Retrieval per se is not sufficient to trigger reconsolidation of human fear memory

https://doi.org/10.1016/j.nlm.2012.01.009Get rights and content

Abstract

Ample evidence suggests that consolidated memories, upon their retrieval, enter a labile state, in which they might be susceptible to change. It has been proposed that memory labilization allows for the integration of relevant information in the established memory trace (memory updating). Memory labilization and reconsolidation do not necessarily occur when a memory is being reactivated, but only when there is something to be learned during memory retrieval (prediction error). Thus, updating of a fear memory trace should not occur under retrieval conditions in which the outcome is fully predictable (no prediction error). Here, we addressed this issue, using a human differential fear conditioning procedure, by eliminating the very possibility of reinforcement of the reminder cue. A previously established fear memory (picture-shock pairings) was reactivated with shock-electrodes attached (Propranolol group, n = 18) or unattached (Propranolol No-Shock Expectation group, n = 19). We additionally tested a placebo-control group with the shock-electrodes attached (Placebo group, n = 18). Reconsolidation was not triggered when nothing could be learned during the reminder trial, as noradrenergic blockade did not affect expression of the fear memory 24 h later in the Propranolol No-Shock Expectation group. Only when the outcome of the retrieval cue was not fully predictable, propranolol, contrary to placebo, reduced the startle fear response and prevented the return of fear (reinstatement) the following day. In line with previous studies, skin conductance response and shock expectancies were not affected by propranolol. Remarkably, a double dissociation emerged between the emotional (startle response) and more cognitive expression (expectancies, SCR) of the fear memory. Our findings have important implications for reconsolidation blockade as treatment strategy for emotional disorders. First, fear reducing procedures that target the emotional component of fear memory do not necessarily affect the cognitive component and vice versa. Second, mere retrieval of the fear memory is not sufficient to induce its labilization and reconsolidation.

Highlights

► Updating of human fear memory depends on prediction error. ► Propranolol impaired the reconsolidation of the affective expression of fear memory. ► A double dissociation emerged between the affective and cognitive fear memory expression.

Introduction

A dynamic balance between stability and plasticity of memory seems to be crucial for adaptation to an ever-changing environment. Stability of fear memory guarantees a fast response to threat and does not require continuous re-learning, whereas plasticity permits modification of an established memory trace, should conditions require such adaptation (Dudai, 2009). Plasticity of fear memory is also of clinical importance, as it provides a window of opportunity to target unwanted, excessive emotional memories such as those that underlie anxiety disorders (e.g., PTSD). While consolidation, the strengthening of a memory trace over time, serves the stability of memory (McGaugh, 1966), the process of reconsolidation, the protein-synthesis dependent restabilization of memory upon retrieval, allows for memory modification (Nader, Schafe, & LeDoux, 2000). Reconsolidation is typically demonstrated by the amnesic effects of protein synthesis inhibitors administered before or after memory reactivation (Nader et al., 2000; Pedreira et al., 2002, Sara, 2000). Recently, we demonstrated that administration of the β-adrenergic receptor antagonist propranolol before memory reactivation disrupted the emotional expression of fear memory (startle fear response) and prevented the return of fear 24 h later, while declarative memory remained intact (Kindt et al., 2009, Soeter and Kindt, 2010, Soeter and Kindt, 2011a, Soeter and Kindt, 2011b).

If retrieval per se were sufficient to render a memory trace labile, memory would be hypermalleable. Indeed, it has been suggested that reconsolidation does not necessarily occur when a memory is being reactivated, but only when something can be learned during memory retrieval (memory updating) (Forcato et al., 2009, Forcato et al., 2010, Lee, 2009, Pedreira et al., 2004). A violation of expectation based on prior learning – the magnitude of the outcome or the outcome itself is not fully predicted (prediction error) – is thus a prerequisite for reconsolidation to take place (Forcato et al., 2009, Forcato et al., 2010, Lee, 2009, Pedreira et al., 2004). We hypothesize that when nothing can be learned during memory reactivation (no prediction error), there is no need for the memory to be updated and reconsolidation will not be triggered.

Here we examined in a differential fear conditioning study in humans whether reconsolidation of a previously established fear memory depends on the outcome of the reactivation trial. We applied a reactivation procedure similar to our previous studies (i.e., unreinforced presentation of the feared stimulus) (Kindt et al., 2009, Soeter and Kindt, 2010), but discarded the reactivation trial of prediction error by excluding the very possibility of reinforcement. Briefly, on day 1, fear acquisition was established with use of two spider pictures (Conditioned Stimuli; CS). One of the pictures (CS1+) was followed by an aversive electrical stimulation (Unconditioned Stimulus; US), while the other was not (CS2). On day 2, participants received propranolol before the fear memory was reactivated with a non-reinforced reminder presentation of the feared stimulus (CS1-R). During memory retrieval, shock-electrodes were attached in one group (Propranolol group), but not in the other (Propranolol No-Shock Expectation). As such, presentation of the CS1 was supposed to elicit active anticipation of the US in the former group but not in the latter. To test whether the observed effect of the manipulation of shock expectation of the reminder trial was indeed related to a propranolol-induced disruption of reconsolidation, an additional group of participants received a placebo pill before memory reactivation. Similar to the Propranolol group, the shock-electrodes were attached during the reminder presentation in the Placebo group. In all groups, expression of the fear memory was tested 24 h later (day 3). We hypothesize that a reactivation session that is devoid of prediction error (i.e., prior knowledge that the presentation of the feared stimulus (CS1-R) cannot be followed by shock) will fail to trigger reconsolidation. Noradrenergic blockade will disrupt reconsolidation of a previously acquired fear memory when learning can occur during reactivation. Therefore, we predict that in contrast to both the Propranolol No-SE and the Placebo group, administration of propranolol before memory reactivation will attenuate subsequent expression of the fear memory (startle fear response) in the Propranolol group. In line with our previous studies, declarative memory will remain unaffected (US-expectancies). Given the close association of the skin conductance response (SCR) with declarative knowledge (Hamm & Weike, 2005), noradrenergic blockade will not affect electrodermal responding either.

Section snippets

Participants

Sixty (19 male; 41 female) healthy undergraduate students participated in the study, ranging in age between 18 and 30 years, with a mean age of 21.08 years (SD = 2.61). All participants were free from any condition contraindicative to the administration of 40 mg propranolol (see Soeter & Kindt, 2010). Participants received either partial course credit or a small amount of money (€ 35, –) for their participation. All participants gave informed consent and were notified that they could withdraw from

Questionnaires, evaluations and blood pressure

The Propranolol group did not differ in reported spider fear, anxiety sensitivity or trait anxiety from the Propranolol No-SE group and the Placebo group (Fs < 1.05). The individually set shock intensity ranged from 4 to 55 mA (M = 16.5, SD = 10.08). Comparing the groups with respect to US-intensity revealed higher shock intensities for the Propranolol group (M = 18.83, SD = 8.7) compared to the Propranolol No-SE group (M = 12.53, SD = 6.95) (F(1, 36) = 5.97, p < .05) but not compared to the Placebo group (M = 

Discussion

The present findings show that reconsolidation of a human fear memory is not triggered when nothing can be learned from the memory retrieval session. That is, administration of the β-adrenergic receptor antagonist propranolol prior to reactivation did not affect the startle fear response 24 h later, when the outcome of the reminder trial was perfectly predictable (no prediction error). Noradrenergic blockade did reduce the startle fear response on day 3, when there was something to be learned

Acknowledgments

The authors thank Bert Molenkamp for his technical assistance. This study was funded by a Vici grant (Merel Kindt) from the Netherlands Organization for Scientific Research.

References (36)

  • S. Duvarci et al.

    Characterization of fear memory reconsolidation

    Journal of Neuroscience

    (2004)
  • A.G. Gilman et al.

    Goodman and Gilman’s the pharmacological basis of therapeutics

    (1996)
  • C.L. Hull

    Principles of behavior

    (1943)
  • M. Kindt et al.

    Beyond extinction: Erasing human fear responses and preventing the return of fear

    Nature Neuroscience

    (2009)
  • P.J. Lang et al.

    International affective picture system (IAPS): Affective ratings of pictures and instruction manual

    (2005)
  • J.L. Lee

    Memory reconsolidation mediates the strengthening of memories by additional learning

    Nature Neuroscience

    (2008)
  • P.F. Lovibond et al.

    Electrodermal and subjective reactions to fear-relevant stimuli under threat of shock

    Australian Journal of Psychology

    (1994)
  • J.L. McGaugh

    Time-dependent processes in memory storage

    Science

    (1966)
  • Cited by (0)

    View full text