Catastrophe models for cognitive workload and fatigue

We reconceptualised several problems concerning the measurement of cognitive workload – fixed versus variable limits on channel capacity, work volume versus time pressure, adaptive strategies, resources demanded by tasks when performed simultaneously, and unclear distinctions between workload and fatigue effects – as two cusp catastrophe models: buckling stress resulting from acute workload, and fatigue resulting from extended engagement. Experimental participants completed a task that was intensive on non-verbal episodic memory and had an automatically speeded component. For buckling stress, the epoch of maximum (speeded) performance was the asymmetry parameter; however, anxiety did not contribute to bifurcation as expected. For fatigue, the bifurcation factor was the total work accomplished, and arithmetic, a compensatory ability, was the asymmetry parameter; R 2 for the cusp models outperformed the linear comparison models in both cases. A research programme is outlined that revolves around the two models with different types of task and resource configurations.