Behavioural and neural evidence for the impact of fluency context on conscious memory

It has been recently suggested that fluency may impact recognition memory performance when the fluency context varies from trial-to-trial. Surprisingly, such an effect has proved difficult to detect in the masked priming paradigm, one of the most popular means to increase fluency-based memory judgements. We conducted a functional magnetic resonance imaging (fMRI) experiment in which participants encoded words at study and, at test, performed a recognition memory task within a masked priming procedure. In order to optimise the chances of finding priming effects on recognition memory performance, we used low-frequency words, which have been shown to increase hits relative to false alarms and enhance masked priming effects. Fluency context was manipulated by either mixing primed and unprimed trials [Random context (RC) experiment] or blocking primed and unprimed trials [Blocked context (BC) experiment]. Behaviourally, priming affected high-confidence memory performance only in the RC experiment. This behavioural effect correlated positively with neural priming in several recognition memory regions. Moreover, we observed a functional coupling between the left middle temporal gyrus and the left parietal and posterior cingulate cortices that was greater for primed relative to unprimed words. In contrast, in the BC experiment, despite similar activity in recognition-memory-related regions, we did not find any significant correlations between neural and behavioural priming. Finally, we observed striking differences in the neural correlates of masked priming between the RC and BC experiments not only in location but also in direction of the neural response. Possible implications of these findings are discussed.