abstract
-
Mental fatigue is usually accompanied by drops in task performance and reduced willingness for further exertion. A value-based theoretical account may help to explain such negative effects. In this view, mental fatigue influences perceived costs and rewards of exerting effort. However, no formal mathematical framework has yet been proposed to model and quantitatively estimate the effects of mental fatigue on subjective evaluations of effort expenditure, subject to possibly imperfect self-perceptions of internal fatigue states. We proposed a mathematical framework to model human cognitive effort allocations, assuming mental fatigue states are partially observable with semi-Markov dynamics. We modeled effort allocation decisions as consistent with the goal of maximizing cumulative subjective values over a given time horizon. We analyzed the proposed model structure and developed an estimation method to identify subjective values and hidden fatigue dynamics, which can be based on self-reports, psychophysiological indices, and behavioral effects associated with fatigue. The modeling and estimation method was tested using a simulated n-back task under a free choice paradigm, with model parameters fine-tuned from past studies. The proposed approach was able to recapitulate task performance and engagement patterns observed under mental fatigue. This work advances a reward/cost trade-off account for explaining the principles of effort exertion and suggests new avenues for both theoretically and empirically relevant understandings of how cognitive operations are affected by mental fatigue.