Detection of obsessive compulsive disorder using resting-state functional connectivity data


Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2013

Öğrenci: SONA KHANEH SHENAS

Danışman: UĞUR HALICI

Özet:

Obsessive Compulsive Disorder (OCD) is a serious psychiatric disease that might be affiliated with abnormal resting-state functional connectivity (rs-FC) in default mode network (DMN) of brain. The aim of this study is to discriminate patients with OCD from healthy individuals by employing pattern recognition methods on rs-FC data obtained through regions of interest (ROIs) such as Posterior Cingulate Cortex (PCC), Left Inferior Posterior Lobe (LIPL) and Right Inferior Posterior Lobe (RIPL). For this purpose, two different approaches were implemented as feature extraction step of pattern recognition. In the first approach the rs-FC fMRI data were subsampled and then the dimensionality of the subsampled data was reduced using the Principal Component Analysis (PCA), Kernel Principal Component Analysis (KPCA) and Linear Discriminant Analysis (LDA) alternatives. In the second approach, feature vectors having already low dimensions were obtained by measuring cosine similarity, dot product similarity and correlation similarity to the separate means of the rs-FC data of subjects in OCD and healthy groups. Afterwards the healthy and OCD groups were classified using Support Vector Machine (SVM) and Gaussian Mixture Models (GMMs). In order to obtain more reliable performance results, Double LOO-CV method that we proposed as a version of Leave-One-Out Cross Validation (LOO-CV) was used and the best performance (73%) was obtained by using cosine similarity for feature extraction and GMMs for classification.