Good Vibrations: Tuning a Systems Dynamics Model of Affect and Cognition in Learning to the Appropriate Frequency Bands of Fine-Grained Temporal Sequences of Data: Frequency Bands of Affect and Cognition

Abstract

Process-oriented studies of cooperative learning from an educational neuroscience perspective has not been firmly quantified experimentally. Within a modeling approach aimed at the development of a systems dynamics model of affect and cognition, the goal of this exploratory study is to identify typical timescales of variation for continuous metrics of affect (Frontal Alpha Asymmetry (FAA): valence) and cognition (Cognitive Load (CL); Index of Cognitive Engagement (ICE); Frontal Midline Theta (FMT): attention). These metrics were obtained from 72 participants paired in dyads (player and watcher) from whom electroencephalography (EEG) was recorded for 2 hours while one participant was playing a serious game to learn Physics, and the other one was watching passively. The results show rather slow cyclical variation for every metric tested, accompanied in certain cases by short bursts of faster variations. This result converges with [Newell 1990] cognitive architecture assuming that psychophysiological measures capture activity at higher levels such as operation tasks and operations. Theoretical, methodological and applied implications are discussed. Also, the need for further fine-grained analyses of the context and other atypical analyses are expressed.