Perceivers can use past experiences to make sense of ambiguous sensory signals. However, this may be inappropriate when the world changes and past experiences no longer predict what the future holds. Optimal learning models propose that observers decide whether to stick with or update their predictions by tracking the uncertainty or “precision” of their expectations. However, contrasting theories of prediction have argued that we are prone to misestimate uncertainty—leading to stubborn predictions that are difficult to dislodge. To compare these possibilities, we had participants learn novel perceptual predictions before using fMRI to record visual brain activity when predictive contingencies were disrupted—meaning that previously “expected” events became objectively improbable. Multivariate pattern analyses revealed that expected events continued to be decoded with greater fidelity from primary visual cortex, despite marked changes in the statistical structure of the environment, which rendered these expectations no longer valid. These results suggest that our perceptual systems do indeed form stubborn predictions even from short periods of learning—and more generally suggest that top–down expectations have the potential to help or hinder perceptual inference in bounded minds like ours.

Remembering delayed intentions can be highly demanding. Accuracy in laboratory paradigms assessing prospective memory (PM) is typically well below ceiling, and failure to remember intended behaviors after a delay is a common occurrence in everyday life. However, relatively little is known of the potential differences in brain activity that distinguish successful versus unsuccessful PM. In this fMRI study, participants repeatedly encoded, stored, and then had the opportunity to retrieve intended behaviors while engaged in a distracting ongoing task. This yielded a success rate of approximately two thirds. Overall levels of brain activity distinguished successful versus unsuccessful trials at all three stages (encoding, storage, and retrieval), suggesting multiple neural determinants of PM success. In addition, the voxelwise similarity between patterns of brain activity at encoding and retrieval was greater for successful than unsuccessful trials. This was true even in posterior cingulate, which showed opposite patterns of signal change between encoding and retrieval. Thus, successful realization of delayed intentions may be associated with reinstatement of encoding context at the time of retrieval.

Reality monitoring refers to the process of discriminating between internally and externally generated information. Two different tasks have often been used to assess this ability: (a) memory for perceived versus imagined stimuli; and (b) memory for participant- versus experimenter-performed operations. However, it is not known whether these two reality monitoring tasks share neural substrates. The present study involved use of a within-subjects functional magnetic resonance imaging design to examine common and distinct brain mechanisms associated with the two reality monitoring conditions. The sole difference between the two lay in greater activation in the medial anterior prefrontal cortex when recollecting whether the participant or the experimenter had carried out an operation during prior encoding as compared to recollecting whether an item had been perceived or imagined. This region has previously been linked with attending to mental states. Task differences were also reflected in the nature of functional connectivity relationships between the medial anterior and right lateral prefrontal cortex: There was a stronger correlation in activity between the two regions during recollection of self/experimenter context. This indicates a role for the medial anterior prefrontal cortex in the monitoring of retrieved information relating to internal or external aspects of context. Finally, given the importance of reality monitoring to understanding psychotic symptoms, brain activity was related to measures of proneness to psychosis and schizotypal traits. The observation of significant correlations between reduced medial anterior prefrontal signal and scores on such measures corroborates these theoretical links.

One of the least well understood regions of the human brain is rostral prefrontal cortex, approximating Brodmann's area 10. Here, we investigate the possibility that there are functional subdivisions within this region by conducting a meta-analysis of 104 functional neuroimaging studies (using positron emission tomography/functional magnetic resonance imaging). Studies involving working memory and episodic memory retrieval were disproportionately associated with lateral activations, whereas studies involving mentalizing (i.e., attending to one's own emotions and mental states or those of other agents) were disproportionately associated with medial activations. Functional variation was also observed along a rostral-caudal axis, with studies involving mentalizing yielding relatively caudal activations and studies involving multiple-task coordination yielding relatively rostral activations. A classification algorithm was trained to predict the task, given the coordinates of each activation peak. Performance was well above chance levels (74% for the three most common tasks; 45% across all eight tasks investigated) and generalized to data not included in the training set. These results point to considerable functional segregation within rostral prefrontal cortex.