Abstract
Post-error increases in accuracy (PEIA) refers to the increase in the probability of making a correct perceptual decision following an error, and post-uncertainty increases in accuracy (PUIA) refers to the increase in the probability of a making a correct perceptual decision following an uncertain perceptual decision. Although there is ample evidence for both PEIA and PUIA, it is not clear whether they are the result of perceptual or decisional processes. For this reason, in this study we evaluated PEIA and PUIA using a multidimensional extension of signal detection theory called general recognition theory (GRT). To this end, we presented participants with vertically or non-vertically oriented Gabor patches, and asked them to judge each Gabor patch as vertical or non-vertical (discrimination task) and to report their level of confidence on their discrimination decision using two alternatives (Certain vs. Uncertain. confidence task). After collecting the behavioral data, we fit two GRT models to evaluate PEIA and PUIA, separately. To evaluate PEIA, we fit a GRT model with orientation of Gabor patch in trials N and N-1 as the dimensions of interest. To evaluate PUIA, we fit a GRT model with orientation of Gabor patch in trial N and participants confidence judgment in trial N-1 as the dimensions of interests. According to both analyses, PEIA and PUIA are the result of decisional effects. That is, the increase in the likelihood of making a correct decision following an inaccurate or an uncertain perceptual decision is due to changes in decisional criterion and not due to changes in perceptual sensitivity. Collectively, these results suggest that the human cognitive control system strategically adapts perceptual responses in similar ways following either errors or correct-but uncertain-perceptual responses. In both instances, individuals appear to engage in a shift towards a more cautious response strategy, as opposed to perceptual modulations.