Pattern Analysis for the Neuroimaging Based Diagnosis of Schizophrenia


DEMİRHAN A.

25th Signal Processing and Communications Applications Conference (SIU), Antalya, Türkiye, 15 - 18 Mayıs 2017 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/siu.2017.7960364
  • Basıldığı Şehir: Antalya
  • Basıldığı Ülke: Türkiye

Özet

Accurate diagnosis at an early stage of schizophrenia, a chronic and severe mental illness that affects people's thoughts, feelings, and behavior, has a critical importance for the initiation of the treatment and monitoring of the disease. In this study, a pattern analysis study is performed for the diagnosis of schizophrenia using structural magnetic resonance (MR) images. Morphometric features like volumes and thickness of the anatomical structures of the brain that reveal the disease-related patterns are used to discriminate between schizophrenia patients and healthy controls. Morphometric brain features that are obtained from the MR images of 150 subjects are analyzed with support vector machines (SVM) trained with 5-fold cross validation. Results obtained from the study showed that using SVM together with morphometric brain features could be a successful complementary method for the early clinical diagnosis of schizophrenia.