Machine learning-supported MRI analysis of brain asymmetry for early diagnosis of dementia
Herzog, Nitsa J. and Magoulas, George D. (2022) Machine learning-supported MRI analysis of brain asymmetry for early diagnosis of dementia. In: Hassanien, A.E. and Bhatnagar, R. and Snášel, V. and Yasin Shams, M. (eds.) Medical Informatics and Bioimaging Using Artificial Intelligence. Studies in Computational Intelligence 1005. Springer, pp. 29-52. ISBN 9783030911034.
|
Text
NH_GM_book_chapter1.pdf - Author's Accepted Manuscript Download (685kB) | Preview |
Abstract
The chapter focuses on the detection of early degenerative processes in the human brain using computational algorithms and machine learning classification techniques. The research is consistent with the hypothesis that there are changes in brain asymmetry across stages of dementia and Alzheimer’s Disease. The proposed approach considers the pattern of changes in the degree of asymmetry between the left and right hemispheres of the brain using structural magnetic resonance imaging of the ADNI database and image analysis techniques. An analysis of levels of asymmetry is performed with the help of statistical features extracted from the segmented asymmetry images. The diagnostic potential of these features is explored using variants of Support Vector Machines and a Convolutional Neural Network. The proposed approach produces very promising results in distinguishing between cognitively normal subjects and patients with early mild cognitive impairment and Alzheimer’s Disease, providing evidence that image asymmetry features or MRI images of segmented asymmetry can offer insight on early diagnosis of dementia.
Metadata
Item Type: | Book Section |
---|---|
School: | Birkbeck Faculties and Schools > Faculty of Science > School of Computing and Mathematical Sciences |
Research Centres and Institutes: | Birkbeck Knowledge Lab |
Depositing User: | George Magoulas |
Date Deposited: | 04 Apr 2022 17:53 |
Last Modified: | 05 Apr 2024 00:10 |
URI: | https://eprints.bbk.ac.uk/id/eprint/47964 |
Statistics
Additional statistics are available via IRStats2.