Metacognition: ideas and insights from neuro- and educational sciences

Metacognition is defined as “thinking about thinking” or the ability to monitor and control one’s cognitive processes and plays an important role in learning and education. For instance, high performers tend to present better metacognitive abilities (especially control) than low performers in diverse educational activities. Recently, there has been a lot of progress in studying the neural mechanisms of metacognition, yet it is unclear at this point how these results may inform educational sciences or interventions. Given the potential benefits of metacognition, it is important to get a better understanding of how metacognition works and of how training can be useful.

The interest in bridging cognitive neuroscience and educational practices has increased in the past two decades, spanning a large number of studies grouped under the umbrella term of educational neuroscience. With it, researchers have brought forward issues that are viewed as critical for the discipline to improve education. Recurring issues that may impede the relevance of neural insights for educational practices concern external validity, theoretical discrepancies and differences in terms of the domains of (meta)cognition operationalised (specific or general). This is important because, in recent years, brain research is starting to orient itself towards training metacognitive abilities that would translate into real-life benefits. However, direct links between metacognition in the brain and metacognition in domains such as education have still to be made. As for educational sciences, a large body of literature on metacognitive training is available, yet we still need clear insights about what works and why. While studies suggest that training metacognitive abilities results in higher academic achievement, other interventions show mixed results. Moreover, little is known about the long-term effects of, or transfer effects, of these interventions. A better understanding of the cognitive processes involved in metacognition and how they are expressed in the brain may provide insights in these regards.

Within cognitive neuroscience, there has been a long tradition of studying executive functions (EF), which are closely related to metacognitive processes. Similar to metacognition, EF shows a positive relationship with learning at school. For instance, performance in laboratory tasks involving error monitoring, inhibition and working memory (i.e. processes that monitor and regulate cognition) are associated with academic achievement in pre-school children. More recently, researchers have studied metacognition in terms of introspective judgements about performance in a task. Although the neural correlates of such behaviour are being revealed, little is known about how behaviour during such tasks relates to academic achievement.

Educational and cognitive neuroscientists study metacognition in different contexts using different methods. Indeed, while the latter investigate metacognition via behavioural task, the former mainly rely on introspective questionnaires. The extent to which these different operationalisations of metacognition match and reflect the same processes is unclear. As a result, the external validity of methodologies used in cognitive neuroscience is also unclear. We argue that neurocognitive research on metacognition has a lot of potential to provide insights in mechanisms relevant in educational contexts, and that theoretical and methodological exchange between the two disciplines can benefit neuroscientific research in terms of ecological validity.

For these reasons, we investigate the literature through the lenses of external validity, theoretical discrepancies, domain generality and metacognitive training. Research on metacognition in cognitive neuroscience and educational sciences are reviewed separately. First, we investigate how metacognition is operationalised with respect to the common framework introduced by Nelson and Narens (see Fig. 1). We then discuss the existing body of evidence regarding metacognitive training. Finally, we compare findings in both fields, highlight gaps and shortcomings, and propose avenues for research relying on crossovers of the two disciplines.

In cognitive neuroscience, metacognition is divided into two main components, which originate from the seminal works of Flavell on metamemory. First, metacognitive knowledge (henceforth, meta-knowledge) is defined as the knowledge individuals have of their own cognitive processes and their ability to monitor and reflect on them. Second, metacognitive control (henceforth, meta-control) consists of someone’s self-regulatory mechanisms, such as planning and adapting behaviour based on outcomes. Following Nelson and Narens’ definition, meta-knowledge is characterised as the flow and processing of information from the object level to the meta-level, and meta-control as the flow from the meta-level to the object level (Fig. 1). The object-level encompasses cognitive functions such as recognition and discrimination of objects, decision-making, semantic encoding, and spatial representation. On the meta-level, information originating from the object level is processed and top-down regulation on object-level functions is imposed.

Educational researchers have mainly investigated metacognition through the lens of Self-Regulated Learning theory (SRL), which shares common conceptual roots with the theoretical framework used in cognitive neuroscience but varies from it in several ways. First, SRL is constrained to learning activities, usually within educational settings. Second, metacognition is merely one of three components, with “motivation to learn” and “behavioural processes”, that enable individuals to learn in a self-directed manner. In SRL, metacognition is defined as setting goals, planning, organising, self-monitoring and self-evaluating “at various points during the acquisition”. The distinction between meta-knowledge and meta-control is not formally laid down although reference is often made to a “self-oriented feedback loop” describing the relationship between reflecting and regulating processes that resembles Nelson and Narens’ model (Fig. 1). In order to facilitate the comparison of operational definitions, we will refer to meta-knowledge in educational sciences when protocols operationalise self-awareness and knowledge of strategies, and to meta-control when they operationalise the selection and use of learning strategies and planning. For an in-depth discussion on metacognition and SRL, we refer to Dinsmore et al..