Metacognition and self-awareness are commonly assumed to operate as global capacities. However, there have been few attempts to test this assumption across multiple cognitive domains and metacognitive evaluations. Here, we assessed the covariance between "online" metacognitive processes, as measured by decision confidence judgments in the domains of perception and memory, and error awareness in the domain of attention to action. Previous research investigating metacognition across task domains have not matched stimulus characteristics across tasks raising the possibility that any differences in metacognitive accuracy may be influenced by local task properties. The current experiment measured metacognition in perceptual, memorial and attention tasks that were closely matched for stimulus characteristics. We found that metacognitive accuracy across the three tasks was dissociated suggesting that domain specific networks support an individual's capacity for accurate metacognition. This finding was independent of objective performance, which was controlled using a staircase procedure. However, response times for metacognitive judgments and error awareness were associated suggesting that shared mechanisms determining how these meta-level evaluations unfold in time may underlie these different types of decision. In addition, the relationship between these laboratory measures of metacognition and reports of everyday functioning from participants and their significant others (informants) was investigated. We found that informant reports, but not self reports, predicted metacognitive accuracy on the perceptual task and participants who underreported cognitive difficulties relative to their informants also showed poorer metacognitive accuracy on the perceptual task. These results are discussed in the context of models of metacognitive regulation and neuropsychological evidence for dissociable metacognitive systems. The potential for the refinement of metacognitive assessment in clinical populations is also discussed.

Objective: Metacognition reflects our capacity to monitor or evaluate other cognitive states as they unfold during task performance, for example, our level of confidence in the veracity of a memory. Impaired metacognition is seen in patients with traumatic brain injury (TBI) and substantially impacts their ability to manage functional difficulties during recovery. Recent evidence suggests that metacognitive representations reflect domain-specific processes (e.g., memory vs. perception) acting jointly with generic confidence signals mediated by widespread frontoparietal networks. The impact of neurological insult on metacognitive processes across different cognitive domains following TBI remains unknown. Method: To assess metacognitive accuracy, we measured decision confidence across both a perceptual and memory task in patients with TBI (n = 27) and controls (n = 28). During the metacognitive tasks, continuous electroencephalography was recorded, and event-related potentials (ERP) were analyzed. Results: First, we observed a deficit in metacognitive efficiency across both tasks suggesting that patients show a loss of perceptual and memorial evidence available for confidence judgments despite equivalent accuracy levels to controls. Second, a late positive-going ERP waveform (500-700 ms) was greater in amplitude for high versus lowconfidence judgements for controls across both task domains. By contrast, in patients with TBI, the same ERP waveform did not vary by confidence level suggesting a deficient or attenuated neural marker of decision confidence postinjury. Conclusions: These findings suggest that diffuse damage to putative frontoparietal regions in patients disrupts domain-general metacognitive accuracy and electrophysiological signals that accumulate evidence of decision confidence. Key PointsQuestion: What are the neurocognitive correlates that underlie impaired self-awareness and confidence (metacognition) in patients with traumatic brain injury (TBI). Findings: We report a domain general deficit in metacognitive ability in TBI patients across two cognitive domains (perception and memory) associated with reduced modulation of a centro-parietal late positivity waveform in TBI patients. Importance: These findings provide insight into how disruption of metacognitive processes may underlie inflexible or perseverative behaviors that are common features following brain injury. Next Steps: Future studies could examine these candidate cognitive and electrophysiological markers in the monitoring of targeted treatment options.