EEG affective modelling using dysphoria model of affect (DMOA) for dysphoria detection / Mohd Hafiz Mohd Nasir

• Main Author: Mohd Nasir, Mohd Hafiz
• Format: Thesis
• Language: English
• Published: 2020
• Subjects: Computer applications to medicine. Medical informatics; Computer engineering. Computer hardware
• Online Access: https://ir.uitm.edu.my/id/eprint/59570/1/59570.pdf; https://ir.uitm.edu.my/id/eprint/59570/

Dysphoria is a state when one experienced disappointment. If it is not handled properly, dysphoria may trigger acute stress, anxiety and depression. However, the individual who experienced dysphoria typically are in-denial because dysphoria is always associated with negative connotation such as incompetent to handle pressure, weak personality and others. To complicate matters, there is no widely and easily available computational tool that can be used to measure the dysphoria tendency except using well-established questionnaire by psychologists, such as: Depression, Anxiety and Stress Scale (DASS) and Nepean Dysphoria Scale (NDS-24 instruments. Participants may suppress or exaggerate their answers resulting in misdiagnosis. Moreover, the instruments need well-trained psychiatrists and psychologists to interpret the results. With limited number of psychiatrists and psychologists as well as insufficient number of infrastructure and facilities, the result interpretation is posed as a challenge to accommodate the patients for mental health services. In this work, a conceptual Dysphoria Model of Affect (DMoA) is developed for dysphoria detection. Based on the hypothesis that dysphoria is related to negative emotion, the input from brain signal is captured from electroencephalogram (EEG) is used to detect negative emotions. This is because the brain signals are considered ‘pure’ because participants cannot control or alter the signals. The results from the EEG signals are also compared with the DASS and NDS questionnaires for correlation analysis. From the experimental result, the DMoA model can identify negative emotions ranging from 63.14% to 78.98%. The results also show that the NDS questionnaire result is inline with the EEG emotion identification result in particularly for Subject 9 that showed the highest accuracy for fear indicating the potential of using the DMoA to detect dysphoria. Hence, it can be observed that the EEG approach can be an alternative approach to detect dysphoria. It is hoped that the proposed approach can be used for early dysphoria detection for early intervention to identify the mental states of patients before it become worse and thus necessary treatment/help can be provided. Subsequently, the DMoA can be used to complement NDS-24 instrument to give empirical insight that can facilitate psychologists and psychiatrist in their patient’s assessment in a timely manner.