AUTHOR=Chou Po-Han , Yao Yun-Han , Zheng Rui-Xuan , Liou Yi-Long , Liu Tsung-Te , Lane Hsien-Yuan , Yang Albert C. , Wang Shao-Cheng 

TITLE=Deep Neural Network to Differentiate Brain Activity Between Patients With First-Episode Schizophrenia and Healthy Individuals: A Multi-Channel Near Infrared Spectroscopy Study

JOURNAL=Frontiers in Psychiatry

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2021.655292

DOI=10.3389/fpsyt.2021.655292

ISSN=1664-0640

ABSTRACT=Backgrounds: Reduced brain cortical activity over the frontotemporal regions measured by near infrared spectroscopy (NIRS) has been reported in patients with first-episode schizophrenia (FES). The aim of this study was to differentiate patients with FES and healthy controls (HCs) based on the frontotemporal activity measured by NIRS with a support vector machine (SVM) and deep neural network (DNN) classifier. In addition, we compared the accuracy of performance of SVM and DNN.
Methods: A total of 33 FES patients and 34 HCs were recruited. Their brain cortical activities were measured by NIRS while performing letter and category versions of verbal fluency tests (VFTs). The integral and centroid values of brain cortical activity in the bilateral frontemporal regions during the VFTs were selected as features in SVM and DNN classifier. 
Results: Compared to HCs, FES patients displayed reduced brain cortical activity over the bilateral frontotemporal regions during both types of VFTs. Regarding the classifier performance, SVM reached an accuracy of 68.6%, while DNN reached an accuracy of 79.7% with a sensitivity 88.8% and specificity 74.9% in the classification of patients and normal controls. 
Conclusions: Similar to previous studies, we found that brain activity during the VFT measured by NIRS could be used as a potential marker to classify FES patients from HCs. Future additional independent dataset are needed to confirm the validity of our model.