Modeling dynamic brainwave patterns using multi agent technology to effectively identify human emotions

Abstract

Research in interfacing silicon computers (typical computers) with carbon computers (human and animals) has recorded an exponential growth in the modern world. At present, these research ranges from implanting of micro chips inside brain, using brain waves for wireless communication with computers and introducing natural brain cells in silicon computers. Among these trends, use of the brain waves through Electroencephalography signals (EEG) has been a fascinating area of research in the field of computing. In general, EEG Signals are very weak signals with lots of noise unless they are filtered and amplified. This project envisages the use of EEG signals for interfacing our minds with computers. As the primary objective, it has been investigated the nature of real world scenarios or systems for which we can apply the EEG technology for analysis and understanding. We proceed to use EEG in applicable scenarios and model selected real world scenarios, allowing persons to use our computer-based solution for monitoring and guiding people in similar scenarios. EEG patterns detected during a particular scenario have been modeled by Multi Agent Systems technology, which has a proven potential in modeling real world systems with large number of interconnected entities changing over time. In a given scenario, EEG signals are detected from multiple channels with varying strengths. In our design, each channel of EEG signals has been assigned with an agent. Our experiments have shown that brain waves generated for scenarios such as typing in the keyboard, invoking a mouse action or watching television while eating are very noisy, hence it’s difficult to analyze. This has been the case for activities at which people are not adequately mindful. In contrast, experiment shows that activities requiring considerable mindfulness generate sharp EEG patterns. Listening to a lecture, meditation and solving a mathematical problem are some examples for such scenarios. These experimental results have been used to bench mark EEG pattern for given scenario and analyze situations such as the extent a person is engaged in a task, how soon a person gets into the expected mood, at which points a person is distracted. The current system has been tested by considering scenario such as meditation and making a call. The system can be used by people such as those who are curious about their study skills, novice in meditations and those who wish to know about their mental behavior in certain tasks.