Electrooculogram detection of eye movements on gaze displacement.

Changes in potential are known to occur in the orbital area during saccades. The sign of these changes depends on the position of the electrode and the direction of eye rotation, while their amplitude depends on the rotation angle. The patterns of potentials can be used to resolve the reverse task, i.e., that of describing the gaze trajectory, such that the eye can be used to control a computer in an on-line regime. This requires a screen cursor to be placed at the calculated gaze fixation point, i.e., the point at which the observer is looking. Electrodes beneath the eyes were used to assess the vertical component of gaze displacement, while side electrodes at the corners of the orbit were used to assess the horizontal component. Sharp unipolar changes in potential occurring on saccades were apparent as steps which could be detected and measured. The signal was digitally filtered using a complex filter constructed by ourselves. The ratio of the amplitudes at the four points was then used to calculate the direction and angle of eye rotation. Characteristic changes in potential during spontaneous blinking were identified and ignored. Voluntary blinks of one eye were used to simulate mouse clicks. Subjects were initially told to make changes in gaze through specified angles in eight directions. This allowed calibration of standard saccades (in microV). Calibration curves were used to resolve the reverse task - changes in potential (in microV) were used to calculate the point on the screen (the pixel) to which the cursor was to be moved. Subjects were then trained to control the cursor using their eyes, and control of the computer was then completely handed over to the subject. The apparatus described here provides a brain-computer interface. Some interesting data on eye coordination were obtained during these studies: saccades were preceded by short negative electrooculogram (EOG) potentials lasting 10-15 msec. With age, the amplitude of saccade-related EOG potentials decreased. When gaze was shifted to the left, deviation of the eyes was more significant than when gaze was shifted to the right, while on shifting of gaze to the right, the lateral deviations of the eyes were similar. On diagonal right-down and left-up movements, right eye skew was greater than left eye skew, while on right-up and left-down movements, left eye skew was greater than right eye skew. Differences in eye coordination between genders were minor.