Eye Tracking Lab
Location:Department of Psychology
Social Sciences Building
Level 4, Room D403Tel: +44 (0)20 7040 4211
A brief introduction to eye tracking in cognitive neuroscience
Eye tracking provides an objective and quantitative measure of a person's point of gaze in a visual scene displayed e.g on a computer screen. Recording and subsequently analysing even tiniest eye movements when participants perform experimental tasks with varying sensory, motor or cognitive demands prove very revealing with respect to many cognitive and attentional processes investigated in cognitive neuroscience. Eye tracking devices such as the one we are using in our lab collect up to 360 gaze data points per second and therefore allow comprehensive analyses of eye movement characteristics such as fixation time and frequency, length and direction of saccades as a direct function of the experimental task.
What does an eye tracking device look like?
There are different types of eye tracking devices. The one we are using in our lab is a model where all crucial optic components for eye tracking are mounted to a chinrest: An eye camera that captures the features of the eye gaze tracking is based on, an illuminator that allows a good discrimination of the pupil by means of an infra-red beam and a small scence camera that records the visual display that is to be viewed. A small mirror also attached to the chinrest fulfils an important function: Properly positioned slightly below the eye, it allows having the eye camera and the illuminator mounted above the eye (instead of in front of the eye where these components would be in the way).
How does eye tracking work?
Eye tracking takes advantage of the distinct light-reflective properties of two features of the eye: The pupil and the cornea. On its way to the retina, light passes the cornea, a transparent layer that covers the front of the eye, and enters the eye through the pupil, a small translucent opening in the otherwise optically opaque iris. Traversing the gelantinous liquid filling in the inner part of the eye, it eventually reaches the photo-sensitive receptor cells in the retina where the incoming light is converted into neural signals.
The retina at the back of the eye usually reflects a portion of the incoming light back to its source along the very same path the light initially came in. Usually, bright pupil technique the pupil nevertheless appears dark because the location the eye is observed from rarely ever coincides with the location of the light source.
Eye tracking, however, takes advantage of this reflective phenomenon with a careful alignment of the optic components: The illuminator projects a harmless beam of infra-red light into the eye on a path that is coaxial with the imaging direction of the eye camera. Due to the retinal reflection of this light beam, the otherwise dark pupil subsequently appears bright to the eye camera. This allows a robust and reliable discrimination of the pupil from the rest of the eye.
Proper tracking of a person's point of gaze, however, requires the extraction of a second feature of the eye. Once again, it is the infra-red beam from the illuminator that comes into play here: although the cornea is the overall translucent, a small portion of the light beam is nevertheless reflected from it. This corneal reflection (CR) appears as a very small, very bright dot somewhere on the eye image. Ideally, the brightness of the illuminator is adjusted in a way that the pupil can clearly be distinguished from the darker background and the CR from the pupil.
The illuminated pupil is considered the "center" of the eye and shifts whenever the eye moves. The corneal reflection, however, remains in the same location and is therefore used as an anchor point for the head position with respect to the camera. For each eye picture the camera captures during eye tracking, both the pupil and the CR are identified and the angle of the eye with respect to the visual stimuli is computed from these two elements.
What will an eye tracking study involve?
To ensure that eye position and gaze direction are the same, it is crucial that the participant's head position does not vary during stimulus presentation. To this end, participants are asked to put their head into a chinrest. After finding a comfortable seated position with the head in the chinrest, the small mirror attached to the chinrest is positioned so the eye camera captures a complete and sharp close up image of the participants left eye and the illuminator is adjusted in a way that allows an optimal discrimination between the pupil, the CR and the rest of the eye.
Before starting the actual experiment, the eye tracking system also needs to be calibrated. The calibration procedure accounts for individual differences in eye anatomy and consists of collecting point of gaze data for a set of predefined target points in a stationary visual display. Participants are simply asked to look at various target points while the respective eye data get stored in the system. An accurate calibration of the system for each individual participant is crucial for obtaining valid eye movement data.