Although it tends toward the sci-fi side of tracking eye movement, particularly in it’s discussion of “brain-to-machine interface (BMI),” the Fricke, Sobot, and Dounavis in their article “Analogue portable electrooculargram real-time processor” show that measuring eye movement is feasible on a small scale, opening the possibility of future applications.
The authors looked at a simple electrooculargram with five electrodes: one centered on the forehead, two above either eyebrow to measure vertical movement, and two on the outside of either eye to measure horizontal movement. Although it would be difficult to determine exactly where someone is looking, this instrument should give enough information for the purposes of studying where students are looking in a classroom and what sort of stimuli they respond to, such as the teacher’s gaze and emoting or material on the board. Ideally, five electrodes would not be too invasive or distracting for a period of about an hour, and the low power requirement and real time data would be helpful in the field.
I cannot speak for the technical details of the circuits and signal processing, but the authors show some favorable results. Their primary focus was on clear, precise movements involved in tracking a white dot that moves only horizontally or vertically, which do not match up with the more varied eye movement in a classroom setting, but could be adapted. They note that trained subjects produce clearer results, which is not unexpected. Nevertheless, with some signal boosting and processing, the electrooculargram gives a clear indication of eye movement with even a simple instrument.
Looking at their results, researchers would have to have significant training to interpret them, which might be a significant setback for a real proposal. Unless their instrument can be adapted to produce results that are more intuitive to interpret or can be processed by another program to produce clearer graphs and figures, then the utility of eye tracking instruments might be limited.
Fricke, K., Sobot, R., & Dounavis, A. (n.d). Analogue portable electrooculogram real-time signal processor. International Journal Of Circuit Theory And Applications, 42(2), 195-208.