Computerized Technique for Computer-Aided Diagnosis of Dementia Based on Structural MRI Data

Neurodegenerative diseases are frequently associated with structural changes in the brain. Magnetic Resonance Imaging (MRI) scans can show these variations and therefore be used as a supportive feature for a number of neurodegenerative diseases. The hippocampus has been known to be a biomarker for Alzheimers disease and other neurological and psychiatric diseases. However, it requires accurate, robust and reproducible delineation of hippocampal structures. This work utilises a datasets consisting of MR images shared by EADC-ADNI working group . Hippocampus volume is a feature used in this analysis. For the other features we used publicly available brain segmentation package FreeSurfer v.5.1 (FS) (freesurfer.nmr.mgh.harvard.edu) [17] to process the structural brain MRI scans and compute morphological measurements. The FreeSurfer pipeline is fully automatic and provides 184 features per MRI scan in total. Volumes of cortical and sub-cortical structures such as the caudate and average thickness measurements within cortical regions, such as the precuneus. We use the FS features but for hippocampus volume we use the segmentation proposed in [16]. For the diagnosis classification we passed all the features to a C-Support Vector Classifier (C-SVC) with a linear kernel on a 5-fold cross validation. The goal is evaluating the performance of an algorithms for multi-class classification: AD, MCI and controls. Methods that are developed for binary classification can be used for three-way classification by using either a one-vs-one (ovo) or one-vs-all (ova) strategy. In this approach, three classifiers are trained for the three binary problems using the ovo methodology and thereafter their outputs are combined into three predictions.