Implementation and use of the electrooculogram in sleep monitoring

This paper describes the semester group project for the North Carolina State University Biomedical Instrumentation class, BME/ECE 522 from the spring semester of 2013. The topic of sleep, and more specifically Rapid Eye Movement (REM) during sleep, was selected by the group, and from that, the electrooculogram was chosen for its particular focus on nonintrusive eye movement detection. Knowledge of the theory behind electrooculography is widely available; however, the consumer market does not favor the electrooculogram, instead preferring electroencephalography or other approaches such as inferring sleep stages from motion. This paper presents the stages involved in our process of understanding the theory and applying it to create a prototype electrooculogram sleep mask. Our sleep mask obtains electrical recordings from the face and, following analog filtering, records this signal via an analog to digital converter. Later, a MATLAB algorithm uses Fourier analysis on this data to determine when REM sleep occurred. We observed that REM detection in the recordings from this prototype sleep mask correlate well with a commercial electroencephalogram device: the Zeo sleep monitor. The results show that the electrooculogram is a promising approach to REM sleep detection.