The Emotiv mind: Investigating the accuracy of the Emotiv EPOC in identifying emotions and its use in an Intelligent Tutoring System
Tegan Harrison
Department of Computer Science and Software Engineering
University of Canterbury
Abstract
Emotion is one of the key elements involved in learning and education; it affects our decision making, our communication and our ability to learn. Therefore in this project we proposed the use of an electroencephalography (EEG) device called the Emotiv EPOC to detect a user's emotions via brainwaves in an Intelligent Tutoring System (ITS). We aimed to test the accuracy and applicability of this device in an ITS called EER-Tutor. First we aimed to test the EPOC's accuracy in detecting emotions by comparing EPOC's emotional values with participants' self-reported scores. Emotions were induced using the International Affective Picture System (IAPS). Due to a lack of reliable data this study was then improved upon and re-conducted. The improved study still resulted in no significant findings as a relationship between self-report scores and EPOC emotional values could not be found. Therefore a different approach was adopted to test the accuracy and applicability of the EPOC. In our third study we tested the EPOC in EER-Tutor using a think-aloud protocol. The utterances of participants were compared to the emotional values EPOC produced. These utterances were defined by context of statement, action taken and timing and placed into categories. No significant difference was found between categories. The variances of the readings for various categories were also quite large, showing that individuals emotional responses can vary greatly and a single emotional model cannot be developed for a group of people. Even though no significant findings were found, certain limitations of the EPOC allowed us conclude its unsuitability for use in ITSs.
