Neural Reinstatement of Overlapping Memories in Young and Older Adults

Abstract

When we update our episodic memories with new information, mnemonic competition between old and new memories may result because of the presence of shared features. Behavioral studies suggest that this competition can lead to proactive interference, resulting in unsuccessful memory updating, particularly for older adults. It is difficult with behavioral data alone to measure the reactivation of old, overlapping memories during retrieval and its impact on memory for new memories. Here, we applied encoding–retrieval representational similarity (ERS) analysis to EEG data to estimate event-specific encoding-related neural reinstatement of old associations during the retrieval of new ones and its impact on memory for new associations in young and older adults. Our results showed that older adults' new associative memory performance was more negatively impacted by proactive interference from old memories than that of young adults. In both age groups, ERS for old associative memories was greater for trials for which new associative memories were forgotten than remembered. In contrast, ERS for new associative memories was greater when they were remembered than forgotten. In addition, older adults showed relatively attenuated target (i.e., new associates) and lure (i.e., old associates) ERS effects compared to younger adults. Collectively, these results suggest that the neural reinstatement of interfering memories during retrieval contributes to proactive interference across age, whereas overall attenuated ERS effect in older adults might reflect their reduced memory fidelity.

Effects of age on neural reinstatement during memory retrieval

“…These include electrophysiological methods such as electroencephalography (EEG) and event-related potentials (ERPs), when neural activity is measured directly, and indirect 'hemodynamic' methods such as functional magnetic resonance imaging (fMRI). Since very few published EEG or ERP studies have examined the effects of age on reinstatement (although see Lee et al, 2022), we focus here on fMRI. We focus on event-related fMRI, which permits neural activity (or more correctly, the blood oxygen level dependent [BOLD] signal, a hemodynamic correlate of neural activity) to be estimated at the single trial level.…”