Human Computer Interface (HCI) using EEG signals.

Abstract

Abstract Human Computer Interface (HCI) is a field which studies about the applied information-processing psychology that occurs when humans interact with computers. A Brain Computer Interface (BCI) is one method to achieve this, and a cursor-controlled application is proposed in this paper. This prototype allows users to control the cursor by converting real-time synaptic commands. An electroencephalogram (EEG) data collection exercise was conducted and the experimental subjects were trained to learn how to control the EMOTIV Insight. Each of them had to decide their mental commands and calibrate the headset before the experiment started. Two different training methods, namely normal training and special training, were applied to the experimental group and control group respectively, and the progress of learning and using the cursor-controlled application were recorded in questionnaires. The control group used the normal training method, which has been widely used in other studies while the experimental group used the special training method, which allows them to decide the content and strategy of mental commands. The objective of this project is to have all the subjects from both experimental groups control 2 commands with an accuracy above 80% after 4 hours of training and observe which group gets better results. It was a randomized controlled trial (RCT) and double-blind trial, and one of the subjects turned his strategy from control group to experimental group during the experiment without informing the experimenter in advanced. In the outcome, the highest accuracy obtained by the 3 subjects in the experimental group are 50%, 81.25% and 100% respectively, and the highest accuracy obtained by the subject in the control group is 62.5%. In short, the accuracy of the experimental group is averagely higher compared to the control group, and 66.6% of the experimental group managed to control 2 commands with an accuracy above 80%.