Target selection using single electroencephalogram electrode based on mental task

Abstract

Electroencephalogram (EEG)-based BCI studies have been done since early 21st century. The main focus is to improve in areas of neuroscience. It has lead to the use of electrical activity on brain into controlling devices such as wheelchair, cursor movement on computer screen, etc. The purpose of this study is to detect the brain activation on scalp by EEG task-based Brain Computer Interface (BCI) using wireless control robotto determine one scalp location which will give signals that can be used to control the wireless robot without subject training. EEG was measured in 8 normal subjects for control and two tasks conditions. In control condition, subjects were asked to relax but in Task 1 condition, subjects were asked to imagine a star rotating clockwise at position of 45 degree direction pointed by the wireless robot where at this angle the target is located. Same conditions applied for Task 2 but subjects were asked to solve multiplication number 13 without saying or moving, just solve in their mind. At position 0 and 90 degrees angle for both tasks, subjects were asked to relax since there is no target on these directions. Using EEG spectral power (analysis I) and normalization (analysis II), the optimum location for these two tasks were detected and compared. For Task 1, the significant brain activation occurred at position F8, which is atthe frontal cortex area and the significant rhythm was alpha frequency band with P < 0.001 and F(10,70)= 5.31. However, Task 2 which is mental task of solving multiplication number did not give significant results as Task 1(P > 0.05).In conclusion, it is possible to direct wireless robot towards direction of target location using an imaginary star rotation movement without prior training given to subject and with only single electrode.