Selective Rehearsal and Active Inhibition Allow us to Intentionally Forget New Information

Post by Stephanie Williams 

What's the science?

The mechanisms that allow us to intentionally forget or remember new information have been a topic of debate in the neuroscience community. There is some evidence that we may reduce our selective rehearsal of that information. There is also evidence that we may actively suppress unwanted memory traces. This week in Current Biology, Fellner and colleagues characterize electrophysiological signatures that show how we recruit both active suppression and passive reduced recruitment in order to suppress unwanted memories. 

How did they do it?                             

The authors recorded electroencephalography (EEG) while 23 subjects performed a task in which they tried to forget or remember real-world objects. The authors used EEG to track how objects were represented, and how those representations changed between forgotten and remembered items. The memory task consisted of a large number of objects, which would flash briefly on a screen. Participants were cued to either forget or remember the item that they had just seen. After viewing all objects, subjects were distracted by a 3-minute counting exercise, and then viewed a new series of images consisting of a mixture of old and new objects. They were then asked to rate how well they recognized each of the images on a 1-6 scale that represented their confidence in their recognition. The authors used the participant’s recognition responses to track whether objects were “forgotten” or “remembered”. Before the experiment began, all subjects were told that they should voluntarily forget the to-be-forgotten objects as they wouldn’t be tested on those objects, so that they would be able to better remember the to-be-remembered objects. 

Although subjects were always cued to remember or forget, sometimes participants forgot objects they had been cued to remember and remembered objects they had been cued to forget. This allowed the authors to group trials into unintentional forgetting (subjects intended to remember, but didn’t), and intentional forgetting (subjects intended to forget, and did indeed forget). The same applied to unintentional remembering and intentional remembering. To understand how participants’ representations of objects changed depending on whether they were asked to forget to remember the object, the authors calculated “item-cue similarity”: the correlation between the electrophysiological activity 1) during the object presentation and 2) right after the memory cue appeared. The authors used different combinations of time windows to calculate the item-cue similarity scores, which resulted in a matrix of scores. They then applied cluster-based permutation statistics to the matrix of scores to identify clusters of activity. Their results led them to focus on a spectral band between 8 and 13 Hz called the alpha band. They used source localization techniques in order to understand which brain areas (eg. occipital vs. parietal areas) contributed to the patterns that they found. They used these results to compare the topography of item representation changes and alpha power changes. 

What did they find?

As expected, the participant’s ratings showed that on average they did remember the “to be remembered’ items better than they remembered the “to be forgotten” items. When the authors applied cluster-based permutation statistics to the matrix of item similarity score patterns, they found two major clusters, an early and a late cluster. When the authors compared the item-cue-similarity scores during the early cluster in the intentional and unintentional forgetting groups, they found that the scores were reduced during intentional forgetting compared to unintentional forgetting, consistent with the changes in object representation expected for active suppression. If participants were actively suppressing item representations, then the item representations would be lower when subjects were intentionally forgetting objects than when they were attempting (but failing) to remember objects. When the authors compared the item-cue-similarity scores during the late cluster period, they found that the scores were significantly higher for the intentionally remembered versus the unintentionally remembered items, consistent with changes in object representation expected for reduced rehearsal. 

The authors also found a significant increase in spectral power in the alpha band during forgotten trials. The authors split the alpha power changes into two broad categories, early and late. Early alpha power was mostly confined to the occipital electrodes, while late alpha power extended across occipital, posterior temporal, parietal, and posterior midline. The authors found that early alpha power only increased for forgotten items. Their results suggest that early occipital alpha power increases during successful suppression of new information, and may represent active downregulation of memory traces. Late alpha power, which was accompanied by beta power increases, was increased for intentionally remembered objects relative to unintentionally remembered objects. These results suggest that late alpha power may index selective rehearsal, and represent a suppression of potentially distracting information. Topographically, the authors found that the overlap of alpha power changes with item representations tracked successful voluntary forgetting.

What's the impact?

The authors show that two distinct processes, reduced rehearsal, and increased suppression of memory traces, both independently contribute to intentional control of memories. Their findings advance our understanding of how memory formation can be voluntarily controlled. 

Fellner et al. Tracking Selective Rehearsal and Active Inhibition of Memory Traces in Directed Forgetting. Current Biology. (2020). Access the original scientific publication here.