Efficacy, Transfer, and Neuro-functional Basis of a Memory Training Targeting Episodic Retrieval in Older Adults.

Aging leads to cognitive changes that affect memory, particularly episodic retrieval, which is involved in remembering an event within a specific spatio-temporal context. These impairments are detrimental to seniors' quality of life. Recently, cognitive training has attracted considerable interest in the scientific community, as it improves cognition in older adults, and produces structural and functional changes. They are considered a form of "late education" likely to provide neuroprotection. A better understanding of training-induced brain changes is therefore essential to a better understanding of the neuroplastic potential of the aging brain. A distinction is made between training that aims to teach strategies, and training that aims to repeatedly practice a task that involves a key process to facilitate it. The latter are recognized as having greater potential to induce a transfer of their benefits to untrained abilities. According to the INTERACTIVE model, they would lead to a decrease in activation in the brain regions involved in the task, due to more efficient processing within specialized regions. Yet there are very few training programs that target episodic retrieval. Developing such trainings would therefore contribute to our understanding of training-induced brain changes. One key process involved in episodic retrieval that such training could target is scene construction. This corresponds to the ability to mentally generate and maintain the coherence of a complex scene. It would engage the anterior hippocampus, especially its medial part, and a "core" network comprising the medial temporal lobes, parahippocampal cortex, and posterior parietal regions such as the retrosplenial cortex and precuneus. Studies suggest that this process may be targeted and facilitated when participants are interviewed about their memory of a video via an Episodic Specificity Induction (ESI) - a manipulation based on a well-established police interviewing technique knwon as the Cognitive Interview (CI). This technique promotes mental imagery, structures recall, and emphasizes spatial relationships between entities in a scene. At the behavioral level, it has been shown in both young and old adults that receiving a single ESI just before performing a task of interest involving scene construction temporarily improves performance on that task. At the brain level, the ESI has been shown in young adults to increase activation of brain regions involved in scene construction during the task of interest, specifically during the construction phase. This phase corresponds to the retrieval of the event representation, and precedes the elaboration phase, associated with the retrieval of additional details (see construction-elaboration paradigm). These two phases have common and distinct neural correlates. The construction phase is associated with stronger connectivity of the left anterior hippocampus with the right anterior hippocampus and left frontotemporal regions, while the elaboration phase is associated with stronger connectivity of the left anterior hippocampus with the bilateral posterior hippocampus and visual perception regions. However, the ESI has never been adapted within a training program to study its effects on the aged brain. In a proof-of-concept study, we tested on 16 healthy older adults an ESI-based training program that we co-created with potential future end-users. We observed transfer to free recall and recognition tasks, but also to a problem-solving and divergent creative thinking tasks, which are not purely memory tasks, but which would involve scene construction. In the present study, we aim to compare the ESI-based training with an active control training and investigate the neural substrates involved.

60 healthy older adults (60-85 years) will be recruited for this study. All participants will be recruited from the community and living in the Montreal area. A telephone interview will provide initial selection information. Eligible persons will be invited to come to the laboratory for a short neuropsychological assessment to evaluate their clinical status and cognitive functioning, as well as a behavioural assessment (PRE). Concerning behavioural tasks, three tasks from the proof-of-concept study will be used: free recall and recognition (as nearest- and near-transfer outcomes) and problem solving (as far-transfer outcome). These tasks will be completed twice, once in the ESI condition right after receiving one ESI; once in the NoESI condition right after receiving a general thoughts interview. Behavioural assessment session will last 2 hours. The pre-training behavioural assessment session will be followed on average two days later by a functional magnetic resonance imaging (fMRI) scan. Concerning the fMRI task: outside the scan, participants will first encode short videos of scenes from everyday life along with their titles. Then, in the scan, each title will be presented and participants will answer questions about the spatial relationships between objects that were present in the video associated with the title. The fMRI sessions will last 1 hour (including 30 minutes in the scanner). Participants will be randomly assigned to one of two training groups (ESI vs. Active control). The ESI-based training consists in recalling short video clips using the ESI, first supervised, then unsupervised (self-administered). The active control training consists in recalling associations between pictures and words. Participants will be trained in small groups of 4 individuals, for a total of 15 groups. On average, outcome measures will be taken a week prior the first training session (PRE), and the day after the last training session (POST). Different version of the tasks will be used in the pre- and post training assessment, as well as, in the ESI and NoESI conditions.

The effect of training on behavioural measures will be assessed. For the three tasks (free recall, recognition and problem solving), higher pre-training performance should be observed in the ESI than in the NoESI condition, and pre-to-post-training improvement should be observed only after the ESI training, especially in the NoESI condition. The effect of training on brain activation and static and dynamic functional connectivity (SFC \& DFC) will also be assessed. Concerning brain activation, activation of the medial temporal lobe during the construction phase, and especially the left anterior hippocampus and inferior parietal lobe involved in scene construction, should decrease after the ESI training, suggesting an increase in brain efficiency. For static FC, the ESI training should increase the modularity of the core brain network related to episodic processing, including the hippocampus, bilateral parahippocampal gyrus, angular gyrus, medial prefrontal cortex, precuneus and retrosplenial/posterior cortex. It should also reduce communication with the fronto-parietal brain network related to cognitive control (i.e., bilateral lateral prefrontal cortex, anterior cingulate and precuneus), reflecting greater segregation and more automated processing. For dynamic FC, we will look at temporal reconfigurations of FC within and between task-associated networks, before and after training. ESI-trained participants should be faster to switch from an integrated network for the construction phase, to an integrated network for the elaboration phase, suggesting an adaptation of cognitive processing strategies.