National and Kapodistrian University of Athens
Διαγνωστικά μοντέλα γεροντικής κατάθλιψης βασισμένα στην απεικόνιση της λευκής ουσίας του εγκεφάλου με μαγνητικό συντονισμό
Abstract
dc:descriptionThe aim of the present research was to develop a machine learning model for the diagnosis of late-life major depression based on an advanced framework for brain white matter segmentation. Twenty six subjects with major depression and 12 healthy controls, aged 55 years or above matched for age, MMSE score, and education years, underwent brain diffusion magnetic resonance imaging and whole brain white matter segmentation using multiple diffusion sensor imaging indices and atlases within the MRI cloud platform. Fractional anisotropy volume and mean values of fractional anisotropy, diffusion trace, and diffusivities along the principal diffusion axis and perpendicular to it (radial diffusivity) were extracted from 146 white matter parcels and used to train and test the Adaboost classifier with cross validation. The classifier’s performance was evaluated using appropriate measures of accuracy and sensitivity, the F1 score, and the Area Under the Receiving Operative Curve (ROC-AUC). The statistical significance of the classifier was tested using standard permutation testing with cross validation. The classifier achieved a balanced accuracy, of 71% and a ROC-AUC, of 0.81 with the diffusion trace index and a balanced accuracy, of 70% and a ROC-AUC, of 0.80 with RD, within limbic paths, the cortex-basal ganglia-thalamus-cortex loop, the brainstem, the internal and external capsules, the corpus callosum, and cerebellum. Both indices shared common structures in the classification, while the Fornix was the most important structure for the classification with the two indices. The classifier proved statistically significant, as ROC-AUC scores after permutation were lower than those with the actual data (P = 0.022, for trace and P = 0.024, for RD). The results encourage further investigation of the applied methods in the characterization of geriatric depression, as well as the computer-assisted diagnostics and therapeutics based on anatomical brain changes.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Siarkos, Kostas
- Σιάρκος, Κώστας
Subjects
dc:subject × 17- Γεροντική κατάθλιψη
- Διαχωρισμός της λευκής ουσίας
- Ανάλυση με βάση άτλαντα
- Ταξινόμηση Adaboost
- Μηχανική μάθηση
- Late-life depression
- White matter parcellation
- Atlas-based analysis
- Adaboost classification
- Επιστήμες Υγείας
- Ιατρική και Επιστήμες Υγείας
- Επιστήμες υγείας, άλλοι τομείς
- Ιατρική Πληροφορική
- Health Sciences
- Medical and Health Sciences
- Health sciences, miscellaneous
- Medical Informatics
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/57738