Metacognition and theory of mind

Listen to audio

Conscious metacognition involves representing one's own mind as representing information (" I think I've forgotten my keys"). The format of these meta-representations seems very similar to that assumed to underlie the representation of others' thoughts (" he thinks I've forgotten my keys"). In both cases, the mental representation must specify the agent (me or someone else), the mental attitude (believe, know...), and the proposition being examined. Could it be, then, that we use the same mental representation format and the same brain areas to represent our minds and those of others? Do conscious metacognitive thinking and "theoryof mind" involve, at least in part, the same mechanisms?

A number of empirical arguments suggest that self-knowledge and knowledge of others are closely linked. First of all, they develop simultaneously in children: it's at the same age that children begin to understand the minds of others and to have a metacognitive representation of their own competence (Gopnik & Astington, 1988). Very recent findings suggest that it is at a very early age, around 7 months, that the theory of others' minds is established. From this age, children represent their own knowledge and that of others in the same format, so that they interfere with each other (Kovacs, Teglas & Endress, 2010).

There is also evidence of a generalization of self-knowledge to knowledge of the other in 12-month-olds (Meltzoff & Brooks, 2008).
At this age, children follow an adult with their gaze when the adult turns their head with eyes open, but not with eyes closed - suggesting that they may understand what " see " means. However, they also follow a person wearing a blindfold. Perhaps this means they need personal experience to understand how others feel. After all, they've already had the experience of closing their eyes, but since they've never had the experience of being blindfolded, they don't understand that wearing a blindfold prevents them from seeing. Meltzoff and Brooks (2008) confirm this hypothesis by training children either with an opaque blindfold, or with a transparent or apertured blindfold. The child's personal experience alters his or her understanding of other people's minds - which implies that self-representation and other people's theory of mind share common representations.

Another argument comes from the analysis of the cerebral networks involved. A network comprising the antero-mesial prefrontal cortex, the precuneus, the temporo-parietal junction (particularly on the right) and the anterior part of the temporal lobe is involved in theory of mind, particularly in false-belief tasks. Metacognitive self-reflection also activates a fraction of this network (Jenkins, Macrae & Mitchell, 2008; Ochsner et al., 2004; Vogeley
et al., 2004), particularly the frontopolar and ventromesial prefrontal cortex. The representation of other people's mistakes evokes an "error negativity", associated with the anterior cingulate cortex, similar to that evoked more classically by our own mistakes (van Schie, Mars, Coles & Bekkering, 2004). People with autism, who have impaired representation of the thoughts of others, also suffer from abnormalities of self-representation in the mesial prefrontal cortex (Lombardo et al., 2010).

Thus, metacognition and the representation of others share certain brain mechanisms. However, the interpretation of this overlap remains ambiguous. Either we have a detailed representation of ourselves and use this " network of the self " to simulate the minds of others and try to understand them; or we don't have a specific system of introspection, but our knowledge is based on repeated self-observation: we represent our behavior and infer our state of mind as we would from that of another person, but we simply have a little more data about ourselves than about others.

In any case, the ability to be aware of our capabilities and limitations plays an essential role in dialogue with others. A recent experiment shows that two people observing the same stimulus can achieve, through dialogue, an optimal psychophysical decision, better than that made by only one of the two protagonists (Bahrami et al., 2010). This optimal success depends on each of the protagonists explicitly representing their own performance and the degree of confidence they can place in it. The verbal exchange of this subjective information implies that it is put into a representation format common to self and other. We can therefore speculate that introspection (fictional or otherwise) and theory of mind are two facets of the same system of mental representation, which plays an essential role in the social dialogue specific to the human species.