Interindividual Variability in Cognitive Aging: Neural Mechanisms, Reserve Processes and Malleability of Neuroprotective Processes.

Aging leads to cognitive changes that affect memory, particularly episodic retrieval. These impairments are detrimental to seniors' quality of life. Cognitive trainings are of great interest to the scientific community because they improve cognition in older people, and produce structural and functional changes likely to provide neuroprotection. Identifying the brain changes induced by cognitive training could therefore provide a better understanding of the neuroplastic processes of the aging brain. Some training programs aim to improve key processes underlying cognitive functioning to lead to transfer, but these most often target working memory or processing speed. Our aim is to understand the brain changes associated with a training program targeting episodic retrieval, and likely to engage a core network for memory, including the anterior hippocampus. 60 healthy older adults will be randomly divided into two groups; one receiving a training based on the Episodic Specificity Induction (ESI) - a manipulation based on a well-established police interviewing technique thought to target and facilitate episodic construction; the other receiving a control training consisting of recalling pairs of words and images. Before and after training, behavioural and brain measures will be taken. Behavioural measures will be taken during recall, recognition, and problem solving tasks. These tasks will be completed once in the ESI condition (after one ESI) and once in the NoESI condition (after a general thoughts interview). Measures of brain activation as well as static and dynamic functional connectivity (SFC \& DFC) will be taken using magnetic resonance imaging (MRI) during a recognition task. For behavioural measures, 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. For brain measures, ESI training should decrease activation of the task network targeted by training, reflecting an increase in efficiency. ESI training should also increase the SFC of the task network and reduce its connectivity with the cognitive control network, suggesting more automated processing. Finally, ESI training should increase DFC by increasing the speed of transition between the networks associated with the two phases of episodic retrieval: the construction phase and the elaboration phase.