Modification of episodic memories by novel learning: a failed replication study

Background: After reactivation, memories can become unstable and sensitive to modification before they are restored into long-term memory. Using behavioural manipulations, reactivated memories may be disrupted via the mechanism of interference (i.e. novel learning). In a laboratory study, Wichert et al. (2013a) showed that new learning after reactivation changed episodic memory, while new learning alone or reactivation alone did not. Objective: Given the potential clinical application of such a procedure in trauma-focused psychological treatments, such as CBT or EMDR, the aim of this study was to replicate Wichert et al. Method: On Day 1, participants (N = 96) viewed and recalled a series of emotional and non-emotional pictures. Then, participants were randomized to one of four groups. One week later, on Day 8, Group 1 reactivated the previously learned pictures and learned new pictures. To control for specific effects of reactivation or new learning, Group 2 only reactivated the previously learned pictures, and Group 3 only learned new pictures. Group 4 received no reactivation and no new learning. On Day 9, all groups indicated for each picture out of a series whether they had seen it on Day 1. Results: The data were analysed using Bayesian hypothesis testing, which allows for quantifying the evidence in favour of the alternative and the null hypothesis. In general, results showed that Group 1 recognized fewer pictures from Day 1 compared to Groups 2 and 4 on Day 9. However, the expected difference between new learning following reactivation (i.e. Group 1) and new learning alone (i.e. Group 3) was not substantially supported by the data for any of our dependent measures. Conclusions: We replicated some of the findings by Wichert et al., but did not find substantial support for the critical difference between new learning following reactivation and new learning alone.