An Application of Affective Computing on Mental Disorders: A Resting State fNIRS Study⁎⁎This work was supported in part by the National Key Research and Development Program of China (Grant No.2019YFA0706200), in part by the National Natural Science Foundation of China (Grant No.61632014, No.61627808, No.61210010), in part by the National Basic Research Program of China (973 Program, Grant No.2014CB744600), in part by the Program of Beijing Municipal Science & Technology Commission (Grant No.Z171

• Published in: IFAC-PapersOnLine Vol. 53; no. 5; pp. 464 - 469
• Main Authors: Wu, Chunyun, Sun, Jieqiong, Wang, Tao, Zhao, Chengjian, Zheng, Shuzhen, Lei, Chang, Peng, Hong
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 2020
• Subjects: Affective Computing; Brain Network; fNIRS; Mental Disorder; fNIRS; Brain Network; Mental Disorder; Affective Computing
• ISSN: 2405-8963
• DOI: 10.1016/j.ifacol.2021.04.195

Affective computing is important for making computers smarter. When emotion can be quantified, machines can understand it. This study aims to apply affective computing to mental disorders, and to classify healthy people and mentally illnesses. For this purpose, 85 subjects, including major depressive disorder patients, schizophrenia patients, and health control people were recruited to participate in resting state functional near infrared spectroscopy (fNIRS) experiment. We measured the changes in oxygenated blood concentration in the prefrontal cortex (PFC). We then used three types of correlation analysis methods to construct the functional connectivity matrices: Pearson correlation analysis (CORR), amplitude squared coherence coefficient (COH), and phase locking value (PLV). We performed the small-world model and centrality analysis based on these matrices. The results demonstrated the existence of a small-world model in both patients and healthy people’s brain networks. Furthermore, features such as the characteristic path length and betweenness centrality extracted from the functional connectivity matrix are helpful for classifying patients and healthy people, thus providing a method for detecting and identifying mental disorders.