AD early diagnosis. In this study, we propose a new method for the classification of Magnetic Resonance (MR) images in both cognitively normal (CN) versus mild cognitive impairment (MCI) and CN versus mild-AD patient groups using only asymmetrical features extracted from the hippocampal MR image hemispheres. The features, devised from the magnitude response images resulting from applying 3-D log-Gabor filters to an MR input image, are used to train Support Vector Machine classifiers for the MR image classification. Quantitative evaluation of our proposed method applied to MR image classification resulted in accuracy, Fl-score, and AUC average values of 71.23%, 0.67, and 0.77 for the CN x MCI case, and 80.43%, 0.75, and 0.88 for the CNx AD case. These results are very promising, considering we used only asymmetrical features from the hippocampal regions in this study. (AU)

• FAPESP's process: 18/09972-9 - Detection and analysis of hippocampal structural asymmetries in magnetic resonance images with application to aid in the diagnosis of Alzheimer's Disease
• Grantee: Italo Antonio Duarte de Oliveira
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• FAPESP's process: 18/08826-9 - Development of feature engineering and deep learning techniques applied to the classification of magnetic resonance images in healthy cognitive aging, mild cognitive impairment and Alzheimer's Disease
• Grantee: Ricardo José Ferrari
• Support Opportunities: Regular Research Grants
• FAPESP's process: 18/06049-5 - Automatic computational scheme for the detection, identification and classification of cerebral structural changes in magnetic resonance images to aid the diagnosis of patients with mild cognitive impairment and mild Alzheimer's disease
• Grantee: Katia Maria Poloni
• Support Opportunities: Scholarships in Brazil - Doctorate