Modeling and analysis tools for brain study

• Main Author: Tan, Eileen Yun Rui
• Other Authors: Olga Sourina
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/64186
• Institution: Nanyang Technological University

Many stress related studies and researches are springing up in attempt to diagnose stress accurately. Although stress is a familiar topic, it however is difficult to capture due to the different markers of stress that individual has. Several researches were done to venture into stress recognition models. These models mostly look at physiological indicators such as heartbeat and blood pressure. There is however little research in developing integrated tool for stress monitoring and recognition through human’s electroencephalogram signals. With the advancement of electroencephalography detection tools, adaptable brain wave sensors have mature, and is increasing becoming affordable equipment. In this project, an experiment was designed and carried out with 9 subjects. The Stroop colour-word test was used as a stressor to induce stress in the subjects. The EEG data are recorded to propose an algorithm for stress monitoring. By using fractal dimension, statistical, and power features, with Support Vector Machine as the classifier, four levels of stress can be recognized with an average accuracy of 60.71 %, three levels of stress can be recognized with an accuracy of 69.82%, and two levels of stress can be recognized with an accuracy of 80.96%. The algorithm is integrated into a user interface CogniMeter for the stress state monitoring.