Research Reports

• Research Reports Home
• UC Research Repository
• Honours Reports
• Reports Manager (CSSE Staff Only)

HONS 01/12

Investigating the Emotiv EPOC for cognitive control in liminted training time

Matthew Lang
Department of Computer Science and Software Engineering
University of Canterbury

Abstract

Brain-computer interfaces (BCI) that utilise electroencephalography (EEG) have been studied for many years as a means of communication and control for physically disabled individuals. Through training, people can learn to use imagined motor movements as computer input or to control assistive technology. The Emotiv EPOC is an inexpensive, lightweight, wireless BCI headset, and provides systems for control and aective measurement. Some studies have successfully used this device for control, however most fail to indicate the time and effort required to achieve the reported results. As an initial step to provide a quantification of the training time required, we ran two studies investigating cognitive control. Using just their mind, participants in the first experiment moved a virtual cube left and right, and in the latter made three-choice selections. The first study employed a fixed training scheme approximating 11 minutes, and resulted in a 36% average success rate. The latter task allowed 15 minutes of self-directed training, increasing the average success to 46.8%. We detail our investigations with the EPOC, including analysis of detection filters for maximising cognitive control, and we review the Aectiv suite and how to analyse the user's emotional measurements it provides. Future applications should be aware that a more practical level of control requires a greater training time than one short session.