www.nature.

com/npjscilearn

ARTICLE OPEN

The mitigating effect of repeated memory reactivations on

forgetting
Sydney MacLeod1, Michael G. Reynolds1 and Hugo Lehmann1

Memory reactivation is a process whereby cueing or recalling a long-term memory makes it enter a new active and labile state.
Substantial evidence suggests that during this state the memory can be updated (e.g., adding information) and can become more
vulnerable to disruption (e.g., brain insult). Memory reactivations can also prevent memory decay or forgetting. However, it is
unclear whether cueing recall of a feature or component of the memory can benefit retention similarly to promoting recall of the
entire memory. We examined this possibility by having participants view a series of neutral images and then randomly assigning
them to one of four reactivation groups: control (no reactivation), distractor (reactivation of experimental procedures), component
(image category reactivation), and descriptive (effortful description of the images). The experiment also included three retention
intervals: 1 h, 9 days, and 28 days. Importantly, the participants received three reactivations equally spaced within their respective
retention interval. At the end of the interval, all the participants were given an in-lab free-recall test in which they were asked to
write down each image they remembered with as many details as possible. The data revealed that both the participants in the
descriptive reactivation and component reactivation groups remembered significantly more than the participants in the control
groups, with the effect being most pronounced in the 28-day retention interval condition. These findings suggest that memory
reactivation, even component reactivation of a memory, makes memories more resistant to decay.
npj Science of Learning (2018)3:9 ; doi:10.1038/s41539-018-0025-x

INTRODUCTION modified by the presentation of novel information. Hupbach

Newly learned information goes through a consolidation period, et al.19 convincingly demonstrated the updating effect following
where it is labile and subject to modification, before stabilizing episodic memory reactivation in humans. Their participants
into a lasting memory.1,2 It was believed that, once consolidated, learned a list of objects, followed by a second list of objects
the memory became fairly permanent and resistant to change.3,4 48 h later. When the memory of List 1 was reactivated before
Now, however, it is well established that a consolidated memory learning List 2, their participants reported significantly more List 2
can re-enter a labile state induced by retrieval, a process called items intrusions in List 1 than their participants who did not
memory reactivation.5 During this new labile state, the memory receive a reactivation. Hence, when new information was
becomes susceptible to modification6–8 and the reactivation presented immediately following memory reactivation, informa-
generates a new bout of consolidation (i.e., reconsolidation).9,10 tion from the second list integrated with the original memory.
Amongst the first neuropsychological demonstrations that Memory updating is now considered to be a robust phenomenon
reactivated memories could be modified were studies examining and has been demonstrated to occur for several types of
the effects of electroconvulsive shock on fear memory.11–13 For memories, including episodic and procedural.19–22
instance, Misanin et al.12 showed that reactivation of a fear Reactivation and consequent reconsolidation processes may, in
memory immediately before electroconvulsive shock treatment addition to updating, benefit the original memory by increasing its
caused amnesia for that memory on future retention tests, strength,6,8,23 a possibility that is gaining substantial scientific
whereas the memory was unaffected if the electroconvulsive interest. Compelling evidence suggests that repeated reactiva-
shock treatment was given without prior reactivation. Studies tions increase the neural network that supports a memory, even
using protein synthesis inhibitors have demonstrated similar making it more resistant to brain insult.24–26 Other studies have
effects.14–17 Indeed, infusions of anisomycin immediately follow- demonstrated that reactivating a memory by brief re-exposure to
ing a memory reactivation in rats causes amnesic effects, but the the acquired information promotes new consolidation bouts that
same anisomycin treatment fails to impair the memory if not mitigate forgetting by increasing the persistence of the memory
preceded by a reactivation.17 Thus, reactivated memories can be and/or its precision.27–33
disrupted, even potentially erased, following a reactivation. The well-established cognitive psychology phenomenon known
Memories can also be updated following a reactivation, as the Testing Effect,34,35 in which repeated testing enhances
meaning that new information can be integrated into the original retention performance beyond new study episodes, should
memory during reconsolidation.18 In this instance, memories are likewise be taken as evidence that memory reactivations promote
not being erased or disrupted by exogenous manipulations such the strengthening of memory during reconsolidation.7 The
as electroconvulsive shock or pharmacological agents, but standard procedures used to demonstrate the Testing Effect

1
Psychology Department, Trent University, Peterborough, ON, Canada
Correspondence: Hugo Lehmann ([email protected])

Received: 13 October 2017 Revised: 27 March 2018 Accepted: 4 April 2018

Published in partnership with The University of Queensland

 The mitigating effect of repeated memory reactivations
S MacLeod et al.
2
involve effortful retrieval of the originally studied information and
can be accompanied by new studying of that original informa-
tion.34,35 This implies, for instance, that an individual providing a
description of the tuxedo they wore at their wedding would
benefit the retention of that memory. We often engage, however,
in partial or component recall of an event and it is unclear
whether this component reactivation elicits a reconsolidation
episode capable of strengthening the memory. For example,
would an individual mentioning the number of times they have
worn a tuxedo also improve their memory for the tuxedo they
wore at their wedding? Evidence pertaining to this question is Fig. 1 Schematic illustration of the experimental design. The
sparse and inconclusive. One study examining reconsolidation in participants briefly viewed a series of neutral images (Learning)
non-human animals suggests that reactivating a memory “linked” and were then assigned to one of three retention intervals: 1 h,
to the target memory is insufficient to promote reconsolidation 9 days, and 28 days. During these intervals, the participants received
processes in the target memory.36 Yet, in another study involving three equally spaced memory reactivations. However, the type of
reactivations was manipulated and either involved distractor,
human participants, recall of partial information from a studied component, or descriptive reactivations. A no reactivation Control
text increased memory for that text beyond the elements that group was also included in the design. At the end, the participants
were reactivated.37 Thus, the benefits of component or indirect were given an in-lab free-recall test
reactivations remain unresolved. The focus of this study was to
partially elucidate this issue by examining whether repeated of variance (ANOVA) revealed a significant main effect of interval,
component reactivations, and associated reconsolidation pro- F (2, 155) = 108.72, p < 0.001, a significant main effect of
cesses, could mitigate forgetting in humans similar to what is reactivation, F (3, 155) = 5.50, p = 0.001, but no significant
observed after effortful retrieval in the Testing Effect. interaction, F (6, 155) = 1.29, p = 0.263.
We assessed the above possibility by having participants The main effect of interval showed a typical time-dependent
reactivate viewed images from the International Affective Picture
1234567890():,;

forgetting curve. Specifically, least significant difference (LSD)

System (IAPS) using two different approaches: (1) descriptive
pairwise comparisons revealed that the participants in the 28-day
reactivations, which required the recall of as much information as
condition remembered significantly fewer images than the
possible about the viewed images (i.e., detailed written recall of
the list of studied images), and (2) component reactivations that participants in the 9-day condition (p < 0.01), and the 1-h
cued image categories without requiring detailed feature recall condition (p < 0.001). The participants in the 9-day condition also
(i.e., provide the number of images that included a category remembered significantly fewer images than the participants in
defining attribute). It is well established that repeated testing the 1-h condition (p < 0.001).
involving effortful retrieval of information improves retention.35 When examining the Reactivation main effect, it was found, as
Our descriptive reactivations are analogous to the effortful predicted, that the participants in the Descriptive group remem-
retrieval procedures in the Testing Effect studies, leading us to bered significantly more images than the participants in the
predict a reduction in forgetting. Our component reactivations, Distractor group (p < 0.001) and the no reactivation Control group
however, involved partial-reactivation processes that are different (p < 0.05), demonstrating mitigated forgetting. The component
than those required at retrieval during the retention test. reactivations also mitigated forgetting as the participants in this
Importantly, we predicted that this component reactivation, group remembered significantly more images than the partici-
despite the lack of required effortful descriptive retrieval, would pants in the Distractor group (p < 0.01). However, the Component
also lessen forgetting and demonstrate that reconsolidation group did not perform significantly better than the no reactivation
processes can increase memory strength. Moreover, we examined Control group (p = 0.25). Note that the Distractor group and the
the possible benefits of these reactivations over different no reactivation Control group did not significantly differ (p = 0.13).
retention intervals (1 h, 9 days, and 28 days) with the expectation
that the reactivation benefits would be the most apparent after Correct details
the longest retention interval involving the most forgetting. Two
control conditions were also included in the study: one condition Figure 3 shows the average number of correctly recalled details for
involved reactivation of procedures without reference to the each reactivation group across all three retention intervals. The
viewed images (Distractor group), whereas in the other the ANOVA revealed a significant main effect of Interval, F (2, 155) =
participants did not receive any reactivation manipulation (Control 89.07, p < 0.001, and a significant main effect of reactivation, F (3,
group). See Fig. 1 for a depiction of the experimental design. 155) = 5.24, p = 0.002, but no interaction, F (6, 155) = 0.64, p =
0.70.
Again, consistent with time-dependent forgetting, LSD pairwise
RESULTS comparisons revealed that the participants in the 28-day and 9-
The participants viewed a total of 26 neutral images and were day conditions remembered significantly fewer details than the
then given a free-recall test, either 1 h, 9 days, or 28 days later. participants in the 1-h condition (ps < 0.001). The performance of
During the retention interval, the participants were given three the participants in the 28-day condition, despite a tendency, did
equally spaced memory reactivations or no reactivation (Control). not significantly differ from those in the 9-day condition (p = 0.09).
The reactivations either involved retrieval of descriptive, compo- When examining the Reactivation main effect, it was found that
nent, or distractor information. The dependent variables used to the participants in the Descriptive group remembered significantly
assess memory performance included correctly recalled images, more details than the participants in the Distractor group (p <
correctly recalled details, and number of correctly recalled details 0.001) and the no reactivation Control group (p < 0.05), demon-
per image, as well as the number of falsely recalled items. strating mitigated forgetting. The participants in the Component
group remembered more details than the Distractor group (p <
Correct images 0.05), but did not significantly differ from the no reactivation
Figure 2 shows the number of correctly recalled images for each Control group (p = 0.31). The Descriptive group did not signifi-
reactivation group across all three retention intervals. The analysis cantly differ from the Component group (p = 0.19).

npj Science of Learning (2018) 9 Published in partnership with The University of Queensland
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
3

Fig. 2 Mean (±SEM) number of correctly remembered images during the free-recall test for each reactivation group across all three retention
intervals: 1 h, 9 days, and 28 days. Retention performance significantly decreased across retention intervals (ps < 0.05), suggesting time-
dependent forgetting. Yet, overall, the participants in the Descriptive and Component groups recalled significantly more images than the
participants in the Distractor group (ps < 0.05). More strikingly, in the 28-day condition, the participants in both the Descriptive and
Component groups recalled significantly more images than the participants in the Distractor and Control groups (ps < 0.05), with an relative
enhancement exceeding 59%. Hence, the descriptive and component reactivations successfully mitigated forgetting on this index of memory

Fig. 3 Mean (±SEM) number of correctly remembered details during the free-recall test for each reactivation group across all three retention
intervals: 1 h, 9 days, and 28 days. Retention performance significantly decreased across retention intervals (ps < 0.05), again suggesting
forgetting over time. However, overall, both the Descriptive and Component groups recalled significantly more details than the participants in
the Distractor group (ps < 0.05). At the longest retention interval, the participants in both the Descriptive and Component groups recalled
significantly more details than the participants in both the Distractor and Control groups (ps < 0.05) and the relative retention enhancement
exceeded 100%. Thus, the descriptive and component reactivations successfully mitigated forgetting on this memory measure as well

Correct details per image determine whether the reactivation method increased the quality
The total number of details recalled may be confounded with the of the retained mnemonic representation. The ANOVA revealed a
number of images recalled, therefore Fig. 4 shows the average significant main effect of interval, F (2, 155) = 13.97, p < 0.001, and
number of details recalled per correctly recalled image for each a significant main effect of reactivation, F (3, 155) = 2.94, p = 0.04,
group across all three retention intervals. This was conducted to but no interaction, F (6, 155) = 0.59, p = 0.74.

Published in partnership with The University of Queensland npj Science of Learning (2018) 9
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
4

Fig. 4 Mean (±SEM) number of correctly remembered details per recalled image during the free-recall test for each reactivation group across
all three retention intervals: 1 h, 9 days, and 28 days. This index is more revealing of the quality/precision of the memory for each recalled
image. In the 28-day condition, the participants in the Descriptive group recalled significantly more details per image than the participants in
the Distractor and Control groups (ps < 0.05), a finding that was not observed in the Component group. Hence, only the descriptive
reactivation improved precision retention

Consistent with time-dependent forgetting, LSD pairwise com-

parisons revealed that the participants in the 28-day and 9-day Table 1. Descriptive statistics for falsely recalled images and details
conditions remembered significantly fewer details per image than
the participants in the 1-h condition (ps < 0.001). The performance Group Condition Mean SD
of the participants in the 28-day condition did not significantly
differ from that of the participants in the 9-day condition (p = 0.82). False images
When examining the Reactivation main effect, it was found that Control 1-h 0.08 0.28
the Descriptive group remembered significantly more details per 9-day 2.75 4.67
image than the Distractor group (p < 0.01). There were no other 28-day 3.00 3.52
significant differences amongst the other groups (ps > 0.05).
Distractor 1-h 0.23 0.44
9-day 1.33 1.18
False recall
28-day 2.79 2.94
Table 1 provides the descriptive statistics for the number of falsely
Component 1-h 0.15 0.38
recalled images as well as the number of falsely recalled details.
These errors were analyzed to determine whether reactivations 9-day 1.93 1.49
caused intrusion or added noise to the mnemonic representations. 28-day 1.77 2.56
When examining the number of false images reported, the Descriptive 1-h 0.08 0.28
ANOVA revealed a significant main effect of Interval, F (2, 155) = 9-day 0.79 1.12
13.57, p < 0.001. The participants in the 1-h condition reported
28-day 2.5 3.27
significantly fewer false images than the participants in the 9-day
condition (p < 0.001) or the 28-day condition (p < 0.001). There False details
were no significant differences between the 9-day condition and Control 1-h 1.92 2.10
28-day condition (p = 0.07). Neither the main effect of reactiva- 9-day 1.94 1.80
tion, F (3, 155) = 0.40, p = 0.76, nor the interaction, F (6, 155) = 28-day 1.16 1.19
0.77, p = 0.59, approached significance. Distractor 1-h 1.00 1.47
When examining the number of falsely recalled details, the
9-day 1.40 1.50
ANOVA failed to reveal any significant effects (interval, F (2, 155) =
0.08, p = 0.93; reactivation, F (3, 155) = 0.40, p = 0.76; and 28-day 1.50 1.40
interaction, F (6, 155) = 0.82, p = 0.56). Component 1-h 1.46 1.81
9-day 1.4 2.06
28-Day condition analysis 28-day 1.31 1.11
It was predicted that the mitigating effect of reactivations would Descriptive 1-h 1.15 1.63
be the most pronounced at the longest time interval as this is 9-day 1.21 1.12
where participants would experience the highest likelihood of 28-day 2.00 1.79
forgetting. These data were, therefore, analyzed independently of

npj Science of Learning (2018) 9 Published in partnership with The University of Queensland
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
5
the other retention intervals using pairwise comparisons to results of our Descriptive group support this view, the component
examine group differences. Both the Descriptive and the reactivation findings, in which the reactivation process differs from
Component groups remembered significantly more images than that of the final test, do not. The component findings suggest that
the Distractor group (ps < 0.05) and the no reactivation Control promoting reactivation of the target memories, even in a different
group (ps < 0.05). Similarly, the Descriptive and Component retrieval form, is sufficient to promote long-term retention. Our
groups remembered significantly more correct details than the combined descriptive and component reactivation findings may
Distractor and no reactivation Control groups (ps < 0.05). These provide better support for theories such as the Elaborative
findings suggest that both types of reactivation mitigated Retrieval Hypothesis or the Construction-Integration Model, which
forgetting. The Descriptive and Component groups did not posits that retrieval, regardless of setting and features, causes
significantly differ from each other (ps > 0.05) nor did the widespread network activation that increases the number of
Distractor and no reactivation Control groups (ps > 0.05) on either traces supporting the memory and/or promotes more pathways to
of the two measures. In addition, the participants in the it.34,39,40 With this greater network, access to the target memories
Descriptive group remembered significantly more details per at the time of testing is facilitated and benefits retention
image than participants in the Distractor group (p < 0.05). There performance.39
was no effect of reactivation group on number of false images or Our findings also extend our knowledge about the Testing
false details recalled (ps > 0.05). Effect on other fronts. First, our demonstration that reactivations
To better depict the magnitude of the reactivation benefits on can improve memory by mitigating forgetting up to a month after
retention, we transformed the absolute retention score of each the initial learning goes beyond the typical retention intervals, of
participant in the Descriptive and Component groups into a hours to a few days, examined in most studies.35,41,42 Second,
relative score using the no reactivation Control group average learning only occurred once without the possibility of additional
(change relative to Control). This revealed that image recall was study in our experimental design, whereas other studies often
enhanced by 58.8% (SEM = 18.4) for the Descriptive group and intermix new study episodes with their reactivations.41,43,44 Third,
67.2% (SEM = 23.9) for the Component group. Similarly, for both the Testing Effect is typically demonstrated using word lists,45
groups the number of details recalled increased by 134.5% (SEM facts,46 word pairs,34,39 prose,47 and course material.41,43,44 Here,
= 27.34) and 113.1% (SEM = 43.4), respectively. The correct details we show that the benefits of reactivations on long-term retention
per image was also increased by 44.5% (SEM = 19.4) for the also applies to visual stimuli, such as pictures of scenes without
Descriptive group, but only by 16.6% (SEM = 17.6) for the the use of verbal information.48
Component group. All these increases were significantly above We believe that the reactivations in the current study, whether
what would be expected by chance (ps < 0.05), with the exception descriptive or component, strengthened long-term memory
of the Component group on the correct details per image (28 days) by initiating cellular/synaptic reconsolidation bouts.
measure (p > 0.05). Importantly, the Distractor group did not have When a memory is initially consolidated, it undergoes protein
any significant relative increases in performance on these synthesis resulting in neuronal changes that convert the memory
measures (ps > 0.05). Actually, all relative changes were negative to long-term storage.49,50 This process, called cellular consolida-
for this group. tion, takes place hours after an initial learning period and
strengthens the memory trace. Despite the original belief that
this was a one-time process,3,4 it has been convincingly
DISCUSSION demonstrated that each time a memory is reactivated it under-
The current study examined whether memory reactivations goes a new bout of consolidation.5 Though these reconsolidation
strengthen memory as indicated by mitigated forgetting. Impor- bouts involve plastic changes that differ from those of the initial
tantly, the effects of repeated memory reactivations on long-term consolidation,15,51 they contribute to memory maintenance.9,52 In
retention were assessed using two different approaches: descrip- contrast, the reactivation-induced memory enhancement within
tive and component reactivations. The descriptive reactivations the 1-h condition cannot be explained by the same plastic
required recall of as much information as possible about the changes because the interval was too short to benefit from
viewed images. The component reactivations, on the other hand, protein thesis. The reactivations, within this retention interval,
required partial recall of categorical features of the previously likely promoted short-term synaptic facilitation mechanisms
viewed images, with no detailed description. Both types of known to support working memory.53
reactivations mitigated forgetting. The benefits of the reactiva- The current reactivation findings have other important implica-
tions, however, were the most prominent at the 28-day interval tions for understanding the neural organization of long-term
when the memories were the weakest. In this instance, image memory. For instance, it has been strongly argued that memories
recall, relative to control performance, was increased by 59% and initially dependent on the hippocampus, such as episodic
67%, respectively, for the Descriptive and Component groups. The memories, become independent of the hippocampus over
relative effect was even greater for the number of recalled details time.54–56 Evidence supporting this argument comes from studies
with an increase exceeding 100% for both reactivation groups. reporting temporally graded retrograde amnesia in patients with
Thus, these findings confirm the view that reactivations hippocampal lesions, meaning that recently, but not remotely,
strengthen memories. acquired memories are lost following the damage.57,58 Distributed
The beneficial memory strengthening effects following the Reinstatement Theory, however, suggests that time may not be
descriptive reactivations are consistent with those of numerous the most critical factor involved in making memories independent
studies that have examined the Testing Effect, in which detailed of the hippocampus.59,60 With more remote memories, there are
effortful retrieval of previously studied information results in more opportunities for reactivations, and thus repeated bouts of
improved retention.34,35 Importantly, our component reactivation reconsolidation which strengthen the memory beyond the
findings add to this literature by demonstrating that cueing hippocampus. Indeed, it has been shown that repeated distrib-
category retrieval also benefits the memory. In addition, the uted reactivations can preserve a memory following hippocampal
component findings may elucidate some of the debated damage,26 but not in all situations.27 Moreover, evidence from
mechanisms supporting the Testing Effect. Transfer appropriate functional magnetic resonance imaging or functional studies in
processing has been proposed as one possible explanation for the human participants suggests that repeated retrievals (reactiva-
Testing Effect.38 This perspective suggests that effortful tions) of a memory increase neuronal activity in areas aside from
reactivation-induced retrievals must be similar in form to the final the hippocampus.24,25 Combined, the evidence from these studies
test to promote better retention performance. Although the suggest that reactivations extend the neural network supporting a

Published in partnership with The University of Queensland npj Science of Learning (2018) 9
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
6
memory, and arguably did the same for the image memories in of 167. This resulted in a group sample size ranging between 12 and 16,
the current study. which is consistent with that of other studies in the field.19,29
The complex features pertaining to the images we used
provided the advantage of an in-depth recall assessment (e.g., Materials
gist vs. detailed information). The descriptive and component Twenty-six neutral images were selected from the IAPS. Specifically, all
reactivations increased the number of images and details recalled, images ranged, according to IAPS, between four and six on the emotional
suggesting improved persistence/duration of the memory. The valence dimension (M = 5.4, SD = 0.45) and below four on the emotional
number of correct details recalled per image, however, is arguably arousal dimension (M = 3.3, SD = 0.42). Nine images depicted landscapes,
a better measure of the qualitative properties (precision) of the 3 were of animals, and the other 14 images were of household objects. The
image memories and only the descriptive reactivations improved images were presented on a 15-in. Dell LED monitor, using a Dell computer
performance on this index. Thus, the descriptive reactivations and Microsoft PowerPoint 2013.
improved long-term memory for the gist as well as the content of
the images. This is consistent with the evidence suggesting that Procedure
more effortful and detailed recall shows better Testing Effect Acquisition phase. The participants were told they would view a series of
benefits,41,44 as well as the enhanced mnemonic precision images, and were asked to remember as much about each image as
reported in studies that examined the effects of reconsolidation possible. The participants sat in front of, and approximately 60 cm away
in non-human animal studies.27,28 In contrast, the component from, the monitor, and when ready, pressed the “enter” key on the
reactivation strengthened the memory for the gist of the images. computer keyboard to begin the experiment. The 26 images were
We cannot know how much of the target memory was reactivated presented individually for 10 s, with a black screen shown between each
image for 3 s.
in the component condition. However, because only one
categorical feature was cued at the time of each reactivation, Reactivation manipulation. Each participant was randomly assigned to
we can assume that the recall was partial and lacked detailed one of four different groups: Descriptive Reactivation (Testing Effect
content. The finding that a memory can be strengthened without condition), Component Reactivation (Cueing Effect condition), Distractor,
detail-specific enhancement may also lend support to the memory and Control (no reactivation).
transformation view, which states that episodic memories become The Descriptive Reactivation group received three separate memory
more semantically represented, at least in part, because of reactivations by asking them to describe the images they had previously
repeated reactivation of gist-like information.61,62 viewed. In each instance, they were asked the same question, “Please
The reactivation benefits, whether from the Descriptive or describe, in as much detail as you recall, the images you viewed in Part 1
(acquisition phase) of this experiment.” The aim of this question was to
Component group, were most prominent when contrasted with
promote explicit effortful recall of detailed information of the images,
the Distractor group than the Control (no reactivation) group. The similar to the process that is followed in studies examining the Testing
former provided a control for experimenter–participant interac- Effect.47
tions during the reactivation sessions. Perhaps reactivating The Component Reactivation group also received three separate
irrelevant information of the experimental procedures added memory reactivations, but by cueing a component/categorical feature of
“noise” to the original image memories similar to memory the images. In each instance, the participants needed to provide the
updating, meaning the incorporation of new information into number of images that included the component. The first question
the target memory. But the absence of a statistical differences participants received was, “How many images included animals? Please do
between the Distractor and Control groups as well as the null not describe the images, only state the number within the category,”
question two, “How many images included landscapes? Please do not
findings on the false memory measures fail to support this describe the images, only state the number within the category” and
argument. Most importantly, the fact that the distractor reactiva- question three “How many images include household objects? Please do
tions did not strengthen the memory of the viewed images, but not describe the images, only state then number within the category.”
that the descriptive and component reactivations did, suggests Importantly, these reactivations did not require an explicit description of
that the reactivation content needs to pertain to the target the target memory.
memory in order to enhance retention. The Distractor group received three questions related to the experi-
In conclusion, we have replicated the Testing Effect through mental procedure which were not intended to trigger image reactivation.
descriptive reactivations and, more importantly, we have shown The objective of this condition was to provide an image-irrelevant
reactivation control, meaning controlling for questioning of the partici-
that component reactivations, in which only categorical features pants during the retention phase. The first question was “What was the
are used to induce reactivation, can similarly mitigate forgetting. date and time when you completed Part 1 of this experiment?” question
This suggests that an effortful description of the target memory is two, “What form did the experimenter go through with you before
not critical to enhance retention, at least gist-like information beginning the experiment?,” and question three “Please describe the
retention. In addition, the findings suggest that memory reactiva- procedure for Part 1 of this experiment.”
tion, whether descriptive or component, leads to memory The Control group did not receive any questions throughout the period
strengthening, a process commonly proposed within the memory from viewing the images to completing the recall test. The objective of this
reconsolidation literature. control condition was to examine the typical rate of forgetting.

Retention interval. In addition to the reactivation condition, the partici-

METHODS pants were randomly assigned to one of three retention interval
conditions: 1 h, 9 days, and 28 days.
All procedures were approved by Trent University’s Research Ethics Board,
1-h condition: In the 1-h condition, the acquisition-to-test interval was
which follows the guidelines of the Tri-Council Policy Statement of the
approximately 1 h. Specifically, after viewing the images, the participants
Government of Canada.
were given a 10-min break where they were provided reading material
from magazines such as Psychology Today, Scientific American Mind, and
Participants New Scientist. Following the break, the participants responded to a
A sample of 183 participants from Trent University (156 female) with a question on the computer using Qualtrics survey software. The question
mean age of 20.5 years (SD = 5.5) provide informed consent and differed according to the reactivation group to which the participant was
completed this experiment. The participants were recruited via SONA, an assigned: Descriptive, Component, or Distractor. This was repeated another
online research participation system, and were given credit towards a two times, again separated by a 10-min interval each time. After the third
psychology course as compensation. The data of 16 participants were question, there was a final 10-min break before the retention test. Note
excluded from the analyses due to issues with contacting the participants that the participants in the Control group were not questioned and read
for the reactivation manipulation or the testing, resulting in a sample size for 50 min without interruption until the test.

npj Science of Learning (2018) 9 Published in partnership with The University of Queensland
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
7
9-day condition: In the 9-day condition, the participants left the lab Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims
after viewing the images. Two days following the acquisition phase, the in published maps and institutional affiliations.
participants were sent a reactivation question via e-mail using the SONA
system. The question differed according to the reactivation group to which
the participant was assigned. The participants were given 12 h to respond REFERENCES
to the question. If they did not respond within this period their data were 1. Dudai, Y. The neurobiology of consolidations, or, how stable is the engram? Annu.
excluded from the study. This procedure was repeated on the fifth and Rev. Psychol. 55, 51–86 (2004).
seventh day. Thus, the participant received a total of three equally spaced 2. McGaugh, J. L. Memory—a century of consolidation. Science 287, 248–251 (2000).
reactivations. Responses to all questions were recorded using the Qualtrics 3. McGaugh, J. L. Time-dependent processes in memory storage. Science 153,
survey software. On the ninth day, the participants returned to the lab and 1351–1358 (1966).
completed a free-recall test. The participants in the Control group were not 4. Glickman, S. E. Perseverative neural processes and consolidation of the memory
contacted during this period, with the exception of an attendance trace. Psychol. Bull. 58, 218–233 (1961).
reminder e-mail sent the day prior to the free-recall test. 5. Sara, S. J. Retrieval and reconsolidation: toward a neurobiology of remembering.
29-day condition: The 28-day condition followed the same procedure Learn. Mem. 7, 73–84 (2000).
as the 9-day condition, with reactivation questions received on days 7, 14, 6. Alberini, C. M. The role of reconsolidation and the dynamic process of long-term
and 21. The participants returned to the lab on day 28 to complete the memory formation and storage. Front. Behav. Neurosci. 5, https://doi.org/10.3389/
free-recall test. The participants in the Control group were not contacted fnbeh.2011.00012 (2011).
during this period, with the exception of an attendance reminder e-mail 7. Dudai, Y. The restless engram: consolidations never end. Annu. Rev. Neurosci. 35,
sent the day prior to the free-recall test. 227–247 (2012).
8. Hardt, O., Einarsson, E. O. & Nader, K. A bridge over troubled water: reconsoli-
Retention test. Each participant completed a pen and paper free-recall dation as a link between cognitive and neuroscientific memory research tradi-
test while seated at the location used in the acquisition phase. They were tions. Annu. Rev. Psychol. 61, 141–167 (2009).
asked to recall as many images with as many details as possible. 9. Lee, J. L. Reconsolidation: maintaining memory relevance. Trends Neurosci. 32,
Specifically, the participants were asked “In part 1 of this experiment, you 413–420 (2009).
observed a series of images. In the space below, describe, in as much detail 10. Alberini, C. M., Milekic, M. H. & Tronel, S. Mechanisms of memory stabilization and
as possible, the images you recall.” The participants were given as much de-stabilization. Cell. Mol. Life. Sci. 63, 999–1008 (2006).
time as they required. Following completion of the test, the participants 11. DeVietti, T. L., Holliday, J. H. & Larson, R. C. Comparison of amnesias induced by
were debriefed on the purpose of the experiment. electroconvulsive shock administered after training-trial footshock or non-
contingent footshock in rats. J. Comp. Physiol. Psychol. 84, 579–585 (1973).
12. Misanin, J. R., Miller, R. R. & Lewis, D. J. Retrograde amnesia produced by elec-
Indices of memory and statistics
troconvulsive shock after reactivation of a consolidated memory trace. Science
The number of correctly recalled images, as well as the number of correct 160, 554–555 (1968).
details listed for each image from the free-recall test were used as indices 13. Devietti, T. L. & Hopfer, T. M. Complete amnesia induced by ECS and complete
of memory. An image was considered correctly recalled if it described the recovery of memory following reinstatement treatment. Physiol. Behav. 12,
gist of the image. A correct detail was tabulated using descriptors such as 599–603 (1974).
color, location, size, orientation, description, and items pertaining to the 14. Nader, K., Schafe, G. E. & Le Doux, J. E. Fear memories require protein synthesis in
image. The number of falsely recalled images and the number of incorrect the amygdala for reconsolidation after retrieval. Nature 406, 722–726 (2000).
details were also tabulated to assess intrusions. 15. Lee, J. L., Everitt, B. J. & Thomas, K. L. Independent cellular processes for hippo-
campal memory consolidation and reconsolidation. Science 304, 839–843 (2004).
Statistics 16. Milekic, M. H. & Alberini, C. M. Temporally graded requirement for protein
synthesis following memory reactivation. Neuron 36, 521–525 (2002).
The data were trimmed by removing the scores (on each dependent
17. Debiec, J., LeDoux, J. E. & Nader, K. Cellular and systems reconsolidation in the
measure) of the top two and bottom two participants on the correct image
hippocampus. Neuron 36, 527–538 (2002).
measure from each group. Trimming the tails of the distribution has been
18. Nadel, L., Hupbach, A., Gomez, R. & Newman-Smith, K. Memory formation, con-
shown to produce more robust findings.63
solidation and transformation. Neurosci. Biobehav. Rev. 36, 1640–1645 (2012).
The data were analyzed using a 3 × 4 between-subject ANOVA with
19. Hupbach, A., Gomez, R., Hardt, O. & Nadel, L. Reconsolidation of episodic mem-
interval conditions (1 h, 9 days, or 28 days) and reactivation groups
ories: a subtle reminder triggers integration of new information. Learn. Mem. 14,
(Descriptive, Component, Distractor, Control/no reactivation) as between-
47–53 (2007).
subject factors. LSD pairwise comparisons were also performed to examine
20. Schwabe, L. & Wolf, O. T. New episodic learning interferes with the reconsoli-
specific group differences that followed the postulated hypotheses. An α-
dation of autobiographical memories. PLoS ONE 4, e7519 (2009).
level of 0.05 was used in all instances.
21. Galluccio, L. & Rovee-Collier, C. Updating reactivated memories in infancy: II. Time
passage and repetition effects. Dev. Psychobiol. 47, 18–30 (2005).
Data availability 22. Walker, M. P., Brakefield, T., Hobson, J. A. & Stickgold, R. Dissociable stages of
All data collected for the current study are available, upon reasonable human memory consolidation and reconsolidation. Nature 425, 616–620 (2003).
request, from the corresponding author. 23. Sara, S. J. Strengthening the shaky trace through retrieval. Nat. Rev. Neurosci. 1,
212–213 (2000).
24. Eriksson, J., Kalpouzos, G. & Nyberg, L. Rewiring the brain with repeated retrieval:
a parametric fMRI study of the testing effect. Neurosci. Lett. 505, 36–40 (2011).
ACKNOWLEDGEMENTS 25. Nadel, L., Campbell, J. & Ryan, L. Autobiographical memory retrieval and hip-
We thank the Natural Sciences and Engineering Research Council of Canada (NSERC; pocampal activation as a function of repetition and the passage of time. Neural
355934 to H.L.; 341586 to M.G.R.; http://www.nserc-crsng.gc.ca) for funding this Plast. 2007, 1–15 (2007).
research. The funders had no role in study design, data collection and analysis, 26. Lehmann, H. & McNamara, K. C. Repeatedly reactivated memories become more
decision to publish, or preparation of the manuscript. resistant to hippocampal damage. Learn. Mem. 18, 132–135 (2011).
27. Alvares Lde, O. et al. Periodically reactivated context memory retains its precision
and dependence on the hippocampus. Hippocampus 22, 1092–1095 (2012).
AUTHOR CONTRIBUTIONS 28. De Oliveira Alvares, L. et al. Reactivation enables memory updating, precision-
keeping and strengthening: exploring the possible biological roles of reconsoli-
All authors contributed to the experimental design. The data were collected and
dation. Neuroscience 244, 42–48 (2013).
analyzed by S.M., but under the supervision of M.R. and H.L. The manuscript was
29. Forcato, C., Fernandez, R. S. & Pedreira, M. E. The role and dynamic of
written by S.M. and H.L., whereas M.R. provided major comments and revisions. All
strengthening in the reconsolidation process in a human declarative memory:
authors approved the final version of the manuscript for submission.
what decides the fate of recent and older memories? PLoS ONE 8, e61688 (2013).
30. Fukushima, H. et al. Enhancement of fear memory by retrieval through recon-
solidation. eLife 3, e02736 (2014).
ADDITIONAL INFORMATION 31. Haubrich, J. et al. Enhancement of extinction memory by pharmacological and
Competing interests: The authors declare no competing financial interests. behavioral interventions targeted to its reactivation. Sci. Rep. 7, 10960 (2017).

Published in partnership with The University of Queensland npj Science of Learning (2018) 9
 The mitigating effect of repeated memory reactivations
S MacLeod et al.
8
32. Inda, M. C., Muravieva, E. V. & Alberini, C. M. Memory retrieval and the passage of 52. Alberini, C. M. Mechanisms of memory stabilization: are consolidation and
time: from reconsolidation and strengthening to extinction. J. Neurosci. 31, reconsolidation similar or distinct processes? Trends Neurosci. 28, 51–56 (2005).
1635–1643 (2011). 53. Mongillo, G., Barak, O. & Tsodyks, M. Synaptic theory of working memory. Science
33. Lee, J. L. Memory reconsolidation mediates the strengthening of memories by 319, 1543–1546 (2008).
additional learning. Nat. Neurosci. 11, 1264–1266 (2008). 54. Squire, L. & Wixted, J. T. The cognitive neuroscience of human memory since H.M.
34. Karpicke, J. D. & Roediger, H. L. III The critical importance of retrieval for learning. Annu. Rev. Neurosci. 34, https://doi.org/10.1146/annurev-neuro-061010-113720
Science 319, 966–968 (2008). (2010).
35. Rowland, C. A. The effect of testing versus restudy on retention: a meta-analytic 55. Winocur, G., Moscovitch, M. & Bontempi, B. Memory formation and long-term
review of the testing effect. Psychol. Bull. 140, 1432–1463 (2014). retention in humans and animals: convergence towards a transformation
36. Debiec, J., Doyere, V., Nader, K. & Ledoux, J. E. Directly reactivated, but not account of hippocampal-neocortical interactions. Neuropsychologia 48,
indirectly reactivated, memories undergo reconsolidation in the amygdala. Proc. 2339–2356 (2010).
Natl. Acad. Sci. USA 103, 3428–3433 (2006). 56. Nadel, L. & Moscovitch, M. The hippocampal complex and long-term memory
37. Chan, J. C., McDermott, K. B. & Roediger, H. L. III Retrieval-induced facilitation: revisited. Trends Cogn. Sci. 5, 228–230 (2001).
initially nontested material can benefit from prior testing of related material. J. 57. Scoville, W. B. & Milner, B. Loss of recent memory after bilateral hippocampal
Exp. Psychol. Gen. 135, 553–571 (2006). lesions. J. Neurochem. 20, 11–21 (1957).
38. Morris, C. D., Bransford, J. D. & Franks, J. J. Levels of processing versus transfer 58. Rempel-Clower, N. L., Zola, S. M., Squire, L. R. & Amaral, D. G. Three cases of
appropriate processing. J. Verbal Learn. Verbal Behav. 16, 519–533 (1977). enduring memory impairment after bilateral damage limited to the hippocampal
39. Carpenter, S. K. Cue strength as a moderator of the testing effect: the benefits of formation. J. Neurosci. 16, 5233–5255 (1996).
elaborative retrieval. J. Exp. Psychol. Learn. Mem. Cogn. 35, 1563–1569 (2009). 59. Sutherland, R. J. & Lehmann, H. Alternative conceptions of memory consolidation
40. Kintsch, W. & Welsch, D. M. in Relating Theory and Data: Essays on Human Memory and the role of the hippocampus at the systems level in rodents. Curr. Opin.
in Honor of Bennet B. Murdock. (eds William, E. H. & Stephan, L.) 367–385 Neurobiol. 21, 446–451 (2011).
(Lawrence Erlbaum Associates, Hillsdale, NJ, 1991). 60. Sutherland, R. J., Sparks, F. T. & Lehmann, H. Hippocampus and retrograde
41. Butler, A. C. & Roediger, H. L. Testing improves long-term retention in a simulated amnesia in the rat model: a modest proposal for the situation of systems con-
classroom setting. Eur. J. Cogn. Psychol. 19, 514–527 (2007). solidation. Neuropsychologia 48, 2357–2369 (2010).
42. Larsen, D. P. et al. The effects of test-enhanced learning on long-term retention in 61. Moscovitch, M., Nadel, L., Winocur, G., Gilboa, A. & Rosenbaum, R. S. The cognitive
AAN annual meeting courses. Neurology 84, 748–754 (2015). neuroscience of remote episodic, semantic and spatial memory. Curr. Opin.
43. Larsen, D. P., Butler, A. C. & Roediger, H. L. 3rd Repeated testing improves long- Neurobiol. 16, 179–190 (2006).
term retention relative to repeated study: a randomised controlled trial. Med. 62. Nadel, L. & Moscovitch, M. Memory consolidation, retrograde amnesia and the
Educ. 43, 1174–1181 (2009). hippocampal complex. Curr. Opin. Neurobiol. 7, 217–227 (1997).
44. McDaniel, M. A., Roediger, H. L. 3rd & McDermott, K. B. Generalizing test- 63. Wilcox, R. R. The goals and strategies of robust methods. Br. J. Math. Stat. Psychol.
enhanced learning from the laboratory to the classroom. Psychon. Bull. Rev. 14, 51, 1–39 (1998).
200–206 (2007).
45. Carpenter, S. K. & DeLosh, E. L. Impoverished cue support enhances subsequent
retention: support for the elaborative retrieval explanation of the testing effect.
Open Access This article is licensed under a Creative Commons
Mem. Cogn. 34, 268–276 (2006).
Attribution 4.0 International License, which permits use, sharing,
46. Carpenter, S. K., Pashler, H., Wixted, J. T. & Vul, E. The effects of tests on learning
adaptation, distribution and reproduction in any medium or format, as long as you give
and forgetting. Mem. Cogn. 36, 438–448 (2008).
appropriate credit to the original author(s) and the source, provide a link to the Creative
47. Roediger, H. L. & Karpicke, J. D. Test-enhanced learning: taking memory tests
Commons license, and indicate if changes were made. The images or other third party
improves long-term retention. Psychol. Sci. 17, 249–255 (2006).
material in this article are included in the article’s Creative Commons license, unless
48. Wheeler, M. A. & Roediger, H. L. Disparate effects of repeated testing: Reconciling
indicated otherwise in a credit line to the material. If material is not included in the
Ballard’s (1913) and Bartlett’s (1932) results. Psychol. Sci. 3, 240–245 (1992).
article’s Creative Commons license and your intended use is not permitted by statutory
49. Roesler, R. Molecular mechanisms controlling protein synthesis in memory
regulation or exceeds the permitted use, you will need to obtain permission directly
reconsolidation. Neurobiol. Learn. Mem. 142, 30–40 (2017).
from the copyright holder. To view a copy of this license, visit http://creativecommons.
50. Davis, H. P. & Squire, L. R. Protein synthesis and memory: a review. Psychol. Bull.
org/licenses/by/4.0/.
96, 518–559 (1984).
51. Milekic, M. H., Pollonini, G. & Alberini, C. M. Temporal requirement of C/EBPbeta in
the amygdala following reactivation but not acquisition of inhibitory avoidance.
Learn. Mem. 14, 504–511 (2007). © The Author(s) 2018

npj Science of Learning (2018) 9 Published in partnership with The University of Queensland