Humans incorporate attention-dependent uncertainty into perceptual decisions and confidence

Abstract

Perceptual decisions are better when they take uncertainty into account. Uncertainty arises not only from the properties of sensory input but also from cognitive sources, such as different levels of attention. However, it is unknown whether humans appropriately adjust for such cognitive sources of uncertainty during perceptual decision-making. Here we show that, in a task in which uncertainty is relevant for performance, human categorization and confidence decisions take into account uncertainty related to attention. We manipulated uncertainty in an orientation categorization task from trial to trial using only an attentional cue. The categorization task was designed to disambiguate decision rules that did or did not depend on attention. Using formal model comparison to evaluate decision behavior, we found that category and confidence decision boundaries shifted as a function of attention in an approximately Bayesian fashion. This means that the observer’s attentional state on each trial contributed probabilistically to the decision computation. This responsiveness of an observer’s decisions to attention-dependent uncertainty should improve perceptual decisions in natural vision, in which attention is unevenly distributed across a scene.

Limitations of proposed signatures of Bayesian confidence

“…One could object to our critique by questioning the relevance of overlapping stimulus distributions, given that non-overlapping stimulus distributions are the norm in the confidence literature 2,17,18,37 . But although overlapping categories are only just beginning to be used to study confidence 1,7 , such categories have a long history in the perceptual categorization literature 4,13,15,23,24,35,36 . It has been argued that overlapping Gaussian stimulus distributions have several properties that make them more naturalistic than non-overlapping distributions 26 .…”

“…Second, the experimenter must determine how much evidence that provides in favor of the BCH, and whether further investigation is warranted. By contrast, model comparison provides a principled quantity (namely, a log likelihood) in favor of the BCH over some other model 1,2,7 . Given the caveats associated with qualitative signatures, it may be that, as a field, we have no choice but to rely on formal model comparison.…”

“…In this alternative standard signal detection theoretical model 13,17 , observers ignore uncertainty, making confidence only a function of the distance between the measurement and the decision bound. Previous studies have referred to similar models as Fixed 1,7,35 or Difference 3 . VII However, because the orientations were drawn such that the mean orientation of each set was the same for all trials in a category, there was no variance over the stimulus variable of interest (the per-trial mean) within categories.…”

Waving goodbye to contrast: Self-generated hand movements attenuate visual sensitivity

“…It has been shown that these aspects can interact with attention and metacognition (Stein & Peelen, 2017;Matthews et al, 2018). In addition, the precise influence of evidence-reliability remains to be elucidated in more extensive work as it has also been shown to affect confidence judgments and metacognition (Boldt, De Gardelle & Yeung, 2017;Bang & Fleming, 2018;Denison & al., 2018).…”

When temporal attention lacks confidence

“…Attention and confidence have already been considered in the spatial domain, leading to mixed findings. Some studies observed a dissociation between the two (Rahnev et al, 2011(Rahnev et al, , 2012Schoenherr, Leth-Steensen, & Petrusic, 2010;Wilimzig, Tsuchiya, Fahle, Einhäuser, & Koch, 2008), while other works tend to suggest that attention-dependent spatial uncertainty is usually well incorporated into confidence (Denison, Adler, Carrasco, & Ma, 2018;Zizlsperger, Sauvigny, & Haarmeier, 2012;Zizlsperger, Sauvigny, Händel, & Haarmeier, 2014). An open question relates to temporal attention and its link to confidence.…”