Using Deep Neural Network to Differentiate Brain Activity Between Patients With Recent-Onset Schizophrenia and Healthy Individuals
Poster B42, Tuesday, October 9, 11:30 am - 1:00 pm, Essex Ballroom
Po Han Chou1, Yun-Han Yao2, Rui-Xuan Zhang2, Yi-Long Liou2, Tsung-Te Liu2; 1Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA., 2Graduate Institute of Electronics Engineering, National Taiwan University, Taipei, Taiwan.
Backgrounds: Reduced brain cortical activity has been reported in patients with schizophrenia, and could reveal the underlying neural correlates of the neuropathophysiology of the disease. The study of schizophrenia at early stage is of particular interest because the confounding role of chronicity and medication can be excluded. The aim of this study was to automatically discriminate patients with recent onset schizophrenia (ROS) and normal controls on the basis of brain activity measured by near infrared spectroscopy (NIRS) employing a support vector machine (SVM) and deep neural network (DNN) classifier. Methods: A total of 33 ROS patients and 34 healthy controls (HCs) underwent NIRS while performing two versions of verbal fluency tests (VFTs). The patterns of hemodynamic changes in brain cortical activity of the bilateral frontemporal regions during the VFTs were selected as features in SVM and DNN classifier. Results Compared to HCs, patients displayed reduced brain cortical activity over the bilateral frontotemporal regions during both types of VFTs. With regard to the classifier performance, SVM reached an accuracy of 68.6%, while DNN reached an accuracy of 79.7% in the classification of patients and normal controls. Conclusions Similar to previous study, we confirmed that brain activity during the VFT measured by NIRS could be used as a potential marker to classify patients with schizophrenia, who show reduced brain activity in respect to normal controls. Moreover, we found DNN approaches have improved ability to learn hidden patterns in brain imaging data.
Topic Area: Neuroimaging