Project Summary / Abstract
Humans possess a remarkable ability to monitor their own mental processes, a property termed
metacognition¿¿. The proposed research investigates the neural mechanisms that underlie this capacity.
Metacognitive judgments are important in everyday life and educational settings, where they play an important
role in helping us to evaluate our own decisions¿1,2¿. Moreover, metacognition is impaired in a number of clinical
conditions¿3-5¿. The proposed research is therefore likely to yield important clinical insights.
The proposed research will address two fundamental questions about the neural mechanisms underlying
metacognition. The first question concerns whether metacognition in the perceptual and memory domains is
supported by the same mechanisms, or by ¿domain-specific mechanisms¿¿. Recent work has suggested that
domain-specific mechanisms may exist in both the ¿anterior prefrontal cortex (aPFC)¿¿13,14,17¿ and the
precuneus¿¿13,14,16¿. The proposed research will provide a causal test of these hypotheses by bidirectionally
manipulating the content of specific representations in these regions.
The second question concerns the distinction between ¿metacognitive bias¿¿ -- one’s overall level of
confidence -- and ¿metacognitive accuracy¿¿ -- the extent to which decision confidence is predictive of decision
accuracy¿23¿. A number of experiments have investigated the brain basis of metacognitive bias, but the brain
basis of metacognitive accuracy is less well understood. Recent proposals have hypothesized that the ¿aPFC
plays an important role in metacognitive accuracy by integrating multiple sources of information that may
inform a confidence judgment¿15,24¿. The proposed research will test this hypothesis by bidirectionally
manipulating the content of specific representations in the aPFC.
To accomplish the proposed research goals, we will use real-time fMRI Decoded Neurofeedback
(DecNef)¿15¿. DecNef allows the content of specific neural representations to be manipulated, even allowing
different representations within the same brain region to be controlled separately. Because many of the
representations of interest overlap within the same region, DecNef is the perfect tool for manipulating these
representations in a precise manner, and provides numerous advantages over other approaches. We plan to
capitalize on these advantages to answer fundamental questions about the brain basis of metacognition which
could not be addressed using previously available approaches. The sponsor’s laboratory is one of a few
research groups with extensive expertise in DecNef¿12,20,22¿, and will be an excellent environment for training in
this technique. Training in DecNef will add an important tool to the applicant’s previous training in fMRI, brain
stimulation techniques, and computational modeling, and will prepare the applicant to transition to their next
stage as an independent researcher.