Research Directions for Methodological Improvement of the Statistical Analysis of Electroencephalography Data Collected in NeuroIS

This proposed research will study and improve the statistical methodology used with neurophysiological data collected from subjects using information systems (IS). This research thus aims to provide guidelines and propose new statistical models constructed explicitly for the analysis of electroencephalography (EEG) data in IS research, where the number of EEG trials is often limited to preserve the ecological validity of the experiment. Two new modeling strategies are proposed: first, we will model explicitly the correlation between repeated trials by finding appropriate correlation structures. Secondly, we will reduce the measurement's error by using explicitly the cyclic behavior of an electrical brain signal. These new models will then be taken into account to derive new formulas for sample size determination.