Metacognition and Cognitive Flexibility in Autistic and Neurotypically-Developing Populations

Whether and how metacognition is altered in individuals with autistic spectrum disorder (ASD) is intensely debated. Metacognitive deficit is claimed to be related to cognitive inflexibility, accounting for restrictive behaviors in ASD individuals. We wanted to test this hypothesis by measuring metacognition in ASD and in matched neurotypically developing (TD) control samples in a task that relies on visuo-spatial cognition, in which ASD allegedly have an advantage.

We measured metacognition in a 3D mental rotation task. Additionally, we administered a trading game: players had to figure out the rules for maximizing the profit on each transaction. These rules changed in the middle of the game, which required that players modify their strategy to keep the profit at maximum. We measured both learning efficiency (how fast players extract the rules) and re-learning speed (cognitive flexibility, how fast learners could adjust their behavioral responses after rules are changed).

TD outperform ASD individuals in terms of accuracy in mental rotation but exhibited lower metacognitive efficiency (i.e., were less aware when they were more likely to make an error). No differences in learning efficiency and cognitive flexibility between TD and ASD individuals were observed. Neither did we observe an association between cognitive flexibility and metacognition. Nevertheless, both in ASD and TD populations, overconfidence in one's decisions is negatively correlated with cognitive flexibility, but not with learning efficiency.

ASD individuals can have superior metacognition in tasks that rely on visuo-spatial cognition. Cognitive flexibility is diminished by overconfidence, not by metacognitive deficit.