Meta-analytic evidence (Vaccaro and Fleming, 2018) analyzing 47 neuroimaging studies identified consistent activation in the medial and lateral prefrontal cortex during metacognitive judgments. The right rostrolateral prefrontal cortex (rlPFC) demonstrates particular importance, showing increased functional connectivity with both contralateral prefrontal regions and task-specific sensory areas during metacognitive reports (Fleming et al., 2012).

Individual differences in metacognitive ability correlate with gray matter volume in the anterior prefrontal cortex (Fleming et al., 2010). Recent transcranial magnetic stimulation studies provide causal evidence for prefrontal involvement: disruption of lateral PFC impairs metacognitive accuracy while preserving first-order task performance (Lapate et al., 2020; Miyamoto et al., 2021). Importantly, different prefrontal subregions contribute to distinct aspects of metacognition (Schurz and Perner, 2015), with more recent work showing the frontal lobe processes difficult metacognitive judgments by reading out performance information from first-order task regions (Soutschek and Mattes, 2025).

The dorsal anterior cingulate cortex (dACC) plays a critical role in monitoring cognitive processes and detecting conflicts. Carter et al. (1998) demonstrated that the ACC activates not only during errors but also during correct responses under high response competition, indicating a monitoring function rather than simple error detection. Subsequent research identified a dACC network specifically involved in metacognitive uncertainty (Qiu et al., 2018a,b; Su et al., 2022).

Meta-analytic evidence confirms that dACC and bilateral anterior insula function as an integrated network hub for self-regulation (Zhang and Peng, 2023). Brain stimulation studies using transcranial direct current stimulation support the causal role of dACC in metacognitive bias (Soutschek and Mattes, 2025). Notably, the ACC demonstrates structural plasticity, with morphological changes related to learning and practice (Newman and McGaughy, 2011).

Posterior parietal regions, particularly the precuneus and inferior parietal lobule, consistently activate during metacognitive judgments across domains (Vaccaro and Fleming, 2018). The parietal cortex may accumulate sensory evidence and confidence-related signals that inform metacognitive decisions. Effective connectivity analyses suggest information flow from posterior to anterior regions during confidence judgments, consistent with hierarchical processing models.

While traditionally viewed as a memory structure, the hippocampus contributes to metacognitive monitoring across domains. Connection strength between hippocampus and precuneus correlates with confidence ratings. Hippocampal microstructure relates to metacognitive ability even in perceptual tasks (Allen et al., 2017).

Ren et al. (2018) demonstrated that effective connectivity from anterior hippocampus to precuneus during stimulus viewing predicts subsequent metacognitive confidence in voluntary recall. Intracranial recordings reveal distinct communication patterns between hippocampus and precuneus during metacognition (Chen et al., 2024). Resting-state connectivity studies confirm that increased hippocampal-precuneus connectivity associates with metamemory ability (Ye et al., 2019).

These findings collectively suggest that metacognitive processes emerge from interactions across distributed networks rather than from discrete localized regions. The specific network configurations engaged likely depend on task domain (perceptual, memory, problem-solving), cognitive load, and individual differences in metacognitive skill.

Metacognitive impairments manifest across a spectrum of neurological and psychiatric conditions, with deficits varying in severity, domain specificity, and amenability to intervention (Prigatano and Altman, 1990; Agnew and Morris, 1998).

In traumatic brain injury (TBI), metacognitive deficits are common and predict functional outcomes more strongly than objective cognitive impairments (Ownsworth et al., 2019). A meta-analysis of 22 studies found that better metacognitive knowledge following acquired brain injury correlated with improved quality of life, family integration, and work outcomes. These deficits span multiple awareness levels, with patients often demonstrating relatively preserved intellectual awareness while showing marked impairments in emergent and anticipatory awareness.

In schizophrenia, metacognitive deficits represent a core feature affecting multiple domains including self-reflectivity, understanding of others’ minds, and integrated sense of identity (Lysaker et al., 2018). These deficits correlate with negative symptoms, functional outcomes, and subjective recovery more strongly than do neurocognitive deficits alone. Neuroimaging studies reveal associations between metacognitive capacity and prefrontal-temporal connectivity (Sapara et al., 2014).

Depression and anxiety disorders show characteristic patterns of metacognitive dysfunction, particularly in the form of maladaptive metacognitive beliefs, worry and rumination. These beliefs include positive beliefs about worry (“worrying helps me cope”) and negative beliefs about uncontrollability (“my worrying could make me go mad”). Such beliefs maintain cognitive-attentional syndrome patterns including repetitive negative thinking, threat monitoring, and maladaptive coping strategies (Wells, 2009).

Anosognosia represents the most dramatic example of metacognitive impairment, causing a complete lack of awareness of neurological deficits. Occurring in 10–18% of acute stroke patients with hemiparesis (Prigatano and Schacter, 1991), anosognosia demonstrates how focal brain damage can eliminate self-awareness while preserving other cognitive functions.

Neuroanatomical studies consistently associate anosognosia with right hemisphere lesions affecting parietal lobe, temporoparietal cortex, insular cortex, and thalamus (Orfei et al., 2010). These regions appear critical for integrating sensory, motor, and cognitive information into coherent self-awareness. The dissociation between preserved motor planning and absent awareness of paresis in anosognosia patients suggests that self-monitoring mechanisms operate independently from action systems.

The Cognitive Awareness Model (Morris and Mograbi, 2013) proposes that anosognosia results from failures in comparator mechanisms that normally detect mismatches between intended and actual performance, potentially modulated by emotional/motivational factors that make awareness psychologically intolerable.

Left hemisphere damage affecting temporal language areas produces Wernicke’s aphasia leading to fluent but often semantically empty speech with severely impaired comprehension. Many patients show limited awareness of their communication failures, providing insight into domain-specific metacognitive mechanisms for language monitoring.

Recent therapeutic research demonstrates the potential for intervention even in severe cases. Raymer et al. (2024) tested a metacognitive intervention in two participants with severe Wernicke’s aphasia, finding improvements in naming and discourse skills. This suggests that metacognitive awareness can be partially restored through targeted training even when language processing is severely compromised. The mechanisms likely involve engaging preserved monitoring systems and establishing compensatory strategies for error detection.

Progressive neurodegenerative conditions produce evolving metacognitive deficits with characteristic temporal patterns. In mild cognitive impairment (MCI) and early Alzheimer’s disease, anosognosia frequency ranges from 20 to 30% (Orfei et al., 2010), with severity correlating with medial temporal and prefrontal atrophy.

Interestingly, metamemory judgments in early stages can show relative preservation despite objective memory decline, suggesting that monitoring mechanisms can operate partially independently from the cognitive functions they monitor. As disease progresses, metacognitive deficits typically worsen, affecting both monitoring accuracy and strategic regulation.

Assessment of metacognitive function requires multiple complementary approaches, each capturing different aspects of metacognitive capacity:

No single measure captures all dimensions of metacognitive function. Comprehensive assessment requires selecting methods matched to the specific clinical question, cognitive domain of interest, and practical constraints of the assessment context (Toglia and Kirk, 2000).

The taxi driver studies demonstrate that intensive spatial learning engaging flexible navigation produces measurable neural reorganization. While not involved in explicit or direct metacognitive training, successful navigation planning and execution of the routines required to develop the skill involves metacognitive processes such as monitoring one’s confidence about route knowledge, regulating search strategies, evaluating whether a route selection was optimal, an various self regulatory functions tied to metacognitive monitoring.

These studies provide proof-of-principle that:

Whether metacognitive training specifically (targeting monitoring and regulation processes) produces comparable neural changes remains an open empirical question, but the involvement of metacognitive processes in coordinating the behaviors necessary to produce meaningful morphological change in specific brain regions to achieve specific cognitive function is clear.

Exercise promotes β-hydroxybutyrate production, which modulates histone deacetylases (HDAC2/HDAC3) to increase BDNF expression, supporting neurogenesis, synaptic plasticity, and cognitive function (Seilman, 2016). Meta-analysis of 22 studies with 552 participants showed that high-intensity exercise significantly increases peripheral BDNF compared to light exercise (effect size = 0.78, p < 0.001) (Fernández-Rodríguez et al., 2022). Importantly, these acute exercise effects persist in older adults (Nilsson et al., 2020).

Mindfulness practices explicitly cultivate metacognitive awareness—observing one’s thoughts and mental states without judgment. Research shows meditation produces measurable neural changes including increased cortical thickness and enhanced connectivity (Calderone et al., 2024). Magnetoencephalography studies of long-term Vipassana meditators reveal significantly higher resting-state hippocampal connectivity (theta band, p = 0.009) compared to controls (Martorano et al., 2019), potentially protecting against age-related cognitive decline.

While exercise and meditation engage neuroplastic mechanisms, the specific causal pathway from these interventions through metacognitive enhancement to structural brain changes remains incompletely characterized. Several potential mechanisms warrant investigation:

Current evidence supports these interventions as neuroplasticity-promoting approaches, but establishing direct causal links to metacognitive enhancement specifically requires studies measuring metacognitive outcomes alongside neural changes with appropriate experimental controls.

Developed by Wells (2009) and Wells and Purdon (1999), MCT focuses on modifying metacognitive beliefs and processes maintaining psychological distress. Unlike cognitive-behavioral therapy (CBT), which targets thought content, MCT addresses how individuals respond to thoughts, targeting the cognitive-attentional syndrome. These are patterns of worry, rumination, and threat monitoring driven by metacognitive beliefs.

Randomized controlled trials provide compelling evidence for efficacy. The largest trial comparing MCT to CBT in 174 adults with major depression (Callesen et al., 2020) found MCT demonstrated superior outcomes: recovery rates of 74% versus 52% for CBT at post-treatment (odds ratio = 2.42, p = 0.014), with significant differences on the Beck Depression Inventory II favoring MCT [−5.49, 95% CI (−8.90 to −2.08), p = 0.002]. Benefits persisted at 6-month follow-up (74% vs. 56% recovery), with MCT requiring fewer sessions on average (5.5 vs. 6.7 sessions). Similar patterns emerge for generalized anxiety disorder (Nordahl et al., 2018).

Developed by Lysaker et al. (2018) and Lysaker et al. (2019) within the Integrative Model of Metacognition, MERIT targets metacognitive capacity as a spectrum of abilities for forming integrated representations of self, others, and mental states. Primarily developed for schizophrenia, MERIT emphasizes narrative development, perspective-taking, and integration of fragmented self-experiences.

Research demonstrates that metacognitive capacity predicts functional outcomes in schizophrenia independent of neurocognition (Lysaker et al., 2018). MERIT aims to enhance: (1) self-reflectivity (recognizing and labeling internal states), (2) awareness of others’ mental states, (3) decentration (recognizing mental states are interpretations), and (4) mastery (using metacognitive knowledge for problem-solving).

Preliminary trials show promise (Lysaker et al., 2018), suggesting moderate effect sizes for functional outcomes. Neuroimaging studies link metacognitive capacity improvements with changes in prefrontal-temporal connectivity (Sapara et al., 2014), though dedicated neuroimaging studies of MERIT specifically are needed.

While both approaches demonstrate clinical efficacy, important limitations exist regarding neuroplastic mechanisms. Most MCT trials have not included neuroimaging protocols, so claims about engaging neuroplastic mechanisms remain speculative. The sustained benefits observed in follow-up studies (Callesen et al., 2020) are consistent with but do not prove lasting neural reorganization.

Recent studies examining related metacognitive interventions provide suggestive evidence. Shan et al. (2021) found that 8 weeks of metacognitive training in schizophrenia produced increased regional brain homogeneity across prefrontal and limbic regions on fMRI. However, these studies typically lack active control conditions that would distinguish specific metacognitive training effects from general therapeutic engagement or attention effects.

Future research incorporating neuroimaging into RCTs of MCT and MERIT, with appropriate control conditions, would significantly advance understanding of whether and how these therapeutic approaches engage neuroplastic mechanisms.

Effective metacognitive educational interventions typically include:

Research by Rueda et al. (2018) demonstrated that children receiving metacognitive scaffolding alongside executive attention training showed not only larger skill gains but also significant increases in electrophysiological indices of conflict processing. Critically, these neural changes directly predicted cognitive improvements, suggesting that metacognitive scaffolding enhanced the effectiveness of cognitive training by engaging monitoring and regulation processes.

Many strategies identified as metacognitive in educational contexts are actually domain-general problem-solving strategies applied to the problem of self-development (Hulbig, 2023, 2024). This reframes metacognitive skill development as teaching students to conceptualize their learning as a problem requiring systematic problem-solving: defining learning goals, selecting and implementing strategies, monitoring progress, and adjusting approaches based on feedback.

Teachers and mentors can support students in developing these skills by making the problem-solving framework explicit, providing models of expert metacognitive processing, and creating opportunities for deliberate practice with feedback. This approach may be particularly valuable for students with learning differences who face more complex learning challenges requiring systematic self-regulation.

Converging evidence from neuroimaging, lesion studies, and brain stimulation research indicates that metacognitive processes rely on distributed networks involving prefrontal cortex (monitoring and control), anterior cingulate (conflict detection and uncertainty), parietal cortex (evidence accumulation and confidence), and hippocampus (particularly for memory-related metacognition). However, substantial heterogeneity exists depending on which specific metacognitive functions are examined, task domain, and individual differences.

Meta-analyses (Vaccaro and Fleming, 2018) provide the most reliable characterization of consistent activation patterns, but show variability reflecting the multifaceted nature of metacognition. Different theoretical frameworks (Fleming’s signal-detection approaches, Lysaker’s integrative model, Wells’ metacognitive therapy framework) emphasize different components of metacognitive function and may engage partially dissociable neural systems.

Multiple therapeutic approaches targeting metacognitive processes show efficacy across various conditions. MCT demonstrates superior outcomes compared to CBT for depression and anxiety, with benefits maintained at follow-up (Callesen et al., 2020; Nordahl et al., 2018). MERIT shows promise for schizophrenia, where metacognitive capacity predicts functional outcomes beyond neurocognition (Lysaker et al., 2018). Metacognitive interventions for TBI improve awareness and functional outcomes (Ownsworth et al., 2019).

The sustained nature of these clinical benefits raises questions about underlying mechanisms. Are improvements maintained through continued application of learned strategies, through neural reorganization making metacognitive processing more efficient, or through both? Most clinical trials have not included neuroimaging, leaving these questions unresolved.

Meta-analyses consistently show positive associations between metacognitive strategy instruction and academic achievement (Xie et al., 2024), with particularly strong effects for students with learning differences. The magnitude of effects (up to 7 months additional progress) suggests educationally meaningful benefits. Neural studies combining metacognitive scaffolding with attention training show both behavioral improvements and enhanced electrophysiological indices of cognitive processing (Rueda et al., 2018).

Evidence that intensive training produces structural brain changes (Maguire et al., 2000, 2006) and that regions critical for metacognition demonstrate experience-dependent plasticity (Newman and McGaughy, 2011) suggests potential mechanisms. Molecular mediators of plasticity including BDNF can be enhanced through exercise and meditation (Fernández-Rodríguez et al., 2022; Calderone et al., 2024), interventions that may complement metacognitive training.

However, direct evidence specifically linking metacognitive training to structural neural reorganization remains limited. Most studies showing neural changes following training involve intensive cognitive tasks that likely engage metacognitive processes but did not explicitly train metacognitive skills. Establishing causal relationships requires studies that: (1) use explicit metacognitive training protocols, (2) include appropriate control conditions, (3) measure both metacognitive outcomes and neural changes, and (4) employ longitudinal designs tracking the time course of change.

Contemporary neuroscience increasingly conceptualizes brain function as hierarchical predictive processing (Seth, 2021). Lower levels generate predictions about sensory input; higher levels monitor prediction errors and update models accordingly. Metacognition can be understood as explicit awareness of this prediction-error process—the ability to consciously detect mismatches between expectations and outcomes and systematically adjust strategies.

In this framework (illustrated in Figure 2), metacognitive knowledge generates predictions about cognitive performance, while metacognitive monitoring detects errors (mismatches between predicted and actual performance). Metacognitive regulation then adjusts strategies to reduce future errors. This maps onto fundamental neural learning mechanisms: prediction errors drive synaptic modifications that improve future predictions.

However, this theoretical alignment does not establish that metacognitive training specifically enhances neuroplasticity beyond domain-specific learning. A critical empirical question remains: Does explicit training in metacognitive monitoring and regulation produce neural changes beyond those resulting from simply practicing the target cognitive skills?

An alternative framework views metacognition as domain-general problem-solving processes applied to cognitive tasks (Hulbig, 2023, 2024). From this perspective, metacognitive training teaches systematic approaches to defining learning problems, generating and testing strategies, monitoring outcomes, and iterative refinement—procedures applicable across domains.

This framework suggests that metacognitive training might support transfer of learning across contexts by establishing generalizable self-regulatory procedures. Neural plasticity would result from both domain-specific skill development and the development of more efficient control processes. Whether such training produces dissociable neural changes in control networks versus task-specific regions requires investigation.

A third possibility is that metacognitive awareness facilitates memory consolidation and skill learning. Explicit awareness of what one has learned and how effectively one performed might enhance hippocampal-mediated memory consolidation and strengthen learning-related synaptic changes. This could explain why metacognitive scaffolding enhances training outcomes (Rueda et al., 2018) even when the primary training targets domain-specific skills.

These frameworks are not mutually exclusive and may capture complementary aspects of how metacognitive processes interface with neural learning mechanisms. Distinguishing among these possibilities requires carefully designed studies manipulating metacognitive training components while measuring both immediate behavioral effects and longer-term neural changes.