Metacognition is the capacity to evaluate the success of one's own cognitive processes in various domains; for example, memory and perception. It remains controversial whether metacognition relies on a domain-general resource that is applied to different tasks or if self-evaluative processes are domain specific. Here, we investigated this issue directly by examining the neural substrates engaged when metacognitive judgments were made by human participants of both sexes during perceptual and memory tasks matched for stimulus and performance characteristics. By comparing patterns of fMRI activity while subjects evaluated their performance, we revealed both domain-specific and domain-general metacognitive representations. Multivoxel activity patterns in anterior prefrontal cortex predicted levels of confidence in a domain-specific fashion, whereas domain-general signals predicting confidence and accuracy were found in a widespread network in the frontal and posterior midline. The demonstration of domain-specific metacognitive representations suggests the presence of a content-rich mechanism available to introspection and cognitive control.SIGNIFICANCE STATEMENT We used human neuroimaging to investigate processes supporting memory and perceptual metacognition. It remains controversial whether metacognition relies on a global resource that is applied to different tasks or if self-evaluative processes are specific to particular tasks. Using multivariate decoding methods, we provide evidence that perceptual- and memory-specific metacognitive representations coexist with generic confidence signals. Our findings reconcile previously conflicting results on the domain specificity/generality of metacognition and lay the groundwork for a mechanistic understanding of metacognitive judgments.

Metacognition is the capacity to evaluate the success of one's own cognitive processes in various domains, e.g. memory and perception. It remains controversial whether metacognition relies on a domain-general resource that is applied to different tasks, or whether self-evaluative processes are domain-specific. Here we directly investigated this issue by examining the neural substrates engaged when metacognitive judgments were made during perceptual and memory tasks matched for stimulus and performance characteristics. By comparing patterns of functional magnetic resonance imaging (fMRI) activity while subjects evaluated their performance, we revealed both domain-specific and domain-general metacognitive representations. Multi-voxel activity patterns in right frontopolar cortex predicted levels of confidence in a domain-specific fashion, whereas domain-general signals predicting confidence and accuracy were found in a widespread network in the frontal and posterior midline. Moreover, individual differences in perceptual and memory metacognitive ability were predicted by the degree of confidence decoding in frontal and parietal cortex, respectively. The demonstration of domain-specific metacognitive representations suggests the presence of a content-rich mechanism available to introspection and cognitive control. Major subject area: Neuroscience 2Metacognition is the capacity to evaluate the success of one's cognitive processes in various do-3 mains, e.g. perception or memory (1-4). Metacognitive ability can be assessed in the laboratory 4 by quantifying the trial-by-trial correspondence between objective performance and subjective 5 confidence (5-8). Anatomical (9-11), functional (12-16) and neuropsychological (1-4,17-19) evi-6 dence indicates specific neural substrates (especially in frontolateral, frontomedial, and parietal 7 regions) contribute to metacognition across a range of task domains, including memory and per-8 ception. However, the neurocognitive architecture supporting metacognition remains controver-9 sial. Does metacognition rely on a common, domain-general resource that is recruited to evaluate 10 performance on a variety of tasks? Or is metacognition supported by domain-specific compo-11 nents? 12Current computational perspectives (20,21) suggest both domain-general and domain-13 specific representations may be important for guiding behavior. On the one hand, one needs to be 14 able to compare confidence estimates in a "common currency" across a range of arbitrary deci-15 sion scenarios (22). One solution to this problem is to maintain a global resource with access to 16 arbitrary sensorimotor mappings (23-25). Candidate neural substrates for a domain-general re-17 source are the frontoparietal and cingulo-opercular networks, known to be involved in arbitrary 18 control operations (23). In particular, the dorsomedial prefrontal cortex (dmPFC; encompassing 19 the paracingulate cortex and pre-supplementary motor area) has been implicated both in repre-20 senting confidence and detecting errors acros...

Arguably the most foundational principle in perception research is that our experience of the world goes beyond the retinal image; we perceive the distal environment itself, not the proximal stimulation it causes. Shape may be the paradigm case of such “unconscious inference”: When a coin is rotated in depth, we infer the circular object it truly is, discarding the perspectival ellipse projected on our eyes. But is this really the fate of such perspectival shapes? Or does a tilted coin retain an elliptical appearance even when we know it’s circular? This question has generated heated debate from Locke and Hume to the present; but whereas extant arguments rely primarily on introspection, this problem is also open to empirical test. If tilted coins bear a representational similarity to elliptical objects, then a circular coin should, when rotated, impair search for a distal ellipse. Here, nine experiments demonstrate that this is so, suggesting that perspectival shapes persist in the mind far longer than traditionally assumed. Subjects saw search arrays of three-dimensional “coins,” and simply had to locate a distally elliptical coin. Surprisingly, rotated circular coins slowed search for elliptical targets, even when subjects clearly knew the rotated coins were circular. This pattern arose with static and dynamic cues, couldn’t be explained by strategic responding or unfamiliarity, generalized across shape classes, and occurred even with sustained viewing. Finally, these effects extended beyond artificial displays to real-world objects viewed in naturalistic, full-cue conditions. We conclude that objects have a remarkably persistent dual character: their objective shape “out there,” and their perspectival shape “from here.”

Whether the prefrontal cortex is part of the neural substrates of consciousness is currently debated. Against prefrontal theories of consciousness, many have argued that neural activity in the prefrontal cortex does not correlate with consciousness but with subjective reports. We defend prefrontal theories of consciousness against this argument. We surmise that the requirement for reports is not a satisfying explanation of the difference in neural activity between conscious and unconscious trials, and that prefrontal theories of consciousness come out of this debate unscathed.

When visual attention is directed away from a stimulus, neural processing is weak and strength and precision of sensory data decreases. From a computational perspective, in such situations observers should give more weight to prior expectations in order to behave optimally during a discrimination task. Here we test a signal detection theoretic model that counter-intuitively predicts subjects will do just the opposite in a discrimination task with two stimuli, one attended and one unattended: when subjects are probed to discriminate the unattended stimulus, they rely less on prior information about the probed stimulus' identity. The model is in part inspired by recent findings that attention reduces trial-by-trial variability of the neuronal population response and that they use a common criterion for attended and unattended trials. In five different visual discrimination experiments, when attention was directed away from the target stimulus, subjects did not adjust their response bias in reaction to a change in stimulus presentation frequency despite being fully informed and despite the presence of performance feedback and monetary and social incentives. This indicates that subjects did not rely more on the priors under conditions of inattention as would be predicted by a Bayes-optimal observer model. These results inform and constrain future models of Bayesian inference in the human brain.

How does the brain give rise to consciousness? A common approach to addressing this question in neuroscience research involves analyzing differences in neural activity in experimental conditions where consciousness of a stimulus differs. However, unless careful measures are taken, conditions that differ in awareness typically also differ in perceptual task performance, e.g. stimulus detection and discrimination. A large body of research demonstrates that task performance and awareness can dissociate, indicating that they are separate mental processes with separate underlying mechanisms. Thus, task performance looms as a potential confound in consciousness science: computational and neural processes attributed to differences in consciousness may actually be better attributed to correlated but distinct differences in task performance. Here we present an extended exploration of the issue of task performance confounds in consciousness research. We describe the approach of performance matching (i.e. creating experimental conditions that yield identical task performance yet different levels of awareness) as a solution to the problem of performance confounds, and discuss why it is not appropriate to artificially match performance by post-hoc selection of trials (e.g. analyzing correct trials only). We review a growing literature demonstrating matched-performance / different-awareness effects using a variety of experimental designs and discuss signal detection theory models that can both explain extant results and guide the design of future research. Finally, we consider caveats and nuances for performance matching approaches and propose that future research could pool across multiple experimental designs with disjoint sets of confounds to triangulate on the confound-free neural substrates of consciousness.

Studying the neural correlates of conscious awareness depends on a reliable comparison between activations associated with awareness and unawareness. One particularly difficult confound to remove is task performance capacity, i.e. the difference in performance between the conditions of interest. While ideally task performance capacity should be matched across different conditions, this is difficult to achieve experimentally. However, differences in performance could theoretically be corrected for mathematically. One such proposal is found in a recent paper by Lamy, Salti and Bar-Haim [Lamy D, Salti M, Bar-Haim Y. Neural correlates of subjective awareness and unconscious processing: an ERP study. J Cognitive Neurosci 2009,21:1435-46], who put forward a corrective method for an electroencephalography experiment. We argue that their analysis is essentially grounded in a version of High Threshold Theory, which has been shown to be inferior in general to Signal Detection Theory. We show through a series of computer simulations that their correction method only partially removes the influence of performance capacity, which can yield misleading results. We present a mathematical correction method based on Signal Detection Theory that is theoretically capable of removing performance capacity confounds. We discuss the limitations of mathematically correcting for performance capacity confounds in imaging studies and its impact for theories about consciousness.

No abstract