On the Antecedents of an Electrophysiological Signature of Retrieval Mode

Neural evidence that disengaging memory retrieval is modulated by stimulus valence and rumination

To remember information from our personal past we need to be in a cognitive state where we treat stimuli as cues for memory retrieval. In this study we considered whether participants could exert control and disengage from a memory state when it was no longer required for the task at hand. In particular, we examined whether this ability was affected by the valence of the stimuli and participant's rumination scores. After a study phase participants completed test blocks where the task switched every two trials between a memory task (retrieving information from the study phase) and a perceptual task. Even though there was no episodic memory requirement in the perceptual task, a well-established event-related potential (ERP) index of memory retrieval was present for both trials when the stimuli were negative valenced pictures but not for neutral pictures. Furthermore, there was a positive correlation between the magnitude of this ERP memory index in the perceptual task and rumination scores but only for neutral stimuli and not negative. Thus, in this study participants generally had difficultly suppressing memory retrieval when negative stimuli were presented. However, for neutral stimuli only ruminators were more susceptible to memory intrusions.

The role of emotion arousal in the retrieval practice effect

Many studies have shown that practicing retrieval produces better memory retention compared to restudy. Though previous literature has provided valuable insights about the retrieval practice effect, it is still unclear how emotion arousal influences the retrieval practice effect, and whether the effect would be manifested in recollection or familiarity processes. To answer these questions, in the current study, negative and neutral words were used as stimuli and participants were asked to perform a recognition test or restudy the words after initial study. At the end of the experiment, a final recognition test with involving the remember-know paradigm was shown. Behavioral data were collected with EEG recorded throughout the experiment. The behavioral retrieval practice effect was only found for the neutral but not the negative words. Consistently, significant ERP differences between the restudy and retrieval practice conditions were only found for neutral, but not negative items, which was a component from 700 to 900 ms at left-posterior electrode cluster. Moreover, we found that the effects of emotion arousal on the retrieval practice effect were mainly reflected in the recollection process. These findings provide behavioral and neural evidence that emotion arousal can influence the retrieval practice effect.

On the sensitivity of event-related potentials to retrieval mode

Event-related potential (ERP) signatures of preparation to retrieve episodic memories have been identified in several studies. A common finding is relatively more positive-going ERP activity over right-frontal sites when people prepare for episodic rather than semantic retrieval. This activity has been linked to the process of retrieval mode - a retrieval set that ensures subsequent events are treated as cues for episodic retrieval. This experiment was designed to test one explanation for why this putative index of retrieval mode was not observed in two recent experiments. Towards this end, ERPs were recorded time-locked to different task-cues indicating which of two retrieval tasks participants should prepare to complete. Each task-cue was followed by a retrieval-cue that required a memory judgment. Departures from the designs of the two studies in which null ERP results were obtained were intra-trial timings and the order in which task cues were presented. Frequentist statistics revealed that ERPs elicited by the task-cues did index preparation to retrieve. The topographies of these activities, however, did not overlap markedly with that of the putative index of retrieval mode reported previously. Bayesian analyses, moreover, provided little compelling evidence for a signature of retrieval mode. These outcomes prompt consideration of how ERP sensitivities to preparatory retrieval processing should be characterized.