hide
Free keywords:
-
Abstract:
Cognitive activity is effortful and we tend to avoid cognitively demanding tasks. Despite the profound socioeconomic implication, we still don’t understand the origin of this subjective cost. We propose that cognitive effort can be formalized in the framework of predictive coding. This influential theory posits that the brain develops a hierarchical model of the world and that percepts and actions are shaped by continuous predictions of the incoming sensory information. Hence, in order to maintain our predictions accurate, we have to dynamically update our internal models. Here, we propose that such model updating incurs a certain cost, which, in turn, determines our perception of cognitive effort. We designed a novel category learning task in order to dissect the link between various aspects of model updating on the one hand, and task aversion and effort perception on the other hand. We used standard questionnaires and pupil size recording to evaluate subjective effort. We also used Hierarchical Gaussian Filters to model the evolution of subjects’ beliefs about stimulus category and hence, infer internal model updating. Our preliminary results (N=28) show that effort estimates were considerably higher for high category overlap as compared to low overlap. Given that high category overlap condition was also characterized by increased model updating, these early findings seem to be in accordance with our initial hypothesis.