Cognitive Workload and Fatigue Dynamics in a Chaotic Forecasting Task

Many real-world tasks require people to forecast chaotic events in order to take adaptive action. This ability is considered rare, and less understood than other cognitive processes. The present study examined how the performance dynamics in a chaotic forecasting task would be affected by stressors such as cognitive workload and fatigue using two cusp catastrophe models. Participants were 147 undergraduates who were shown graphs and brief chaotic number series for which they needed to forecast the next four values. Performance data were complemented by variables known to represent cognitive elasticity versus rigidity, compensatory abilities for fatigue, and NASA TLX ratings of subjective workload. R2 for the workload cusp was .56, which compared favorably to the next best linear alternative model (.12); it contained six bifurcation variables and three measures of workload (asymmetry). R2 for the fatigue cusp was .54, which also compared favorably to the next best linear alternative (.07); it contained one bifurcation variable and two compensatory abilities. The role of field independence as an elasticity variable in the workload model and as a compensatory ability in fatigue was particularly noteworthy. Several elasticity-rigidity variables have now been identified over a series of studies. They appear to be operating in unison to produce a bifurcation effect, and different variables become salient depending on the task. Future research should consider how the ability to forecast chaos and its susceptibility to workload and fatigue carry over to dynamical decisions made while managing a complex system. Key Words.