Suprathreshold perceptual decisions constrain models of confidence

Priors for natural image statistics inform confidence in perceptual decisions

Decision confidence plays a critical role in humans' ability to make adaptive decisions in a noisy perceptual world. Despite its importance, there is currently little consensus about the computations underlying confidence judgements in perceptual decisions. To better understand these mechanisms, we addressed the extent to which confidence is informed by a naturalistic prior distribution. Contrary to previous research, we did not require participants to internalise parameters of an arbitrary prior distribution. We instead used a novel psychophysical paradigm leveraging probability distributions of low-level image features in natural scenes, which are well-known to influence perception. Participants reported the subjective upright of naturalistic image patches, targets, and then reported their confidence in their orientation responses. We used computational modelling to relate the statistics of the low-level features in the targets to the average distribution of these features across many naturalistic images, a prior. Our results showed that participants' perceptual and importantly, their confidence judgments aligned with an internalised prior for image statistics. Overall, our study highlights the importance of naturalistic task designs that capitalise on existing, long-term priors to further understand the computational basis of confidence.

Challenging the Bayesian confidence hypothesis in perceptual decision-making

The Bayesian confidence hypothesis (BCH), which postulates that confidence reflects the posterior probability that a decision is correct, is currently the most prominent theory of confidence. Although several recent studies have found evidence against it in the context of relatively complex tasks, BCH remains dominant for simpler tasks. The major alternative to BCH is the confidence in raw evidence space (CRES) hypothesis, according to which confidence is based directly on the raw sensory evidence without explicit probability computations. Here, we tested these competing hypotheses in the context of perceptual tasks that are assumed to induce Gaussian evidence distributions. We show that providing information about task difficulty gives rise to a basic behavioral signature that distinguishes BCH from CRES models even for simple 2-choice tasks. We examined this signature in three experiments and found that all experiments exhibited behavioral signatures in line with CRES computations but contrary to BCH ones. We further performed an extensive comparison of 16 models that implemented either BCH or CRES confidence computations and systematically differed in their auxiliary assumptions. These model comparisons provided overwhelming support for the CRES models over their BCH counterparts across all model variants and across all three experiments. These observations challenge BCH and instead suggest that humans may make confidence judgments by placing criteria directly in the space of the sensory evidence.

Audiovisual simultaneity windows reflect temporal sensory uncertainty

The ability to judge the temporal alignment of visual and auditory information is a prerequisite for multisensory integration and segregation. However, each temporal measurement is subject to error. Thus, when judging whether a visual and auditory stimulus were presented simultaneously, observers must rely on a subjective decision boundary to distinguish between measurement error and truly misaligned audiovisual signals. Here, we tested whether these decision boundaries are relaxed with increasing temporal sensory uncertainty, i.e., whether participants make the same type of adjustment an ideal observer would make. Participants judged the simultaneity of audiovisual stimulus pairs with varying temporal offset, while being immersed in different virtual environments. To obtain estimates of participants' temporal sensory uncertainty and simultaneity criteria in each environment, an independent-channels model was fitted to their simultaneity judgments. In two experiments, participants' simultaneity decision boundaries were predicted by their temporal uncertainty, which varied unsystematically with the environment. Hence, observers used a flexibly updated estimate of their own audiovisual temporal uncertainty to establish subjective criteria of simultaneity. This finding implies that, under typical circumstances, audiovisual simultaneity windows reflect an observer's cross-modal temporal uncertainty.

A theory of visibility measures in the dissociation paradigm

Research on perception without awareness primarily relies on the dissociation paradigm, which compares a measure of awareness of a critical stimulus (direct measure) with a measure indicating that the stimulus has been processed at all (indirect measure). We argue that dissociations between direct and indirect measures can only be demonstrated with respect to the critical stimulus feature that generates the indirect effect, and the observer's awareness of that feature, the critical cue. We expand Kahneman's (Psychological Bulletin, 70, 404-425, 1968) concept of criterion content to comprise the set of all cues that an observer actually uses to perform the direct task. Different direct measures can then be compared by studying the overlap of their criterion contents and their containment of the critical cue. Because objective and subjective measures may integrate different sets of cues, one measure generally cannot replace the other without sacrificing important information. Using a simple mathematical formalization, we redefine and clarify the concepts of validity, exclusiveness, and exhaustiveness in the dissociation paradigm, show how dissociations among different awareness measures falsify both single-valued measures and monocausal theories of "consciousness," and formulate the demand that theories of visual awareness should be sufficiently specific to explain dissociations among different facets of awareness.

Introspective inference counteracts perceptual distortion

Introspective agents can recognize the extent to which their internal perceptual experiences deviate from the actual states of the external world. This ability, also known as insight, is critically required for reality testing and is impaired in psychosis, yet little is known about its cognitive underpinnings. We develop a Bayesian modeling framework and a psychophysics paradigm to quantitatively characterize this type of insight while people experience a motion after-effect illusion. People can incorporate knowledge about the illusion into their decisions when judging the actual direction of a motion stimulus, compensating for the illusion (and often overcompensating). Furthermore, confidence, reaction-time, and pupil-dilation data all show signatures consistent with inferential adjustments in the Bayesian insight model. Our results suggest that people can question the veracity of what they see by making insightful inferences that incorporate introspective knowledge about internal distortions.

Toward 'Computational-Rationality' Approaches to Arbitrating Models of Cognition: A Case Study Using Perceptual Metacognition

Perceptual confidence results from a metacognitive process which evaluates how likely our percepts are to be correct. Many competing models of perceptual metacognition enjoy strong empirical support. Arbitrating these models traditionally proceeds via researchers conducting experiments and then fitting several models to the data collected. However, such a process often includes conditions or paradigms that may not best arbitrate competing models: Many models make similar predictions under typical experimental conditions. Consequently, many experiments are needed, collectively (sub-optimally) sampling the space of conditions to compare models. Here, instead, we introduce a variant of optimal experimental design which we call a computational-rationality approach to generative models of cognition, using perceptual metacognition as a case study. Instead of designing experiments and post-hoc specifying models, we began with comprehensive model comparison among four competing generative models for perceptual metacognition, drawn from literature. By simulating a simple experiment under each model, we identified conditions where these models made maximally diverging predictions for confidence. We then presented these conditions to human observers, and compared the models' capacity to predict choices and confidence. Results revealed two surprising findings: (1) two models previously reported to differently predict confidence to different degrees, with one predicting better than the other, appeared to predict confidence in a direction opposite to previous findings; and (2) two other models previously reported to equivalently predict confidence showed stark differences in the conditions tested here. Although preliminary with regards to which model is actually 'correct' for perceptual metacognition, our findings reveal the promise of this computational-rationality approach to maximizing experimental utility in model arbitration while minimizing the number of experiments necessary to reveal the winning model, both for perceptual metacognition and in other domains.

On why we lack confidence in some signal-detection-based analyses of confidence

Signal-detection theory (SDT) is one of the most popular frameworks for analyzing data from studies of human behavior - including investigations of confidence. SDT-based analyses of confidence deliver both standard estimates of sensitivity (d'), and a second estimate informed by high-confidence decisions - meta d'. The extent to which meta d' estimates fall short of d' estimates is regarded as a measure of metacognitive inefficiency, quantifying the contamination of confidence by additional noise. These analyses rely on a key but questionable assumption - that repeated exposures to an input will evoke a normally-shaped distribution of perceptual experiences (the normality assumption). Here we show, via analyses inspired by an experiment and modelling, that when distributions of experience do not conform with the normality assumption, meta d' can be systematically underestimated relative to d'. Our data highlight that SDT-based analyses of confidence do not provide a ground truth measure of human metacognitive inefficiency. We explain why deviance from the normality assumption is especially a problem for some popular SDT-based analyses of confidence, in contrast to other analyses inspired by the SDT framework, which are more robust to violations of the normality assumption.

Modality independent or modality specific? Common computations underlie confidence judgements in visual and auditory decisions

The mechanisms that enable humans to evaluate their confidence across a range of different decisions remain poorly understood. To bridge this gap in understanding, we used computational modelling to investigate the processes that underlie confidence judgements for perceptual decisions and the extent to which these computations are the same in the visual and auditory modalities. Participants completed two versions of a categorisation task with visual or auditory stimuli and made confidence judgements about their category decisions. In each modality, we varied both evidence strength, (i.e., the strength of the evidence for a particular category) and sensory uncertainty (i.e., the intensity of the sensory signal). We evaluated several classes of computational models which formalise the mapping of evidence strength and sensory uncertainty to confidence in different ways: 1) unscaled evidence strength models, 2) scaled evidence strength models, and 3) Bayesian models. Our model comparison results showed that across tasks and modalities, participants take evidence strength and sensory uncertainty into account in a way that is consistent with the scaled evidence strength class. Notably, the Bayesian class provided a relatively poor account of the data across modalities, particularly in the more complex categorisation task. Our findings suggest that a common process is used for evaluating confidence in perceptual decisions across domains, but that the parameter settings governing the process are tuned differently in each modality. Overall, our results highlight the impact of sensory uncertainty on confidence and the unity of metacognitive processing across sensory modalities.

Feeling lucky? prospective and retrospective cues for sensorimotor confidence

On a daily basis, humans interact with the outside world using judgments of sensorimotor confidence, constantly evaluating our actions for success. We ask, what sensory and motor-execution cues are used in making these judgements and when are they available? Two sources of temporally distinct information are prospective cues, available prior to the action (e.g., knowledge of motor noise and past performance), and retrospective cues specific to the action itself (e.g., proprioceptive measurements). We investigated the use of these two cues in two tasks, a secondary motor-awareness task and a main task in which participants reached toward a visual target with an unseen hand and then made a continuous judgment of confidence about the success of the reach. Confidence was reported by setting the size of a circle centered on the reach-target location, where a larger circle reflects lower confidence. Points were awarded if the confidence circle enclosed the true endpoint, with fewer points returned for larger circles. This incentivized accurate reaches and attentive reporting to maximize the score. We compared three Bayesian-inference models of sensorimotor confidence based on either prospective cues, retrospective cues, or both sources of information to maximize expected gain (i.e., an ideal-performance model). Our findings showed two distinct strategies: participants either performed as ideal observers, using both prospective and retrospective cues to make the confidence judgment, or relied solely on prospective information, ignoring retrospective cues. Thus, participants can make use of retrospective cues, evidenced by the behavior observed in our motor-awareness task, but these cues are not always included in the computation of sensorimotor confidence.

Confidence reflects a noisy decision reliability estimate

Decisions vary in difficulty. Humans know this and typically report more confidence in easy than in difficult decisions. However, confidence reports do not perfectly track decision accuracy, but also reflect response biases and difficulty misjudgements. To isolate the quality of confidence reports, we developed a model of the decision-making process underlying choice-confidence data. In this model, confidence reflects a subject's estimate of the reliability of their decision. The quality of this estimate is limited by the subject's uncertainty about the uncertainty of the variable that informs their decision ('meta-uncertainty'). This model provides an accurate account of choice-confidence data across a broad range of perceptual and cognitive tasks, investigated in six previous studies. We find meta-uncertainty varies across subjects, is stable over time, generalizes across some domains and can be manipulated experimentally. The model offers a parsimonious explanation for the computational processes that underlie and constrain the sense of confidence.