From maintenance in working memory to long-term consolidation – REFLECTOR

The REFLECTOR project is structured around three main methodologies, each exploring the objectives from complementary angles:

- Behavioral approach - Through “complex span” tasks (where a memorization task is associated with a distraction activity), the effect of cognitive load on memory performance is tested. Participants, confronted with word lists in conditions of low or high cognitive load, are evaluated on their immediate and delayed recall. Other variables are manipulated to assess factors modulating short- and long-term memory interactions: in particular, repetition, familiarity of material and participant motivation.

- EEG neural studies - The neural correlates of refreshment are recorded by EEG, observing theta and alpha oscillations. By varying the number of items to be memorized, we aim to identify the links between neural activity and memory capacity. A recall test is also carried out in the short and long term, and performance is associated with the recorded brain oscillations. A condition in which participants maintain via verbal rehearsal and not attentional refreshing serves as a control condition.

- Animal model for synaptic consolidation - An animal model (rat) explores the relationship between WM and consolidation. Electrodes record activity in the prefrontal cortex and hippocampus during an MDT task, then during sleep to assess the effect of sleep-dependent consolidation. Correlations between brain oscillations, theta-gamma coupling, and synaptic plasticity are analyzed to understand how refreshment may influence durable memory.

These integrated approaches, combining behavior, EEG and animal models, aim to provide multi-dimensional data for understanding MDT processes in consolidation and applications in learning.

As part of our ANR Reflector project, we have published a theoretical review that addresses forgetting in working memory (WM) by bringing together human and animal studies. This review proposes that neural models, and in particular animal models, are essential tools for exploring the precise mechanisms of forgetting in WM. We review theoretical perspectives on forgetting in WM in humans, then explore its neural correlates in animals, from early observations of delayed activity in the prefrontal cortex to more recent theories of synaptic WM. Next, we discuss specific theories of WM, introducing in particular the idea that silent activity might better correspond to the processes of refreshing and decay proposed in human cognitive models. The review concludes by exploring the links between long-term memory and WM, revealing connections via the long-term synaptic hypothesis, which suggests that long-term storage of interference can potentially disrupt WM (Malleret, Salin, Mazza, & Plancher, 2024). We also conducted several studies to better understand the bidirectional links between WM and long-term memory (LTM). We explored a new hypothesis: attentional refreshing requires stable WM representations, irrespective of the presence of associated LTM traces. Although factors such as lexicality, short-term consolidation and multiple repetitions influenced short- and long-term recall performance, both experiments showed that these factors had no effect on the cognitive cost of the simultaneous task, suggesting that refreshing is not influenced by LTM content or the stability of LTM representations (Labaronne, Jarjat, & Plancher, 2023). In another study, we manipulated long-term memory intention, either by informing participants of a final delayed recall, or by offering a monetary reward for immediate or delayed recall performance. Our results show that intention did not alter maintenance mechanisms in MDT, but that cognitive load and rewards increased effects on delayed recall, particularly when intention to memorize in the long term was strong (Labaronne, Ferreri, & Plancher, 2023). Finally, we conducted an EEG study to analyze attentional refreshing oscillatory and verbal rehearsal. The results show that verbal rehearsal enhances short-term recall, while attentional refreshing promotes delayed recall. In addition, an increase in frontal-midline theta power was observed with increased memory load only during verbal rehearsal, while greater occipito-parietal alpha desynchronization was found with attentional refreshing.

The results expected from REFLECTOR could redefine our understanding of memory and influence a number of fields. The prospects are numerous, particularly for education and cognitive management of learning. By proving that active attention and working memory “refreshing” promote information consolidation, we could encourage new pedagogical strategies where regular reactivation of information in working memory would be integrated into curricula to maximize memorization.

On the therapeutic front, a better understanding of consolidation mechanisms could benefit the elderly, where mnesic consolidation is often affected. Targeted interventions, based on “attentional refreshing” and sleep management, could slow age-related memory loss.

Lastly, the animal model opens up the possibility of pharmacological research into synaptic consolidation processes to improve long-term memory in the context of cognitive disorders. Enhancing theta-gamma coupling and hippocampal activity during sleep could pave the way for solutions to pathologies such as memory disorders and Alzheimer's disease.

This interdisciplinary work aims to link the theories of working memory and consolidation, and proposes a breakthrough in the global understanding of memory and its role in learning.

The ambition of the present proposal is to examine the cognitive and neuronal mechanisms of information storage in memory from the very beginning when information is present in working memory, until the late stage of sleep-dependent long-term consolidation of this information. We will investigate these mechanisms in humans and in animals (rats), the animal model offering a more direct measurement of cognitive and neuronal mechanisms of memory. The general and original hypothesis tested in the project is that a particular form of retrieval in working memory, attentional refreshing, influences the way information is consolidated in long-term memory. We will: 1. Measure the impact of maintaining items in working memory through attentional refreshing at various delays: at immediate, delayed and after sleep recalls (Workpackage 1); 2. Investigate the neuronal correlates of attentional refreshing in WM and relate these neuronal correlates with memory performance at various delays (Workpackage 2); 3. Investigate the neuronal correlates of refreshing and long-term consolidation in an animal model (Workpackage 3). Through different experiments, we will collect behavioral data (percentage of correct recalls at various delays) and neuronal data (EEG). Expected findings will: 1. Determine whether maintenance in working memory through attentional refreshing has an impact on long-term memory and whether refreshed items are those that preferentially enter the process of sleep-dependent consolidation; 2. Determine whether brain oscillations are influenced by the use of attentional refreshing in working memory and determine the relation between these oscillations and memory performance; 3. Determine, using an animal model, if an increase in the theta-gamma coupling between the hippocampus and medial prefrontal cortex predicts memory performance; and how this increase could be linked to an increase in synaptic transmission during sleep stages responsible for the long-term consolidation of items previously processed in working memory. By offering a large temporal scale of investigation, our project will provide some crucial information about the role of attentional refreshing in working memory at short and long-term. Finally, a challenge for our project is to bring together the study of memory in humans and in animals. Beyond its utility for our understanding of cognitive and neuronal mechanisms of memory, this association would enable to propose a more general theoretical framework of working memory. In the long-term, because working memory is central in our daily life, our results may help to fight against underperforming at school and improve the well-being of populations with working memory deficits.