Eye-tracking during Dichoptic Presentation
Eye-tracking during dichoptic presentation conditions poses distinct challenges. If a stereoscope is used, the chin rest that supports the mirror mount might be too narrow for a head-mounted eye-tracker (e.g. SR Research EyeLink 1000). HeadSpot chin rest, for example, measures only about 6.5 inches wide. One solution is to construct a mirror support system that is either separate from the chin rest, or has bars wide enough to accommodate the eye-tracker.
Here, the cameras of the eye-tracker are about 4 cm away from the stereoscope. The metal bar on which the mirrors are mounted is 30 cm wide, and the space between the vertical bars on the wood platform is about 33 cm. Another alternative to this set up is hanging the stereoscope upside down (after rearranging its mirrors), and using a stationary eye-tracker like Tobii.
Dichoptic presentation can also be accomplished with the use of prisms. A nice video demonstration of prisms is available here. In this scenario eye-tracking occurs like in a typical experiment, except that calibration should be done using a single eye. The photo below shows how prisms would be used with an eye-tracker; in an actual experiment a chin rest and a divider between the eyes would also be employed.
If prisms are used, lower diopter prisms are generally preferred. First, it is easier to maintain fusion and hence there is less eye fatigue with lower diopter prisms. We find that least amount of fatigue is reported during the initial 20 minutes of a study when using 10 PD lenses. Also, higher diopter prisms tend to produce color fringes on chromatic stimuli; these artifacts become more noticeable the higher the prismatic deviation is.