Positive and neutral updating reconsolidate aversive episodic memories via different routes

Abstract

Aversive memories are long-lasting and prone to burden our emotional wellbeing and mental health. Yet, how to remedy the maladaptive effects of aversive memories remains elusive. Using memory reactivation and emotional updating manipulations, we investigated how positive and neutral emotion may update aversive memories for reconsolidation in humans. We found that positive updating after reactivation was equivalent to neutral updating in impairing true memories of a previous aversive event after a 12-hour wakeful delay, but induced more false memory. Moreover, additional 12-hour delay with overnight sleep did not further enlarge true memory differences, but attenuated the effect of reactivation and updating on false memory. Interestingly, the neutral rather than the positive updating reduced the emotional arousal of the aversive memory 24 h later. Our findings could serve as a reference for real-world therapeutic applications regarding how positive and neutral updating may reshape aversive memories, especially when taking wake- and sleep-filled reconsolidation into account.

Introduction

Emotional episodic memory constitutes our individual lives. While joyful moments sparkle our memory bank with happiness, there are also incidents of aversive and distressing events, even traumatic. Once the aversive memories become maladaptive, they could burden our emotional wellbeing and mental health. For example, the sudden outburst of the global COVID-19 pandemic has infected over 200 million people and led to more than four million deaths till the publication of this work. The part-for-ever caused by this pandemic has become traumatic memories of millions of people worldwide (Rajkumar, 2020). Scientists have warned that overloaded traumatic episodic memory could lead to severe mental disorders like post-traumatic stress disorder (PTSD) and major depression even in the post-pandemic time (Carmassi et al., 2020, Kathirvel, 2020, Xiao et al., 2020). Appropriate and efficient interventions for maladaptive memory to prevent the development of severe mental illnesses are in need. There are pieces of evidence showing pharmacological manipulations like adrenergic receptor antagonist (Lonergan et al., 2013, Soeter and Kindt, 2015), and electroconvulsive therapy (ECT) might interference with aversive memories (Kroes et al., 2014, Misanin et al., 1968). However, these methods are limited for applications due to their invasive nature. Hence, patient-friendly and non-invasive approaches, like behavioral modifications, are under the urgent need for real-world applications.

Convincing evidence has proved that after memory retrieval, the consolidated memories can be rendered into a labile state; a reconsolidation process is then involved to stabilize it into the brain again (Nader & Hardt, 2009). Once a memory is destabilized, it becomes susceptible to new information, leading to the existence of a time window for memory updating (Phelps & Hofmann, 2019). The discovery of memory malleability has generated broad interest to remedy symptoms linking to aversive memories in clinical populations (i.e., PTSD). Emerging evidence has suggested that introducing new learning shortly after memory reactivation can incorporate new information into the already labialized original memory, a process known as memory updating (Lee, Nader, & Schiller, 2017). Among conventional memory updating approaches, counter-conditioning that involves replacing an expected salient outcome with a new outcome of the opposite valence (Keller, Hennings, & Dunsmoor, 2020) has been proved with promising outcomes for appetitive-to-aversive applications. For instance, introducing aversive feelings shortly after memory reactivation had significantly reduced addictive behavior for food (Olshavsky et al., 2013), alcohol (Das, Lawn, & Kamboj, 2015), and cocaine (Goltseker, Bolotin, & Barak, 2017). In contrast, aversive-to-appetitive counter-conditioning studies reported inconsistent findings. On the one hand, utilizing behavioral and optogenetic techniques, researchers had successfully remedied depressive-like or aversive behavior in rodents, by artificially triggering positive memory engrams in the hippocampus during the reactivation of negative experiences (Ramirez et al., 2015, Redondo et al., 2014). On the other hand, findings from the aversive-to-appetitive counter-conditioning paradigm in humans were inconsistent. While many studies have shown a better effect for aversive-to-positive updating than aversive-to-neutral updating or extinction (Eifert et al., 1988, Newall et al., 2017, Reynolds et al., 2018), there are also studies showing that counter-conditioning is not better than the traditional extinction protocol in some aspects (de Jong et al., 2000, Meulders et al., 2015). The counter-conditioning could even prone to renewal the negative memory (Holmes, Leung, & Westbrook, 2016). Hence, although with promising real-world application potentials, the experimental boundary of counter-conditioning updating remains to be explored.

Episodic memory for an emotional event consists of multiple aspects of composited contents, including episodic details on what, where, and when such event occurred (Tulving, 1993), as well as the emotional significance of how we feel about that experience (Christianson, 1992, Dolan et al., 2000, Liu et al., 2016). The recall of episodic memory includes the information that one truly experienced (i.e., true memory), and some fictitious yet plausible information that did not experience (i.e., distorted or false memory) (Guarnieri, Bueno, & Tudesco, 2019; Loftus, 1979). The fuzzy-trace theory (FTT) had proposed that the specific and detailed traits of the episodes are stored in a literal memory system, the recalling of which would generate true memory. In contrast, the gist and central theme of the episodes are stored in the essence memory system, generating false memory (Brainerd & Reyna, 2002, 2005). Besides the memory contents, the subjective emotional feeling is a vital component of emotional episodic memory, and it is also a crucial indicator to evaluate therapeutic modifications (Lane, Ryan, Nadel, & Greenberg, 2015). Ample studies have successfully reduced the individual's negative feelings (e.g., fear) through memory updating operations (Sandkühler and Lee, 2013, Soeter and Kindt, 2015). However, to our knowledge, there is no study yet to investigate how updating with positive emotion reshapes previously acquired aversive memories with either impaired or distorted outcomes, and even less is known how this procedure alters the emotional arousal of original memory representation.

Memory reconsolidation involves protein synthesis processes, requiring a period of hours, days or even longer to complete (Nader, Schafe, & LeDoux, 2000). Some studies suggested that the reconsolidation window lasts at least six hours during wakefulness (Björkstrand et al., 2016). Meanwhile, it is also reported that sleep supports memory reconsolidation (Klinzing et al., 2016, Walker et al., 2003), probably by speeding up or shortening the reconsolidation window (Moyano, Diekelmann, Pedreira, & Forcato, 2019). This indicates that both sleep and wakefulness play an essential role in memory reconsolidation. In the context of emotional episodic memory, sleep is crucial to consolidate newly encoded episodic events for both true (Weber, Wang, Born, & Inostroza, 2014) and false memories (Payne et al., 2009). Importantly, when combining the modulation of emotion, sleep seems to consolidate these two kinds of memory in different ways (McKeon, Pace-Schott, & Spencer, 2012). Reconsolidation works as a follow-up step after memory retrieval rather than encoding, and it naturally involves both wake and sleep. Hence, it is critical to understand how sleep, after the wake reconsolidation, supports reconsolidation for both the true and false memory of episodic events that have been emotionally modified. To our knowledge, this has not been well studied.

Here we attempted to bridge the gaps mentioned above by raising the following questions: First, whether new information introduced shortly after reactivation can alter existing aversive memory under the reconsolidation procedure? Second, whether introducing interference with positive emotion can impair previous negative memory better than updating with neutral one? And third, how the effects of positive and neutral updating on aversive memory evolve over 12-hour wakefulness and 24-hour interval with a night of sleep? We set up a between-subject factorial experiment to address these critical questions, including the independent variables of reactivation and positive/neutral updating procedures. Participants encoded an aversive story presented as picture slides with auditory narratives on the Day 1 morning, with an immediate recall test as the baseline assessment of memory (Fig. 1). Seven days later (on the morning of Day 8), participants returned to the lab in morning and were assigned into four experimental groups with either memory reactivation combining with positive or neutral updating or just updating manipulations (Group A: ReaPos, B: ReaNeu, C: NonReaPos, and D: NonReaNeu; see Methods). For the reactivation, prediction errors (or surprise) are thought to initiate destabilization of memory traces (Exton-McGuinness et al., 2015, Sinclair and Barense, 2018). Hence we instructed participants to perform a memory retrieval session, but surprisingly interrupted the retrieval after the first cue slide was presented (see Methods), a way that had been used for successful memory destabilization (Forcato et al., 2007). All participants were tested 12 h later in the evening of the same day for their memory of the original aversive story. Finally, after a night of sleep (on the morning of Day 9), all participants returned to the lab in the morning for a final test on memory, emotional valence, and arousal of the negative story.

With this design, we hypothesized that (1) an impairment of memory for the original negative (or aversive) story on Day-8 only appears in the groups that received both memory reactivation and updating; (2) Since the strength of the updating material differs the degree of impairment for the old memory, i.e., stronger new learning after reactivation leads to better memory updating, we would expect that the group with positive emotion updating should impair the negative memory more than the neutral group. Moreover, (3) a recent study in starlings showed that sleep promoted the reconsolidation of old memory once it has been reactivated, and this effect was notably more substantial when new interferences existed after the reactivation (Brawn, Nusbaum, & Margoliash, 2018). Hence, we hypothesize that after 24-hour interval involving sleep could favor the recovery of the negative memory, especially in the positive group (as a stronger updating manipulation than the neutral updating) with memory reactivation. Apart from memory impairment, an early study suggested that after reactivation, the destabilized memory turns to be more susceptible to interference, intermixing with new information (Hupbach, Gomez, Hardt, & Nadel, 2007). Thus, we further hypothesized that (4) comparing with non-reactivation groups, memory performance from reactivation groups should have more false memory after the updating stories. Finally, we hypothesized that the emotional components of the updating material should also be integrated into the negative memory after reactivation, leading to a significant reduction of the negative feeling of the aversive memory for subjects who received a positive updating.