关键词: 元认知反思, 神经机制, 大规模脑网络, 时空整合, 特异性-协同模型

Abstract: Metacognitive reflection is a central mechanism supporting self-regulated learning and higher-order cognition. Although substantial progress has been made in identifying neural correlates of individual metacognitive judgments, existing findings remain fragmented, focusing on single judgment types or isolated brain regions. As a result, the field lacks an integrative framework that can explain the neural specificity of different reflective processes and their coordination across large-scale networks. To address this gap, the present article develops a unified conceptual account—the Specificity-Synergy Model—and provides a systematic synthesis of behavioral and neural evidence across four major forms of metacognitive reflection.
A key innovation of this work is the construction of a two-dimensional taxonomy, defined by temporal focus (prospective vs. retrospective) and execution timing (immediate vs. delayed). This taxonomy yields four theoretically meaningful reflection types and clarifies long-standing inconsistencies in how metacognitive evaluations are categorized in prior research. More importantly, it reveals important evidence asymmetries: while delayed prospective judgments have been extensively studied, delayed retrospective reflection remains understudied despite its strong theoretical relevance.
Building on this classification, we integrate findings from functional neuroimaging, electrophysiology, neurostimulation, and lesion research to outline the roles of three key systems: the frontoparietal control network (FPCN), the default mode network (DMN), and the salience network (SN). We highlight that each reflection type involves a characteristic constellation of information demands, which in turn modulates the recruitment of specific neural pathways. Immediate prospective judgments rely primarily on fluency-based heuristics encoded in parietal DMN regions, whereas delayed prospective judgments depend on memory reactivation supported by medial temporal lobe structures and DMN hubs. Immediate retrospective reflection engages SN and dorsal anterior cingulate cortex for rapid error detection, while delayed retrospective judgments depend more on reconstructive retrieval and high-level integration within the FPCN.
The proposed Specificity-Synergy Model accounts for these dissociations by emphasizing the dynamic interplay among DMN, SN, and FPCN. DMN provides internally generated evidence, SN detects uncertainty and signals the need for control, and FPCN integrates evidence to support evaluative decisions. This tri-network coordination mechanism extends existing models of metacognition by linking judgment type to temporal shifts in information sources and network engagement. It also aligns metacognitive monitoring with broader large-scale network theories while specifying the computational contributions of each system.
In addition, the article synthesizes causal evidence demonstrating functional specializations within prefrontal and parietal regions. Disruptive stimulation to rlPFC selectively impairs domain-general metacognitive accuracy, whereas parietal perturbation disproportionately affects metamemory. These dissociations provide essential support for type-specific neural pathways proposed in the model.
Based on this integrative account, we outline several productive directions for future research. One important direction concerns dynamic network modeling, such as time-varying connectivity and dynamic causal modeling, which can directly test whether SN-driven signals precede FPCN recruitment under uncertainty or whether delayed retrospective reflection engages unique DMN-FPCN coupling patterns. A second direction calls for enhancing ecological validity, for example through fNIRS hyperscanning or mobile EEG to capture reflective learning processes in naturalistic classroom environments. A third direction involves examining developmental and individual differences, as longitudinal evidence suggests that maturation of connectivity within and between DMN, SN, and FPCN may shape the trajectory of metacognitive abilities.
Finally, the model offers practical insights for education. Because the large-scale networks supporting metacognitive reflection also underpin emotion regulation, self-awareness, and higher-order reasoning, strengthening reflective skills may have broad benefits beyond academic performance. The Specificity-Synergy Model thus provides a neuroscientifically grounded framework for designing reflective learning activities, tailoring instructional support, and developing individualized interventions.
In summary, this article makes three main contributions: it establishes a theoretically coherent taxonomy of metacognitive reflection, delineates the neural circuits underlying distinct judgment types, and articulates a mechanistic model integrating large-scale neural dynamics with metacognitive monitoring. These advances lay the groundwork for future empirical work and offer new avenues for translating metacognitive theory into educational practice.

Key words: metacognitive reflection, neural mechanisms, large-scale brain networks, spatiotemporal integration, Specificity-Synergy Model