Metamemory monitoring, the process of making subjective assessments of the status of one’s own memory, is crucial to guiding behavior and effective learning. Past cognitive research has shown that subjective confidence judgments are inferential in nature, and based on cues available at the time of the judgment. When confidence is based on cues that are related to objective memory performance, metamemory accuracy is high. However, past studies have shown that metamemory monitoring tends to be inaccurate because individuals base their confidence on information that is not predictive of memory success, such as the fluency with which items were encoded during study, or invalid information about task difficulty from external sources. Brain research has lagged behind cognitive research in establishing the neural bases of metamemory monitoring, but there is evidence that the prefrontal cortex contributes to Judgments of Learning (JOLs), subjective confidence judgments made during encoding about the ability to retrieve information at later test. Retrospective Confidence Judgments (RCJs), which are made at retrieval and reflect one’s confidence in his or her response, have been associated with activity in the prefrontal, parietal, and temporal lobes. However, there is evidence that the roles of each of these regions in metamemory monitoring may vary with the information on which confidence is based (eg, fluency of encoding or information about task difficulty), and here we present two experiments which tested how the neural mechanisms underlying JOLs and RCJs interact with the cognitive bases of the judgments. Experiment 1 (Chapter 2) used transcranial direct current stimulation (tDCS) to directly manipulate activity in the dorsolateral prefrontal cortex (DLPFC) and anterior prefrontal cortex (aPFC) during a JOL task in which subjects studied words that varied in their fluency at encoding. Results showed that DLPFC stimulation impaired encoding for both fluent and disfluent conditions, and aPFC stimulation enhanced JOL accuracy for disfluent encoding conditions. Conversely, DLPFC and aPFC stimulation decreased JOL accuracy for fluently encoded conditions, suggesting the causal roles of the aPFC and DLPFC in JOL accuracy vary depending on whether JOLs are based on the cue of fluency or disfluency.