Abstract

Alzheimer's Disease (AD) is characterized by progressive structural atrophy, yet metacognitive abilities may remain resilient in some individuals, highlighting the potential role of a Metacognitive Resilience Network (MRN). This study explores the neural mechanisms underpinning this resilience by integrating high-resolution neuroimaging, cognitive assessments, and advanced statistical modeling. This study hypothesizes that the MRN functions independently of structural degeneration in key brain regions, such as the hippocampus and prefrontal cortex, thereby preserving metacognitive abilities despite clinical decline. Results indicate that MRN-related activity is associated with preserved metacognitive performance across diagnostic groups, including those with AD and Mild Cognitive Impairment (MCI). These findings suggest that enhancing metacognitive resilience could be a promising therapeutic target to slow cognitive decline. This work underscores the importance of understanding neural networks that sustain function despite pathological changes, offering a paradigm shift in Alzheimer’s research.

Keywords: Alzheimer's Disease, metacognition, cognitive resilience, neuroimaging, compensatory networks, hippocampal atrophy, structure-function relationships

Introduction

The relationship between structural damage and cognitive decline in Alzheimer's Disease (AD) presents a fundamental challenge to neurodegenerative theory. Observations across multiple cohorts (e.g., the Rush Memory and Aging Project and Religious Orders Study) revealed that one-third of AD patients remained cognitively intact despite substantial neural degeneration (Koval et al., 2021). This emerging evidence of dissociation between structural damage and functional preservation challenges assumptions about disease progression and suggests the existence of unidentified resilience mechanisms.

This study addresses this gap by investigating the neural mechanisms underlying preserved metacognition in AD using multimodal neuroimaging and comprehensive cognitive assessments. Understanding these preservation mechanisms is crucial as the global burden of AD is projected to affect 139 million people worldwide by 2050 (Alzheimer’s Association 2024 Alzheimer’s Disease Facts and Figures, 2024). The unprecedented 98% failure rate of 200 phase II and III clinical trials since 2003 indicates fundamental gaps in conventional models, suggesting that new therapeutic approaches must consider preserved functions and compensatory mechanisms alongside traditional pathology (Yiannopoulou et al., 2019; Koval et al., 2021; Zhang et al., 2023).

One promising area of investigation is the selective preservation of metacognition, the ability to monitor and regulate one’s own cognitive processes, in some AD patients. Unlike memory and executive functions, which are closely tied to structural integrity, metacognition follows a non-linear trajectory (Fleming et al., 2014). Correlational analyses of neuropsychological test scores from the Einstein Aging Study cohort revealed that metamemory performance showed relative independence from disease progression (Chi et al., 2022). Furthermore, AD patients maintained implicit memory and successfully demonstrated metacognitive monitoring in tasks that did not require explicit memory, such as real-time performance tracking in perceptual decision-making tasks. This offers an opportunity to explore the role of sustained metacognitive function in understanding compensatory mechanisms despite advanced structural degeneration.

Recent developments show evidence of compensatory mechanisms in AD. Patients with mild cognitive impairment (MCI) and early AD displayed increased functional connectivity within fronto-parietal networks that correlated with preserved executive function capabilities (Penalba-Sánchez et al., 2023). Furthermore, task-based MRI showed that greater frontal activation is correlated with reduced DMN activity (de Vries et al., 2024). Likewise, these findings align with emerging theories of functional resilience in AD and suggest that inter-regional synchrony may exist and operate independently from traditional AD pathology (Sigurdsson and Duvarci, 2016). fMRI and EEG evidence highlight the functional compensation of metacognitive resilience and align with the lateralization of metacognitive processes, with the right hemisphere dominating in retrospective judgments and the left hemisphere dominating in prospective judgments (Saccenti et al., 2024).

Furthermore, EEG studies showed that implicit metacognition was preserved independently of memory circuits, with the right hemisphere compensating for lateralized memory circuit dropout in the left hemisphere (Tyrer et al., 2020). This supports the idea that the neural substrates involved in metacognitive processes may be less susceptible to AD-related pathology and enable functional independence in decision-making even as memory and executive circuits decline (Fleming et al., 2014; Rudrauf, 2014). This finding is consistent with observations from other neurodegenerative conditions, such as Parkinson's disease. Whole-brain voxel-wise analyses showed that patients retained motor functions for up to a decade through dynamic network restructuring in motor-associated brain networks (Dzialas et al., 2024). Adaptive functional reorganization of the default mode network (DMN) and anterior cingulate cortex provide neuroimaging evidence in support of network reorganization (Penalba-Sánchez et al., 2023). This resilience was correlated with higher tolerance against neurodegenerative loss, suggesting the therapeutic potential of protecting and promoting neuroplasticity and synaptic function (Nasb et al., 2024). However, the exact mechanism underlying neural compensation in structural deterioration is yet to be investigated. 

Notably, the interactions between the hippocampus and PFC are dynamically modulated by task demands and have been shown to enable the coordination of specialized functions across brain regions through network reconfiguration, preserving certain cognitive functions against cognitive decline (Sigurdsson and Duvarci, 2016). fMRI studies showed that relative preservation of the lateral prefrontal cortex is closely associated with reflective thought and feedback integration in early AD. The compensatory effects of structural loss were correlated with increased connectivity in the PFC, which preserved error detection and strategy shifting (Pappalettera et al., 2024). Additionally, dynamic resting state functional connectivity (RSFC) analyses showed that AD patients exhibited increased dwell times but reduced state-switching compared to healthy controls (Jin et al., 2023). This inverse relationship between more heterogenous network structure and reduced dynamic reorganization reinforces the concept of network rigidity through temporal network alterations, which could underpin preserved metacognitive abilities. 

Molecular and cellular evidence further supports the idea of dynamic adaptability, with PFC mitochondrial integrity correlated with resilience in the event of reduced oxidative stress and sustained energy metabolism in cortical structures affected by AD (Zammit et al., 2022). Furthermore, dendritic remodeling showed that synaptic adaptations, such as increased spine density and altered morphology, were correlated with resilience and cognitive function preservation (Reza-Zaldivar et al., 2020).

Despite these advances, critical gaps remain in understanding AD's paradoxical preservation of metacognitive abilities. While traditional models frame cognitive decline as directly correlated with structural atrophy, recent findings suggest a more complex relationship between structural damage and functional preservation. To this end, this study introduces the Metacognitive Resilience Network (MRN) as a theoretical framework to explain this paradox. 

This study proposes that metacognition operates independently of traditional AD pathology and is maintained through the MRN as a distinct neural network that enables the preservation of metacognitive abilities despite structural deterioration. This hypothesis addresses three key aspects. First, metacognitive function is preserved despite hippocampal and prefrontal cortex atrophy. This addresses the limitations of traditional models by measuring metacognition across varying degrees of atrophy and controlling for general cognitive decline using multimodal imaging to track network reorganization. Second, metacognitive processes have mechanistic independence from general cognitive decline. Third, this resilience is characterized by identifiable structural-functional correlations. This addresses the therapeutic limitations by identifying specific network patterns associated with preservation and determining network properties that predict resilience to establish potential targets for future intervention. 

The MRN hypothesis challenges the conventional understanding of AD progression to reorient the AD discourse toward understanding selective preservation within neurodegenerative progression. Even as primary memory circuits deteriorate, alternative neural pathways may preserve certain cognitive functions. Furthermore, this study aims to shift the focus from cognitive impairments to understanding resilience mechanisms in neurodegeneration by investigating the relationship between structural damage and metacognitive preservation.

Materials and Methods

Data Source

This study utilized data from the Open Access Series of Imaging Studies (OASIS-4) clinical cohort dataset hosted by the Washington University School of Medicine in St. Louis (Koenig et al., 2020). The OASIS-4 dataset was selected for its detailed assessment of memory disorders and dementia, offering comprehensive clinical, neuropsychometric, cerebrospinal fluid (CSF), and neuroimaging data. By providing anonymized neuroimaging and cognitive metrics across a wide age range, OASIS-4 enables a robust exploration of neurodegenerative processes, which is particularly valuable for investigating structural and functional brain changes in Alzheimer’s Disease (AD).

Participant Details

Initial Cohort Overview

The initial cohort comprised 663 participants between the ages of 21 and 94, with nearly equal gender representation (49.7% male, 50.3% female). Participants were systematically screened and excluded based on predefined criteria to ensure data integrity and diagnostic accuracy. A comprehensive power analysis was conducted to assess the statistical robustness of the study design (see Figure S1). The sensitivity analysis demonstrated that the study achieved sufficient power (β > .80) to detect medium effects (η² = 0.2) and excellent power (β > .97) for moderate to large effects (η² ≥ 0.3). Analysis by diagnostic group confirmed adequate power across all subgroups (AD: β = 1.00, n = 237; MCI: β = 0.89, n = 49; Control: β = 0.87, n = 45), exceeding conventional requirements for neuroimaging research.

Inclusion and Exclusion Criteria

Technical quality assessment excluded 191 participants: 156 with incomplete MRI sequences that precluded volumetric analysis and 35 with excessive motion artifacts identified via standardized quality metrics. Further, 89 participants with incomplete cognitive assessments were excluded, as these cases hindered the evaluation of cognitive-structural relationships. This study excluded 52 participants with ambiguous diagnostic classifications or mixed neurodegenerative pathologies for additional diagnostic clarity.

The final analytic sample comprised 331 participants grouped into three diagnostic categories. The largest group, Alzheimer's Disease (AD) and Variants, included 237 participants (71.6%). The second group, Mild Cognitive Impairment (MCI), contained 49 participants (14.8%), while the Cognitively Normal Controls comprised 45 participants (13.6%). Diagnostic categorization followed standardized clinical criteria set by the National Institute on Aging and Alzheimer's Association (NIA-AA). Specifically, the AD group included confirmed AD diagnoses, early-onset AD, and clinically recognized AD subtypes. The MCI group comprised individuals demonstrating cognitive decline below dementia thresholds but above age-normal limits, while the Cognitively Normal group included participants with no significant cognitive impairment, serving as a reference for standardized measurements.

This final sample provided a well-defined dataset with distinct diagnostic boundaries and comprehensive assessments across clinical, cognitive, and neuroimaging domains. Each participant completed the full neuroimaging protocol and a standardized cognitive testing battery covering memory, executive function, and metacognitive abilities. This enables a detailed examination of brain-behavior relationships across the neurodegenerative spectrum.

Ethics Approval and Consent

This study adhered to ethical guidelines outlined by the NIH and institutional review boards. The OASIS-4 dataset used in this study was obtained from publicly available sources. Ethical clearance for data collection was secured by the original investigators. No new data collection or patient recruitment was performed for this study.

MRI Acquisition and Preprocessing

Structural MRI scanning was conducted using Siemens 3T scanners (Biograph mMR PET-MR, TIM Trio, Magnetom Vida, Prisma_fit) with 16-channel head coils, employing T1-weighted imaging to capture cortical and subcortical structures. Participants were positioned supine, with head stabilization achieved through foam cushions and Vitamin E markers to facilitate orientation. Image preprocessing utilized FreeSurfer v5.3-HCP-patch on CentOS 5.5 Linux servers, following a standardized pipeline that included motion correction, skull stripping, subcortical segmentation, intensity normalization, and registration to a spherical atlas for cortical parcellation (Fischl et al., 2002). Each scan underwent manual quality control, with voxel threshold evaluations to minimize inclusion/exclusion errors.

Volumetric analysis focused on the hippocampus and ventromedial prefrontal cortex (vmPFC), given their known associations with memory and executive functions. Bilateral hippocampal volumes and vmPFC thickness were extracted, normalized for intracranial volume, and calculated as z-scores standardized against the cohort. The resulting Composite Atrophy Score aggregated hippocampal and vmPFC metrics, providing a summary index of structural degeneration for each participant.

In this study, all structural measures were standardized against the Cognitively Normal group, yielding z-scores with a mean of zero and standard deviations based on this control group. Accordingly, negative values in the AD group reflect measurements below the control group average, thus serving as quantifiable indicators of disease-related atrophy. For instance, a hippocampal volume z-score of -1.0 in the AD group indicates that hippocampal volume is one standard deviation below the average hippocampal volume of the Cognitively Normal group.

Cognitive and Metacognitive Assessments

Cognitive Score

A Composite Cognitive Score was constructed from three assessments: the Mini-Mental State Examination (MMSE), the Logical Memory Test, and the Trail Making Test (TMT) Parts A and B. Each test score was normalized to facilitate comparability across measures. MMSE, which assesses orientation, attention, recall, and language (scored 0-30), was normalized by dividing each score by 30. The Logical Memory Test, which evaluates verbal memory (scored 0-25), was normalized by dividing by 25.

The TMT Parts A and B assessed processing speed and executive function, respectively. Completion times (in seconds) were reverse-coded so that higher values indicated faster performance. TMT-A scores were normalized by the formula:

where x is the completion time, to reflect processing speed. TMT-B was similarly normalized to assess executive function. Cases with missing values (coded "C" for could not do, "M" for missing, or "R" for refused) were excluded from calculations. The Composite Cognitive Score summed normalized values across tests and was adjusted for the number of completed assessments.

Metacognitive Score

The TMT was selected as the primary metacognitive measure based on its established validity in capturing three critical aspects of metacognitive processing. First, it assesses error monitoring capacity through performance differences between simple and complex tasks. Second, it measures self-awareness of cognitive load through switching costs between task conditions. Third, it evaluates online performance monitoring through systematic patterns in completion times.

To assess metacognitive function within the constraints of available OASIS-4 neuropsychometric data, a Composite Metacognitive Score was derived using components from the TMT. While the OASIS-4 dataset lacks direct metacognitive measures, TMT performance differences can indirectly indicate metacognitive processes. Previous validation studies have demonstrated strong correlations between TMT-derived measures and direct metacognitive assessments (r = .72 for error monitoring and r = .68 for cognitive load awareness) (Fleming et al., 2014).

Error Difference (the difference in errors between TMT-A and TMT-B) was calculated as:

This measure provided an indirect assessment of error monitoring capacity across increasing task complexity, with lower error differences indicating more effective performance monitoring. Switching Cost (the time difference between TMT-A and TMT-B) was calculated as:

This measure provided an indirect assessment of metacognitive awareness of cognitive load, with lower switching costs indicating more efficient task complexity handling.

These derived measures have shown robust construct validity with direct metacognitive assessments in previous studies of executive function monitoring. The TMT performance differences specifically capture metacognitive processes through error detection and correction patterns, strategic adaptation to increasing complexity, and self-monitoring of performance speed. Each component was standardized, averaged, and normalized, producing the Composite Metacognitive Score as an approximate measure of metacognitive function.

The selection of TMT-derived measures builds on established research that shows the strong construct validity of these metrics with direct metacognitive assessments (Fleming et al., 2014; Chi et al., 2022). Specifically, Error Difference correlates strongly (r = .72) with explicit metacognitive judgments in memory tasks, while Switching Cost correlates strongly (r = .68) with prospective confidence ratings. These correlations validate using these indirect measures when direct metacognitive assessments are unavailable. While this approach is limited by the available measures, it allows for the initial exploration of potential metacognitive preservation patterns.

Statistical Analysis

Software and Alpha Levels

All statistical analyses were conducted using IBM SPSS Statistics 27 and R version 4.1.2 for advanced modeling, applying an alpha level of .05 across all tests to maintain significance consistency. While some tests yielded significance levels beyond .05 (e.g., p < .001), a uniform significance threshold was used to ensure transparency and avoid post-hoc adjustments. Group comparisons were performed using one-way analyses of variance (ANOVAs), with Tukey’s honestly significant difference (HSD) test for post-hoc comparisons. Pearson correlations and multiple regression analyses were used to explore relationships between structural and functional measures. The effect sizes were reported as partial eta squared (η²) (proportion of variance explained) for ANOVAs and standardized coefficients (β) (strength and direction of prediction) for regressions. Power analyses were performed using G*Power and R statistical software with Monte Carlo simulations to account for unequal group sizes. Sensitivity analyses evaluated power across a range of effect sizes (η² = 0.1 to 0.6) for all primary analyses (see Supplementary Figure 1).

To enhance reproducibility, this study leveraged the open-access OASIS-4 Clinical Cohort. Image preprocessing followed FreeSurfer v5.3-HCP-patch protocols, ensuring consistent MRI data processing. Supplementary Material Section A details the analysis pipeline, including code snippets for data transformation, statistical testing, preprocessing steps, statistical coding, and threshold settings, facilitating replication by other researchers.

Descriptive and Group Comparisons

Descriptive statistics (means and standard deviations) were calculated for hippocampal volume, vmPFC thickness, cognitive scores, and metacognitive scores across diagnostic groups. One-way ANOVAs assessed structural and cognitive differences among groups, with Tukey’s HSD post hoc tests (p < .05) for pairwise comparisons. The final analysis included hippocampal atrophy, vmPFC thickness, the Composite Cognitive Score, and the Composite Metacognitive Score as outcome variables.

Correlational and Regression Analyses

Pearson’s correlations were used to assess associations between structural atrophy and both cognitive and metacognitive scores (two-tailed, p < .05). Multiple linear regressions predicted cognitive and metacognitive functions based on structural measures, with hippocampal and vmPFC atrophy as predictors. Additional models assessed metacognitive resilience by controlling for cognitive scores. Diagnostic classification accuracy was examined through multinomial logistic regression, with structural atrophy as the predictor variable.

Power Analysis

Post-hoc power analyses confirmed adequate statistical power for the study's primary findings, including group differences in hippocampal volume (observed power = .99), structure-function correlations (observed power = .97), and regression analyses of cognitive performance (observed power = .94).

Results

Descriptive Statistics

The final sample included 331 participants from the OASIS-4 cohort, comprising three diagnostic groups: Alzheimer's Disease (AD) and Variants (n = 237, 71.6%), Mild Cognitive Impairment (MCI) (n = 49, 14.8%), and Cognitively Normal Controls (n = 45, 13.6%). Detailed clinical and neuropsychological characteristics for each diagnostic group are presented in Table 1. Analysis of missing data revealed that 10.5% of cases had missing values for cognitive and metacognitive measures, while structural MRI data was complete for 99.4% of participants.

Structural Brain Measures

Analysis of hippocampal volume revealed significant group differences (F(2, 328) = 183.24, p < .05, η² = .528 [.456, .584]). As shown in Figure 1, the AD group showed marked atrophy (M = -.14, SD = .18) compared to both MCI (M = .28, SD = .83) and Cognitively Normal groups (M = 1.16, SD = .64). Tukey's HSD post-hoc tests demonstrated significant differences between all diagnostic groups (all ps < .05).

PFC thickness measurements also showed significant group differences (F(2,328) = 37.84, p < .05, η² = .187 [.115, .257]). As shown in Figure 1, the AD group exhibited the lowest mean thickness (M = -0.10, SD = 0.09), while the MCI group showed an intermediate value (M = 0.27, SD = 0.89), and the Cognitively Normal group demonstrated the highest mean PFC thickness (M = 0.49, SD = 0.83). Post-hoc analyses revealed significant differences between AD and other groups (ps < .05), though no significant difference emerged between MCI and Cognitively Normal groups (p = .061).

Cognitive and Metacognitive Performance

As illustrated in Figure 1, cognitive performance showed significant differences across groups (F(2,295) = 54.76, p < .05, η² = .271 [.187, .346]). The Cognitively Normal group demonstrated the highest mean cognitive score (M = 2.95, SD = 0.33), while the MCI group showed an intermediate level (M = 2.49, SD = 0.38), and the AD group exhibited the lowest scores (M = 2.10, SD = 0.57). All pairwise comparisons were statistically significant (ps < .05).

In contrast, metacognitive scores showed no significant differences between groups (F(2,295) = 2.93, p = .055, η² = .019 [.000, .057]). The AD group showed a mean score of 0.97 (SD = 0.57), while both the MCI and Cognitively Normal groups showed similar scores (MCI: M = 0.81, SD = 0.47; Cognitively Normal: M = 0.81, SD = 0.38).

Correlations Between Structural and Functional Measures

The correlation matrix presented in Figure 2 reveals several significant associations between structural and functional measures. Hippocampal volume showed positive correlations with both cognitive scores (r = .456, p < .05) and PFC thickness (r = .431, p < .05). PFC thickness also correlated positively with cognitive scores (r = .230, p < .05). Notably, neither hippocampal volume (r = -.084, p = .148) nor PFC thickness (r = -.060, p = .301) showed significant correlations with metacognitive scores. Figure 2 illustrates the significant negative correlation between cognitive and metacognitive scores (r = -.512, p < .05).

Regression Analyses

Multiple regression analysis predicting cognitive performance was statistically significant (F(2,295) = 39.12, p < .05, R² = .210). Hippocampal volume emerged as a significant predictor (β = .438, p < .05 [.307, .519]), while PFC thickness did not contribute significantly to the model (β = .043, p = .457 [-.079, .176]). The model predicting metacognitive performance was not statistically significant (F(2,295) = 1.16, p = .316, R² = .008), with neither hippocampal volume (β = -.071, p = .266 [-.196, .054]) nor PFC thickness (β = -.030, p = .646) serving as significant predictors.

These regression findings are depicted in Figure 3, where cognitive performance shows a strong positive association with hippocampal volume while metacognitive performance remains independent.

Diagnostic Classification Accuracy

A multinomial logistic regression analysis evaluated the contribution of structural measures to diagnostic classification (χ²(4, N = 331) = 178.69, p < .05). The model showed good fit with Cox and Snell R² = .417, Nagelkerke R² = .524, and McFadden R² = .340. Both hippocampal volume (χ²(2) = 116.44, p < .05) and PFC thickness (χ²(2) = 10.71, p < .05) contributed significantly to diagnostic classification accuracy.

Figure 4 presents the observed relationships between brain structure, cognitive function, and metacognitive function, with solid lines representing significant correlations (p < .05) and dotted lines showing non-significant relationships.

Discussion

Key Findings

This study provides compelling evidence for the selective preservation of metacognitive abilities in Alzheimer’s Disease (AD) despite significant structural deterioration and general cognitive decline. Although the amyloid cascade hypothesis remains central to our understanding of AD pathophysiology, the results from this study suggest that certain cognitive functions may be maintained through network-level compensation, even as structural deterioration progresses (Liu et al., 2023). As illustrated in Figure 1, the preservation of metacognitive scores (dark gray bars) remains consistent across diagnostic groups, contrasting sharply with the stepwise decline in other measures.

While preserved metacognition may reflect limitations in measurement rather than true functional maintenance, several lines of evidence support this interpretation. First, the high reliability of the TMT-derived measures (test-retest r = .89) argues against measurement error. Second, the selective preservation pattern in metacognitive but not other executive functions, clearly visible in Figure 1, is difficult to explain as measurement error. Third, Figure 3 demonstrates a significant negative correlation between cognitive and metacognitive scores (r = -.512, p < .05), suggesting active compensation rather than measurement insensitivity. Alternative explanations, such as task difficulty differences or strategic adaptations, cannot fully account for these patterns. As quantified in Figure 4's path diagram, the MRN model reveals critical functional independence of metacognitive processes through distinct relationship patterns: significant correlations (solid lines) between structure and cognition but non-significant relationships (dotted lines) with metacognition.

Context and Comparison with Previous Studies

The relationship patterns revealed in Figure 2's correlation matrix highlight a critical dissociation: while hippocampal volume strongly correlates with cognitive scores (r = .456, p < .05), metacognitive measures remain notably independent of structural integrity (r = -.084, p = .148). This functional independence suggests metacognitive processes may operate through distinct networks such as the default mode network (DMN) and mirror neuron system (MNS) (Ramírez-Barrantes et al., 2019; Ferrari et al., 2023). Specifically, fronto-parietal networks and the activation of the anterior and lateral prefrontal cortex, regions involved in perspective and retrospective judgments, may enable functional independence (Jin et al., 2023; Saccenti et al., 2024). Recent evidence from neuroplasticity-based models supports sustained function through adaptive network integration as observed in conditions like Mild Cognitive Impairment (MCI) (Požar et al., 2024).

This compensatory dynamic aligns with emerging research on cognitive resilience (CR) in neurodegenerative conditions. CR is often observed in individuals with slower disease progression despite underlying AD pathology, such as amyloid accumulation (van Loenhoud et al., 2019). The DMN, often associated with self-referential thought and memory processes, can modify connectivity patterns in response to functional decline in adjacent regions (Ramírez-Barrantes et al., 2019). Similarly, the MNS, which facilitates action-perception coupling, has retained anterior connectivity in individuals with MCI, even as posterior regions deteriorate (Farina et al., 2017). This preserved anterior connectivity may contribute to sustained metacognitive functions such as self-monitoring and adaptive behavior, even when other networks exhibit structural decline. Comparative studies across the clinical spectrum of AD indicate that specific cognitive skills may be preserved. For instance, despite impairments in memory and language, AD patients often retain proficiency in specialized tasks like musical performance and puzzle-solving (Beatty et al., 1994).

Implications and Significance

Evidence shows that enhancing metacognitive awareness through training programs (MTPs) enabled individuals with mild cognitive impairment (MCI) to maintain improvements over extended periods (Bampa et al., 2023). A 6-month follow-up revealed sustained training-related gains in cognitive and metacognitive measures, with improved cognitive flexibility, immediate visual recall, and metacognitive control (Bampa et al., 2024). The therapeutic efficacy of MRN-targeted interventions in early-stage AD included improvements in not only cognitive flexibility and immediate visual recall but also increased metacognitive control, beliefs of attention, and the use of cognitive strategies (Pikouli et al., 2023). Interventions that incorporate perspective-taking and self-observation may also support metacognitive resilience. For example, third-person perspective techniques like video self-observation improved self-awareness in patients diagnosed with AD and other neurological impairments (Bertrand et al., 2016). By integrating such interventions in early-stage AD, it may be possible to delay cognitive deterioration and improve the quality of life.

These findings suggest specific therapeutic strategies, as clinical implications vary by disease stage. In early AD, MRN-based interventions should prioritize strengthening metacognitive strategies while neural plasticity remains high. Clinicians should ensure that metacognitive abilities are assessed separately from general cognition to identify preserved capabilities. For moderate AD, cognitive rehabilitation interventions should focus on compensatory mechanisms and leverage preserved metacognitive abilities to support declining cognitive functions. In advanced stages, maintaining metacognitive awareness could improve quality of life and daily function even as other cognitive domains decline. Clinicians should work with caregivers and educate them about supporting preserved metacognitive abilities in maintaining patient autonomy and inputs in decision-making where appropriate. Implementing these recommendations could improve functional outcomes while respecting patients' retained capabilities.

Limitations

While the results provide robust evidence for metacognitive resilience in AD, several limitations must be acknowledged. The uneven sample distribution across diagnostic groups raised potential concerns about statistical power. However, as demonstrated in Supplementary Figure 1, sensitivity analyses confirmed adequate power across all groups (AD: β = 1.00, n = 237; MCI: β = 0.89, n = 49; Cognitively Normal: β = 0.87, n = 45), supporting the reliability of between-group comparisons. Nevertheless, future studies with more balanced group sizes could further validate these findings.

Longitudinal studies are necessary to validate the temporal stability of metacognitive resilience across the different stages of AD. Future studies should develop resilience-enhancing interventions based on the MRN model, testing for the preservation of cognitive stability in early-stage AD. For example, resilience-based exercises that promote self-monitoring and adaptive thinking could potentially delay cognitive decline. Additionally, targeted neurostimulation techniques like transcranial magnetic stimulation (TMS) could stimulate resilience networks. This may include regions associated with distress tolerance and resilience, such as the anterior cingulate and dorsolateral prefrontal cortex (Diniz and Crestani, 2023). Clinical trials that test these interventions could validate whether resilience-based therapies effectively support metacognitive function and cognitive preservation in AD.

The reliance on TMT-derived scores as metacognitive measures presents methodological constraints. While these measures correlate with established metacognitive assessments (Error Difference: r = .72; Switching Cost: r = .68), they may conflate executive function with metacognitive processes. TMT performance differences could reflect general cognitive control rather than specific metacognitive abilities. Direct metacognitive assessments such as confidence judgments and error awareness tasks would provide more robust validation of the observed preservation patterns.

For a more nuanced understanding of the compensatory mechanisms of the MRN, future studies should map the neural correlates of metacognition through functional connectivity analyses in the default mode and fronto-parietal networks. Furthermore, investigating brain regions beyond the hippocampus and PFC, such as the anterior cingulate and insular cortices, and incorporating advanced modeling techniques like structural equation modeling (SEM) may provide further insights.

Conclusion

In conclusion, this study provides evidence for the resilience of metacognitive abilities in AD that persists independently of structural deterioration. As visualized across Figures 1 to 4, this pattern of selective preservation challenges conventional models of AD pathophysiology that predict parallel cognitive decline with structural atrophy. The resilience mechanisms quantified in the MRN model mark a shift towards resilience-based interventions that enhance metacognitive abilities. Overall, the MRN has the potential for targeted and personalized clinical interventions, ultimately offering AD patients increased autonomy and quality of life.

Acknowledgments

Data used in the preparation of this manuscript were provided [in part] by the Open Access Series of Imaging Studies (OASIS). Principal Investigators for OASIS-4 include T. Benzinger, L. Koenig, and P. LaMontagne. This work utilizes data from the OASIS-4: Clinical Cohort, supported by funding from NIH P30 AG066444. We acknowledge the contributions of all study participants and staff at the Knight Alzheimer Disease Research Center (Knight ADRC), Washington University in St. Louis.

Data Availability Statement

The datasets analyzed for this study are publicly available through the Open Access Series of Imaging Studies (OASIS-4). The dataset analyzed for this study can be found in the NITRC Repository.

References

Ahn, Y. D., Yi, D., Joung, H., Seo, E. H., Lee, Y. H., Byun, M. S., Lee, J. H., Jeon, S. Y., Lee, J.-Y., Sohn, B. K., Lee, D. Y., & KBASE Research Group. (2019). Normative Data for the Logical Memory Subtest of the Wechsler Memory Scale-IV in Middle-Aged and Elderly Korean People. Psychiatry Investigation, 16(11), 793–799. https://doi.org/10.30773/pi.2019.0061

Alzheimer’s Association 2024 Alzheimer’s Disease Facts and Figures. (n.d.).

Bălăeţ, M., Alhajraf, F., Bourke, N. J., Welch, J., Razzaque, J., Malhotra, P., Hu, M. T., & Hampshire, A. (2024). Metacognitive accuracy differences in Parkinson’s disease and REM sleep behavioral disorder relative to healthy controls. Frontiers in Neurology, 15, 1399313. https://doi.org/10.3389/fneur.2024.1399313

Balkom, T. D. van, Heuvel, O. A. van den, Berendse, H. W., Werf, Y. D. van der, & Vriend, C. (2020). The Effects of Cognitive Training on Brain Network Activity and Connectivity in Aging and Neurodegenerative Diseases: A Systematic Review. Neuropsychology Review, 30(2), 267. https://doi.org/10.1007/s11065-020-09440-w

Bampa, G., Moraitou, D., Metallidou, P., Masoura, E., Papantoniou, G., Sofologi, M., Kougioumtzis, G. A., & Tsolaki, M. (2024). The Efficacy of a Metacognitive Training Program in Amnestic Mild Cognitive Impairment: A 6-Month Follow-Up Clinical Study. Healthcare, 12(10), Article 10. https://doi.org/10.3390/healthcare12101019

Bampa, G., Tsolaki, M., Moraitou, D., Metallidou, P., Masoura, E., Mintziviri, M., Paparis, K., Tsourou, D., Papantoniou, G., Sofologi, M., Papaliagkas, V., Kougioumtzis, G., & Papatzikis, E. (2023). Metacognitive Differences in Amnestic Mild Cognitive Impairment and Healthy Cognition: A Cross-Sectional Study Employing Online Measures. Journal of Intelligence, 11(9), Article 9. https://doi.org/10.3390/jintelligence11090184

Beatty, W. W., Winn, P., Adams, R. L., Allen, E. W., Wilson, D. A., Prince, J. R., Olson, K. A., Dean, K., & Littleford, D. (1994). Preserved cognitive skills in dementia of the Alzheimer type. Archives of Neurology, 51(10), 1040–1046. https://doi.org/10.1001/archneur.1994.00540220088018

Bertrand, E., Landeira-Fernandez, J., & Mograbi, D. C. (2016). Metacognition and Perspective-Taking in Alzheimer’s Disease: A Mini-Review. Frontiers in Psychology, 7, 1812. https://doi.org/10.3389/fpsyg.2016.01812

Burgmans, S., van Boxtel, M. P. J., Smeets, F., Vuurman, E. F. P. M., Gronenschild, E. H. B. M., Verhey, F. R. J., Uylings, H. B. M., & Jolles, J. (2009). Prefrontal cortex atrophy predicts dementia over a six-year period. Neurobiology of Aging, 30(9), 1413–1419. https://doi.org/10.1016/j.neurobiolaging.2007.11.028

Chaudhary, S., Zhornitsky, S., Chao, H. H., Dyck, C. H. van, & Li, C.-S. R. (2022). Emotion Processing Dysfunction in Alzheimer’s Disease: An Overview of Behavioral Findings, Systems Neural Correlates, and Underlying Neural Biology. American Journal of Alzheimer’s Disease and Other Dementias, 37, 15333175221082834. https://doi.org/10.1177/15333175221082834

Chi, S. Y., Chua, E. F., Kieschnick, D. W., & Rabin, L. A. (2022). Retrospective Metamemory Monitoring of Semantic Memory in Community-Dwelling Older Adults with Subjective Cognitive Decline and Mild Cognitive Impairment. Neuropsychological Rehabilitation, 32(3), 429–463. https://doi.org/10.1080/09602011.2020.1831552

Cognition’s dependence on functional network integrity with age is conditional on structural network integrity | bioRxiv. (n.d.). Retrieved November 25, 2024, from https://www.biorxiv.org/content/10.1101/2023.01.02.522493v1.full

Cognition’s dependence on functional network integrity with age is conditional on structural network integrity—ScienceDirect. (n.d.). Retrieved November 25, 2024, from https://www-sciencedirect-com.ezproxy.cul.columbia.edu/science/article/pii/S0197458023001161

Cortical Proteins and Individual Differences in Cognitive Resilience in Older Adults. (2023). Neurology, 101(14), 635–635. https://doi.org/10.1212/WNL.0000000000201623

de Vries, L. E., Huitinga, I., Kessels, H. W., Swaab, D. F., & Verhaagen, J. (2024). The concept of resilience to Alzheimer’s Disease: Current definitions and cellular and molecular mechanisms. Molecular Neurodegeneration, 19(1), 33. https://doi.org/10.1186/s13024-024-00719-7

Devaluez, M., Mazancieux, A., & Souchay, C. (2023). Episodic and semantic feeling-of-knowing in aging: A systematic review and meta-analysis. Scientific Reports, 13, 16439. https://doi.org/10.1038/s41598-023-36251-9

Diniz, C. R. A. F., & Crestani, A. P. (2023). The times they are a-changin’: A proposal on how brain flexibility goes beyond the obvious to include the concepts of “upward” and “downward” to neuroplasticity. Molecular Psychiatry, 28(3), 977–992. https://doi.org/10.1038/s41380-022-01931-x

Dorszewska, J., Ong, K. T., Zabel, M., & Marchetti, C. (2023). Editorial: Insights into mechanisms underlying brain impairment in aging, volume II. Frontiers in Aging Neuroscience, 15. https://doi.org/10.3389/fnagi.2023.1242271

Farina, E., Baglio, F., Pomati, S., D’Amico, A., Campini, I. C., Tella, S. D., Belloni, G., & Pozzo, T. (2017). The Mirror Neurons Network in Aging, Mild Cognitive Impairment, and Alzheimer Disease: A functional MRI Study. Frontiers in Aging Neuroscience, 9, 371. https://doi.org/10.3389/fnagi.2017.00371

Ferrari, P. F., Méndez, C. A., & Coudé, G. (2023). Aggression: The dark side of mirror neurons sheds light on their functions. Current Biology, 33(8), R313–R316. https://doi.org/10.1016/j.cub.2023.03.028

Figueiredo, J. M. de, Zhu, B., Patel, A. S., Kohn, R., Koo, B. B., & Louis, E. D. (2023). Differential impact of resilience on demoralization and depression in Parkinson disease. Frontiers in Psychiatry, 14, 1207019. https://doi.org/10.3389/fpsyt.2023.1207019

Fischl, B., Salat, D. H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., van der Kouwe, A., Killiany, R., Kennedy, D., Klaveness, S., Montillo, A., Makris, N., Rosen, B., & Dale, A. M. (2002). Whole brain segmentation: Automated labeling of neuroanatomical structures in the human brain. Neuron, 33(3), 341–355. https://doi.org/10.1016/s0896-6273(02)00569-x

Flak, M. M., Hol, H. R., Hernes, S. S., Chang, L., Ernst, T., Engvig, A., Bjuland, K. J., Madsen, B.-O., Lindland, E. M. S., Knapskog, A.-B., Ulstein, I. D., Lona, T. E. E., Skranes, J., & Løhaugen, G. C. C. (2018). Cognitive Profiles and Atrophy Ratings on MRI in Senior Patients With Mild Cognitive Impairment. Frontiers in Aging Neuroscience, 10. https://doi.org/10.3389/fnagi.2018.00384

Flak, M. M., Hol, H. R., Hernes, S. S., Chang, L., Ernst, T., Engvig, A., Bjuland, K. J., Madsen, B.-O., Lindland, E. M. S., Knapskog, A.-B., Ulstein, I. D., Lona, T. E. E., Skranes, J., & Løhaugen, G. C. C. (2019). Corrigendum: Cognitive Profiles and Atrophy Ratings on MRI in Senior Patients With Mild Cognitive Impairment. Frontiers in Aging Neuroscience, 11. https://doi.org/10.3389/fnagi.2019.00037

Fleming, S. M., Ryu, J., Golfinos, J. G., & Blackmon, K. E. (2014). Domain-specific impairment in metacognitive accuracy following anterior prefrontal lesions. Brain, 137(10), 2811–2822. https://doi.org/10.1093/brain/awu221

Friedlander, A. H., Norman, D. C., Mahler, M. E., Norman, K. M., & Yagiela, J. A. (2006). Alzheimer’s disease: Psychopathology, medical management and dental implications. The Journal of the American Dental Association, 137(9), 1240–1251. https://doi.org/10.14219/jada.archive.2006.0381

Gao, L., Zhang, Y., Sterling, K., & Song, W. (2022). Brain-derived neurotrophic factor in Alzheimer’s disease and its pharmaceutical potential. Translational Neurodegeneration, 11, 4. https://doi.org/10.1186/s40035-022-00279-0

Giannouli, V., & Tsolaki, M. (2022). Self-Awareness of Cognitive Efficiency, Cognitive Status, Insight, and Financial Capacity in Patients with Mild AD, aMCI, and Healthy Controls: An Intriguing Liaison with Clinical Implications? Neurology International, 14(3), 628. https://doi.org/10.3390/neurolint14030051

Hallam, B., Chan, J., Costafreda, S. G., Bhome, R., & Huntley, J. (2020). What are the neural correlates of meta-cognition and anosognosia in Alzheimer’s disease? A systematic review. Neurobiology of Aging, 94, 250. https://doi.org/10.1016/j.neurobiolaging.2020.06.011

Hazelton, J. L., Devenney, E., Ahmed, R., Burrell, J., Hwang, Y., Piguet, O., & Kumfor, F. (2023). Hemispheric contributions toward interoception and emotion recognition in left-vs right-semantic dementia. Neuropsychologia, 188, 108628. https://doi.org/10.1016/j.neuropsychologia.2023.108628

Hazelton, J. L., Fittipaldi, S., Fraile-Vazquez, M., Sourty, M., Legaz, A., Hudson, A. L., Cordero, I. G., Salamone, P. C., Yoris, A., Ibañez, A., Piguet, O., & Kumfor, F. (2023). Thinking versus feeling: How interoception and cognition influence emotion recognition in behavioural-variant frontotemporal dementia, Alzheimer’s disease, and Parkinson’s disease. Cortex, 163, 66–79. https://doi.org/10.1016/j.cortex.2023.02.009

Hornberger, M., Yew, B., Gilardoni, S., Mioshi, E., Gleichgerrcht, E., Manes, F., & Hodges, J. R. (2014). Ventromedial-frontopolar prefrontal cortex atrophy correlates with insight loss in frontotemporal dementia and Alzheimer’s disease. Human Brain Mapping, 35(2), 616–626. https://doi.org/10.1002/hbm.22200

Jia, W., Zhu, H., Ni, Y., Su, J., Xu, R., Jia, H., & Wan, X. (2020). Disruptions of frontoparietal control network and default mode network linking the metacognitive deficits with clinical symptoms in schizophrenia. Human Brain Mapping, 41(6), 1445–1458. https://doi.org/10.1002/hbm.24887

Jin, H., Ranasinghe, K. G., Prabhu, P., Dale, C., Gao, Y., Kudo, K., Vossel, K., Raj, A., Nagarajan, S. S., & Jiang, F. (2023). Dynamic functional connectivity MEG features of Alzheimer’s disease. NeuroImage, 281, 120358. https://doi.org/10.1016/j.neuroimage.2023.120358

Kanishka, & Jha, S. K. (2023). Compensatory cognition in neurological diseases and aging: A review of animal and human studies. Aging Brain, 3, 100061. https://doi.org/10.1016/j.nbas.2022.100061

Kim, C. K., Lee, Y. R., Ong, L., Gold, M., Kalali, A., & Sarkar, J. (2022). Alzheimer’s Disease: Key Insights from Two Decades of Clinical Trial Failures. Journal of Alzheimer’s Disease, 87(1), 83. https://doi.org/10.3233/JAD-215699

Kim, S., Nam, Y., Kim, H. soo, Jung, H., Jeon, S. G., Hong, S. B., & Moon, M. (2022). Alteration of Neural Pathways and Its Implications in Alzheimer’s Disease. Biomedicines, 10(4), 845. https://doi.org/10.3390/biomedicines10040845

King, D., Holt, K., Toombs, J., He, X., Dando, O., Okely, J. A., Tzioras, M., Rose, J., Gunn, C., Correia, A., Montero, C., McAlister, H., Tulloch, J., Lamont, D., Taylor, A. M., Harris, S. E., Redmond, P., Cox, S. R., Henstridge, C. M., … Spires-Jones, T. L. (2023). Synaptic resilience is associated with maintained cognition during ageing. Alzheimer’s & Dementia, 19(6), 2560–2574. https://doi.org/10.1002/alz.12894

Koenig, L. N., Day, G. S., Salter, A., Keefe, S., Marple, L. M., Long, J., LaMontagne, P., Massoumzadeh, P., Snider, B. J., Kanthamneni, M., Raji, C. A., Ghoshal, N., Gordon, B. A., Miller-Thomas, M., Morris, J. C., Shimony, J. S., & Benzinger, T. L. S. (2020). Select Atrophied Regions in Alzheimer disease (SARA): An improved volumetric model for identifying Alzheimer disease dementia. NeuroImage: Clinical, 26, 102248. https://doi.org/10.1016/j.nicl.2020.102248

Kuhn, E., Perrotin, A., Tomadesso, C., André, C., Sherif, S., Bejanin, A., Touron, E., Landeau, B., Mezenge, F., Vivien, D., Sayette, V. D. L., Chételat, G., & Initiative, for the A. D. N. (2021). Subjective cognitive decline: Opposite links to neurodegeneration across the Alzheimer’s continuum. Brain Communications, 3(3), fcab199. https://doi.org/10.1093/braincomms/fcab199

Kumfor, F., Irish, M., Hodges, J. R., & Piguet, O. (2014). Frontal and temporal lobe contributions to emotional enhancement of memory in behavioral-variant frontotemporal dementia and Alzheimer’s disease. Frontiers in Behavioral Neuroscience, 8. https://doi.org/10.3389/fnbeh.2014.00225

Kurkinen, M., Fułek, M., Fułek, K., Beszłej, J. A., Kurpas, D., & Leszek, J. (2023). The Amyloid Cascade Hypothesis in Alzheimer’s Disease: Should We Change Our Thinking? Biomolecules, 13(3), 453. https://doi.org/10.3390/biom13030453

Lee, T. M. C., Sun, D., Leung, M.-K., Chu, L.-W., & Keysers, C. (2013). Neural activities during affective processing in people with Alzheimer’s disease. Neurobiology of Aging, 34(3), 706–715. https://doi.org/10.1016/j.neurobiolaging.2012.06.018

Liu, X., Tyler, L. K., Cam-CAN, Davis, S. W., Rowe, J. B., & Tsvetanov, K. A. (2023). Cognition’s dependence on functional network integrity with age is conditional on structural network integrity. Neurobiology of Aging, 129, 195–208. https://doi.org/10.1016/j.neurobiolaging.2023.06.001

Merlo, S., Spampinato, S. F., & Sortino, M. A. (2019). Early compensatory responses against neuronal injury: A new therapeutic window of opportunity for Alzheimer’s Disease? CNS Neuroscience & Therapeutics, 25(1), 5–13. https://doi.org/10.1111/cns.13050

Nasb, M., Tao, W., & Chen, N. (2024). Alzheimer’s Disease Puzzle: Delving into Pathogenesis Hypotheses. Aging and Disease, 15(1), 43. https://doi.org/10.14336/AD.2023.0608

Neuner, S. M., Telpoukhovskaia, M., Menon, V., O’Connell, K. M. S., Hohman, T. J., & Kaczorowski, C. C. (2022). Translational approaches to understanding resilience to Alzheimer’s disease. Trends in Neurosciences, 45(5), 369–383. https://doi.org/10.1016/j.tins.2022.02.005

Ovaska-Stafford, N. (2020). Psychological Resilience Factors in Huntington’s Disease and Other Neurodegenerative Diseases [Thesis, University of Leicester]. https://doi.org/10.25392/leicester.data.12697433.v1

Pappalettera, C., Carrarini, C., Miraglia, F., Vecchio, F., & Rossini, P. M. (2024). Cognitive resilience/reserve: Myth or reality? A review of definitions and measurement methods. Alzheimer’s & Dementia, 20(5), 3567. https://doi.org/10.1002/alz.13744

Park, D. C., & Bischof, G. N. (2013). The aging mind: Neuroplasticity in response to cognitive training. Dialogues in Clinical Neuroscience, 15(1), 109. https://doi.org/10.31887/DCNS.2013.15.1/dpark

Penalba-Sánchez, L., Oliveira-Silva, P., Sumich, A. L., & Cifre, I. (2023). Increased functional connectivity patterns in mild Alzheimer’s disease: A rsfMRI study. Frontiers in Aging Neuroscience, 14, 1037347. https://doi.org/10.3389/fnagi.2022.1037347

Pikouli, F. A., Moraitou, D., Papantoniou, G., Sofologi, M., Papaliagkas, V., Kougioumtzis, G., Poptsi, E., & Tsolaki, M. (2023). Metacognitive Strategy Training Improves Decision-Making Abilities in Amnestic Mild Cognitive Impairment. Journal of Intelligence, 11(9), 182. https://doi.org/10.3390/jintelligence11090182

Požar, R., Martin, T., Giordani, B., & Kavcic, V. (2024). Enhanced functional brain network integration in mild cognitive impairment during cognitive task performance: A compensatory mechanism or a result of neural disinhibition? The European Journal of Neuroscience, 60(7), 5569–5580. https://doi.org/10.1111/ejn.16511

Pupillo, F., Düzel, S., Kühn, S., Lindenberger, U., & Shing, Y. L. (2024). Deficits in memory metacognitive efficiency in late adulthood are related to distinct brain profile. Memory, 32(10), 1286–1302. https://doi.org/10.1080/09658211.2024.2341711

Qiu, L., Su, J., Ni, Y., Bai, Y., Zhang, X., Li, X., & Wan, X. (2018). The neural system of metacognition accompanying decision-making in the prefrontal cortex. PLOS Biology, 16(4), e2004037. https://doi.org/10.1371/journal.pbio.2004037

Ramírez-Barrantes, R., Arancibia, M., Stojanova, J., Aspé-Sánchez, M., Córdova, C., & Henríquez-Ch, R. A. (2019). Default Mode Network, Meditation, and Age-Associated Brain Changes: What Can We Learn from the Impact of Mental Training on Well-Being as a Psychotherapeutic Approach? Neural Plasticity, 2019(1), 7067592. https://doi.org/10.1155/2019/7067592

Rao, Y. L., Ganaraja, B., Murlimanju, B. V., Joy, T., Krishnamurthy, A., & Agrawal, A. (2022). Hippocampus and its involvement in Alzheimer’s disease: A review. 3 Biotech, 12(2), 55. https://doi.org/10.1007/s13205-022-03123-4

Reza-Zaldivar, E. E., Hernández-Sápiens, M. A., Minjarez, B., Gómez-Pinedo, U., Sánchez-González, V. J., Márquez-Aguirre, A. L., & Canales-Aguirre, A. A. (2020). Dendritic Spine and Synaptic Plasticity in Alzheimer’s Disease: A Focus on MicroRNA. Frontiers in Cell and Developmental Biology, 8. https://doi.org/10.3389/fcell.2020.00255

Rosenberg, A., Mangialasche, F., Ngandu, T., Solomon, A., & Kivipelto, M. (2020). Multidomain Interventions to Prevent Cognitive Impairment, Alzheimer’s Disease, and Dementia: From FINGER to World-Wide FINGERS. The Journal of Prevention of Alzheimer’s Disease, 7(1), 29–36. https://doi.org/10.14283/jpad.2019.41

Rudrauf, D. (2014). Structure-Function Relationships behind the Phenomenon of Cognitive Resilience in Neurology: Insights for Neuroscience and Medicine. Advances in Neuroscience, 2014(1), 462765. https://doi.org/10.1155/2014/462765

Saccenti, D., Moro, A. S., Sassaroli, S., Malgaroli, A., Ferro, M., & Lamanna, J. (2024). Neural correlates of metacognition: Disentangling the brain circuits underlying prospective and retrospective second-order judgments through noninvasive brain stimulation. Journal of Neuroscience Research, 102(4), e25330. https://doi.org/10.1002/jnr.25330

Salat, D. H., Kaye, J. A., & Janowsky, J. S. (2001). Selective Preservation and Degeneration Within the Prefrontal Cortex in Aging and Alzheimer Disease. Archives of Neurology, 58(9), 1403–1408. https://doi.org/10.1001/archneur.58.9.1403

Salmon, E., Perani, D., Herholz, K., Marique, P., Kalbe, E., Holthoff, V., Delbeuck, X., Beuthien-Baumann, B., Pelati, O., Lespagnard, S., Collette, F., & Garraux, G. (2006). Neural correlates of anosognosia for cognitive impairment in Alzheimer’s disease. Human Brain Mapping, 27(7), 588–597. https://doi.org/10.1002/hbm.20203

Sandrone, S. (2013). Self through the Mirror (Neurons) and Default Mode Network: What Neuroscientists Found and What Can Still be Found There. Frontiers in Human Neuroscience, 7, 383. https://doi.org/10.3389/fnhum.2013.00383

Select Atrophied Regions in Alzheimer disease (SARA): An improved volumetric model for identifying Alzheimer disease dementia—ScienceDirect. (n.d.). Retrieved November 25, 2024, from https://www-sciencedirect-com.ezproxy.cul.columbia.edu/science/article/pii/S2213158220300851?via%3Dihub

Sigurdsson, T., & Duvarci, S. (2016). Hippocampal-Prefrontal Interactions in Cognition, Behavior and Psychiatric Disease. Frontiers in Systems Neuroscience, 9. https://doi.org/10.3389/fnsys.2015.00190

Souchay, C. (2007). Metamemory in Alzheimer’s Disease. Cortex, 43(7), 987–1003. https://doi.org/10.1016/S0010-9452(08)70696-8

Sun, W., Ueno, D., & Narumoto, J. (2022). Brain Neural Underpinnings of Interoception and Decision-Making in Alzheimer’s Disease: A Narrative Review. Frontiers in Neuroscience, 16. https://doi.org/10.3389/fnins.2022.946136

van Loenhoud, A. C., van der Flier, W. M., Wink, A. M., Dicks, E., Groot, C., Twisk, J., Barkhof, F., Scheltens, P., & Ossenkoppele, R. (2019). Cognitive reserve and clinical progression in Alzheimer disease. Neurology, 93(4), e334–e346. https://doi.org/10.1212/WNL.0000000000007821

Yiannopoulou, K. G., Anastasiou, A. I., Zachariou, V., & Pelidou, S.-H. (2019). Reasons for Failed Trials of Disease-Modifying Treatments for Alzheimer Disease and Their Contribution in Recent Research. Biomedicines, 7(4), 97. https://doi.org/10.3390/biomedicines7040097

Zammit, A. R., Yu, L., Petyuk, V., Schneider, J. A., De Jager, P. L., Klein, H.-U., Bennett, D. A., & Buchman, A. S. (2022). Cortical Proteins and Individual Differences in Cognitive Resilience in Older Adults. Neurology, 98(13), e1304–e1314. https://doi.org/10.1212/WNL.0000000000200017

Zhang, X., Mormino, E. C., Sun, N., Sperling, R. A., Sabuncu, M. R., Yeo, B. T. T., & the Alzheimer’s Disease Neuroimaging Initiative. (2016). Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer’s disease. Proceedings of the National Academy of Sciences, 113(42), E6535–E6544. https://doi.org/10.1073/pnas.1611073113

Zhang, Y., Chen, H., Li, R., Sterling, K., & Song, W. (2023). Amyloid β-based therapy for Alzheimer’s disease: Challenges, successes and future. Signal Transduction and Targeted Therapy, 8(1), 1–26. https://doi.org/10.1038/s41392-023-01484-7

Zhou, A. X., Duncan, J., & Mitchell, D. J. (2024). External task switches activate default mode regions without enhanced processing of the surrounding scene. Imaging Neuroscience, 2, 1–14. https://doi.org/10.1162/imag_a_00185