Allocation of Time, EEG-Engagement and EEG-Workload Resources as Scientific Problem Solving Skills Are Acquired in the Classroom

We have studied EEG-derived metrics of Workload (WL) and Engagement (E) as students developed and refined their problem solving approaches to determine the degree to which these are modulated as problem solving experience is gained. The problem solving tasks (IMMEXTM) used for these studies were a series of science and mathematics online simulations designed for middle school students. Comparison of WL and E levels on IMMEXTM tasks and baseline cognitive tasks indicated that the simulations recruited high levels of both WL and E. Detailed second-by second analysis of these metrics during problem solving indicated they were dynamic with cycles of high and low values. Aggregated comparisons of WL and E across students as they gained experience in problem solving showed rapid decreases in time on task while E and particularly WL showed little change. Performances where the solution was missed were significantly lower in WL than when the problem was solved. Analysis of WL and E of individual students showed fluctuations of E with practice with some students showing decreased levels with time and others showing increases. When the levels of WL and E were compared across strategies accounted to be novice or proficient by probabilistic modeling there were no significant differences. These findings indicate that as students practice and refine their problem solving strategies the levels, and changes in the mental effort put into the process is not easily predicted by the changes in the speed of the task, by whether or not the problem was solved, or whether the resulting strategy is more novice or expert.