Trends in Neuroscience and Education 4 (2015) 49–59

Contents lists available at ScienceDirect

Trends in Neuroscience and Education

journal homepage: www.elsevier.com/locate/tine

Review article

The effect of distributed practice: Neuroscience, cognition,

and education
Emilie Gerbier a,n, Thomas C. Toppino b,1
a
Laboratoire d'Etudes des Mécanismes Cognitifs – EA 3082, Université de Lyon , Université Lumière Lyon 2, 5 Avenue Pierre Mendès France,
69500 Bron, France
b
Villanova University, 800 Lancaster Avenue, Villanova, PA 19087, USA

art ic l e i nf o a b s t r a c t

Article history: Education ideally should induce learning that lasts for years and more. A wealth of research indicates
Received 26 July 2014 that, to achieve long-lasting retention, information must be practiced and/or tested repeatedly, with
Received in revised form repeated practice well distributed over time. In this paper we discuss the behavioral, neuroimaging and
24 November 2014
neurophysiological findings related to the effect of distributed practice and testing as well as the
Accepted 4 January 2015
resulting theoretical accounts. Distributed practice and testing appear to be powerful learning tools. We
Available online 14 February 2015
consider implications of these learning principles for educational practice.
Keywords: & 2015 Elsevier GmbH. All rights reserved.
Distributed practice
Spacing effect
Testing effect
Long-term retention
Consolidation
Reconsolidation

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2. Massed versus spaced practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.1. Deficient processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2. Spacing and the cellular bases of learning and memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.3. The effect of longer inter-study intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3. More on the effect of spacing: complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4. The role of retrieval or reactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.1. Study-phase retrieval. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.2. Testing effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.3. Reconsolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5. Implications for education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.1. The promise and limits of educational application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2. Learning for the long term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.3. Implementation of distributed practice in the classroom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.4. Practical pointers for learners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

1. Introduction

Say you need to learn a lesson about the French Revolution for
n
your history class and you only have a limited time to prepare for
Corresponding author. Tel.: þ 334 76 82 71 45.
E-mail addresses: [email protected] (E. Gerbier),
an exam that will take place in one week. What is the best strategy
[email protected] (T.C. Toppino). for scheduling your study sessions during the next seven days?
1
Tel.: þ1 610 519 4722. Should you wait until the day before the exam to study? Should

http://dx.doi.org/10.1016/j.tine.2015.01.001
2211-9493/& 2015 Elsevier GmbH. All rights reserved.
50 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59

you start today and study a different part of your lesson each day? problems [11]. And, very different learning tasks appear to yield
Or, should you organize your studying so that every part of the similar effects across an extended range of spacings [12]. (For
lesson is studied and reviewed in two or three sessions, separated reviews of the distributed practice effect, see [2,13–15]).
by one to several days? It is noteworthy that the effect is valid with tasks that are
Obviously, one way to boost learning and memory is to increase experienced in real school settings. For example, Bloom and Shuell
the number of study opportunities with to-be-learned materials. [16] had high-school students learn foreign language vocabulary
Unfortunately, learning time is necessarily limited because of the either in a single, 30-min long session or with three 10-min sessions
enormous amount of information to be acquired by students in the over 3 days. Retention after four days was 35% better in the spaced
different disciplines, both in and outside school. In addition, we than in the massed condition. The benefit of spaced practice seems
generally want knowledge to be as durable as possible. In this to be even more powerful when multiple repetitions are used. Work
article, we will discuss how to arrange study time efficiently by from Bahrick and colleagues showed that the more the study
scheduling study sessions in a way that maximizes learning and sessions were spaced, the better the result (e.g., [17,18]), although
memory. As we will develop it further in this article, it is most there are limits to this generalization as we will see later.
beneficial to engage in “distributed practice.” That is, it is best to Although most of the research on distributed practice has been
study something by scheduling relatively short study sessions that conducted with young, college-age adults, the benefit of distrib-
can be repeated after an appropriate period of time rather than by uted practice has been observed throughout the lifespan. Children
devoting the same total amount of time to a single study session or display this effect. It has been demonstrated in infancy (e.g.,
to a number of repeated study sessions that occur in immediate 5 months in [19]), in preschool children (e.g., [20]), as well as in
succession. The distribution of practice is under-used in real life elementary-school (e.g., [21]), and middle-school (e.g., [22]) chil-
settings [1], perhaps because it is counter-intuitive [2]. However, it dren. At the other extreme, distributed practice effects have been
is easy to implement and potentially useful in a large number of demonstrated in the elderly (e.g., [23,24]).
contexts and disciplines (see [3]). It is important for the educational The generality of the spacing effect also extends to other species.
community to become more aware of the benefits of distributed Very basic organisms, from drosophila [25], aplysia [26] and bees
practice. We will focus on the cognitive and neuroscientific research [27], to vertebrates like rodents [28] display the spacing effect in
on distributed practice effects, and conclude with implications for simple learning paradigms like conditioning, habituation or sensi-
education. tization. For instance, when drosophila learned to avoid a given
odor that was associated with electric shocks through multiple
exposures, they retained the avoiding behavior over a longer period
2. Massed versus spaced practice when they experienced 15-min spaced exposures to this association
rather than massed exposures [25].
Re-studying a piece of information immediately after the first The uncommon ubiquity of the spacing effect suggests that it
study episode is not an efficient way to proceed in order to learn reflects a fundamental principle of the memory system, shaped by
effectively and retain information over a long period of time. More evolution (e.g., [29]). It calls for an explanation in terms of a very
than 100 years of psychological research have consistently demon- general mechanism or set of mechanisms.
strated that spacing the repetitions of the same piece of informa-
tion over time favors later retention of the material compared to 2.1. Deficient processing
massing the repetitions in immediate succession. Classic experi-
ments typically involve either of two basic designs. Researchers One general mechanism that may contribute to the spacing effect
may present to-be-learned items (e.g., words) one by one in a is described by the deficient processing hypothesis. According to this
single session and vary how many other items are interspersed view, the second occurrence of an immediately repeated item
between two presentations of a repeated item Alternatively, they receives less processing compared to a spaced repetition, resulting
may present the same items in two different sessions and vary the in less efficient encoding and poorer memory. Supporting evidence
interval between sessions (see Fig. 1 for a schematic view of such comes from both psychological and brain imaging studies, some of
studies). Two spaced presentations can be twice as effective as two which are reviewed below [for a more complete review, see [2]].
massed presentations. Research has also shown that this effect is Magliero [30] found that pupil dilation, an indicator of processing
valid for virtually all types of to-be-learned materials: words (e.g., effort, is smaller for the second presentation of a massed item
[4]), word pairs (e.g., [5]), faces (e.g., [6]), pictures (e.g., [7]), texts compared to the second presentation of a spaced item. Johnston and
(e.g., [8]), and so on. The phenomenon also has been observed for Uhl [31] showed that reaction times to a tone are shorter during
the practice of motor skills (e.g., [9]) as well as for cognitive skills repeated occurrences of an item when the repetitions are massed
like grammar in a foreign language [10] or solving mathematics rather than spaced, indicating that less attentional processing is
devoted to the massed repetitions. Deficient processing may be
partially influenced by voluntary strategies used by the learner. For
example, a sense of familiarity may arise when the same informa-
tion is re-presented immediately, creating the impression that the
information is already learned and that no further processing is
necessary. Indeed, when participants could choose how long to
study each item, they spent less time on the second presentation of
a massed item than a spaced item [32]. However, involuntary
processes also seem to play an important role. The spacing effect
is undiminished even when incentives are provided to motivate
people to study massed repetitions to a greater extent [33]. It also is
Fig. 1. Schematic representation of experimental studies of the effect distributed obtained when learning occurs incidentally (unintentionally) (e.g.,
practice. The upper panel represents studies manipulating spacing within a [34]), and when learners are very young children who typically do
learning session, and the dark arrows highlight the spacing interval separating
some of the repeated items. The lower panel describes studies using two distinct
not engage in strategic study behavior [20].
learning sessions involving the same material. The dark arrow represents the Several researchers (e.g., [6,35]) have proposed that deficient
interval between sessions which can be manipulated to vary the degree of spacing. processing may be related to a priming mechanism. Short-term
 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59 51

behavioral priming refers to the fact that an item is processed twice incidentally under the guise of performing a concrete/abstract
more quickly and more easily when a previous presentation has judgment task. Repetitions were separated by a 2-min or a 25-h lag,
occurred than when it is first presented. This priming effect, and a surprise recognition test was administered 48 h after the
however, dissipates as the interval between the first and second second presentation of repeated items. During the judgment task,
presentation increases. One intuitive interpretation is that, when repeated words were categorized more quickly during their second
the interval between repetitions is short, the item's representation presentation than during the first (i.e., behavioral priming) and
is already pre-activated in memory at the time of its second more quickly after the 2-min lag than after the 25-h lag. Mirroring
occurrence, and, therefore, further processing is not needed (or this behavioral priming effect, fMRI results revealed lower activa-
is needed to a lesser extent) in order to activate it. Less processing, tion in the left inferior prefrontal and left fusiform cortices during
however, leads to poorer memory of the item later on. As priming the second occurrence of repeated items compared to their first
dissipates with an increase in the interval between presentations, presentation. In addition, this neural suppression was larger after
the second presentation should require more processing, leading the 2-min lag than after the 25-h lag, and, as expected, recognition
to improved memory. The original findings supporting a priming performance was more accurate in the 25-h condition than in the
explanation of the spacing effect were reported in psychological 2-min condition. Finally, negative correlations were obtained
studies (e.g., [6,35]), but more recent findings consistent with this between the magnitude of behavioral priming (and neural suppres-
view have been reported in neuroscientific studies, primarily using sion) and the final recognition scores of the words; in other words,
imaging techniques. the greater the behavioral priming and/or neural suppression, the
Functional Magnetic Resonance Imagery (fMRI) capitalizes on the poorer the final recognition. However, correlations do not imply
fact that blood flow is greater in active regions of the brain. The causality, and the possibility cannot be ruled out that the effect of
technique reveals the degree of activation in various neural struc- lag on priming and the effect of lag on subsequent memory act in
tures by measuring the magnetic properties of oxygenated and parallel without a causal effect of the former on the latter (for
deoxygenated hemoglobin. In this case, fMRI studies identified a example, see Stark et al.'s study [46] that replicated Wagner et al.'s
neural correlate of the priming effect in that the magnitude of the [45] overall results but not the correlations between priming and
neural activation observed in particular brain areas during a repeti- subsequent memory).
tion is smaller than it is during the first presentation. This is called Studies supporting a priming account of repetition effects have
the repetition suppression effect [36], and which brain areas display included some in which priming and repetition suppression occur on a
this effect depends on the type of stimuli and type of memory task timescale of seconds to minutes [6,35,37,38,40,42–44], whereas others
used. The amount of spacing between repetitions has been shown to occur on a timescale of hours and days [39,45,46]. This raises the
affect the magnitude of the repetition suppression effect [37–39]. possibility, suggested by Henson [36], that these effects may reflect
After a null or short lag, the second occurrence is associated with a multiple mechanisms that may operate over different timescales.
low neural activation (i.e., a high level of repetition suppression). As
the lag increases, the level of activation gradually increases until it 2.2. Spacing and the cellular bases of learning and memory
approaches the level observed during the first occurrence (i.e., no
more repetition suppression). Thus, repetition suppression is stron- It is widely agreed that learning occurs when experience produces
gest with massed repetitions and becomes weaker and weaker as changes in an organism that last for an extended period of time, thus
the lag increases. producing a relatively durable memory. In terms of brain processes, a
Xue et al. [40] have observed this phenomenon using faces that critical role is thought to be played by synaptic plasticity or changes in
were repeated four times, either in a massed fashion or spaced with a the synaptic connections between pre- and post-synaptic neurons.
small number of other faces intervening between repetitions. The Synapses that are active during a learning task presumably undergo
behavioral spacing effect was observed in that spaced faces were more structural and functional changes that strengthen the connection
often recognized in a final memory test than massed faces. On the between neurons such that the activation of pre-synaptic neurons in
neural level, three important results were found in the fusiform cortex the future will more readily result in a corresponding activation of
(FC), which is the region of interest for face processing [41]: (a) the FC post-synaptic neurons. Although experience may initiate synaptic
was more activated for those items that subsequently would be changes, their longevity depends on a number of processes including
recognized than for those that would not be recognized; (b) the FC gene expression and protein synthesis that play out over time after
displayed the repetition suppression pattern since its activation level the initial experience [47]. Thus, learning and the formation of long-
decreased from repetition to repetition; and (c) the activation level term memories must be conceptualized not as a process that is
observed in the FC was affected by how repetitions were spaced, with instantaneous but as one that takes place over an extended period.
the spaced condition yielding a lesser magnitude of repetition sup- Initial memories are labile and easily disrupted, but they become
pression. Therefore, spacing of repetitions led to reduced repetition stabilized through a temporally-related process known as consolida-
suppression which, in turn, was associated with better final recogni- tion. Usually two phases of consolidation are identified: Synaptic
tion. These findings suggest a relationship between the amplitude of consolidation and systems consolidation [48,49]. Synaptic consolida-
the repetition suppression and the spacing effect. Similar results have tion helps establish long-lasting cellular and synaptic changes and
been obtained by Xue et al. [42] with Korean characters, by Callan and usually is complete in a period ranging from minutes to several hours.
Schweighofer [43] with pseudoword-word pairs. Still more evidence Systems consolidation involves the reorganization of information in
consistent with a deficient processing account of the spacing effect the brain. Thus, memory that originally relied on one region of the
was obtained by Van Strien et al. [44] who used electroencephalogram brain (e.g., the hippocampus) may come to rely primarily on another
(EEG) technology to measure electrical activity in the brain. In terms of brain region (e.g., the neocortex). This process occurs over a long time
behavior, they obtained a typical spacing effect in the free recall of period that may range from days to years depending on specific
words. While learners were studying the words, however, event- circumstances.
related potentials (electrical activity in the brain which is time locked The role of synaptic plasticity in learning and the formation of
to stimulus presentation) revealed reduced activity associated with the long-term memory has been studied in the context of long-term
occurrence of the second massed presentation in comparison with the potentiation (LTP), a phenomenon that has been demonstrated both
second spaced presentation. in vivo and in vitro to create lasting changes in the synaptic strength
Wagner et al. [45] investigated the relationship between spacing, between two neurons [50–54]. When both pre- and post-synaptic
priming and neural suppression using word lists that were studied neurons are simultaneously stimulated or when intense electrical
52 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59

stimulation (e.g., theta burst stimulation) is applied to the pre- initiated by the first stimulus application or because high-threshold
synaptic neuron, the post-synaptic neuron reacts by changing synapses have not had time to become sensitized and, thus, remain
its future reactivity to the pre-synaptic stimulation, such that resistant to the effect of stimulation.
less pre-synaptic activation is necessary to trigger a post-synaptic Spacing trials also produces other cellular changes that may
response [55]. Two major phases of LTP have been identified. Early contribute to superior learning. Wu et al. [65] showed that it can
LTP (E-LTP) is characterized by the activation of protein kinases but affect the morphology of neurons. Spaced stimulations of mammalian
not protein synthesis and, by itself, produces relatively short-lived hippocampal cells triggered the development of new neuronal den-
changes in synaptic strength. In contrast, Late LTP (L-LTP) requires dritic extensions that were durable, whereas massed stimulations
both protein synthesis and gene expression and produces more failed to do so. Because dendrites bear the post-synaptic receptors,
enduring changes in synaptic strength. The processes involved in new dendritic extensions suggest the creation of new synapses.
L-LTP are essentially identical to processes involved in synaptic Spacing also appears to boost the survival of newly-born neural cells.
consolidation. New neurons are continuously being created in the hippocampus, and
LTP studies with animals suggest that differences in the effects some of the new neurons survive while some others do not. Sisti et al.
of massed and spaced practice on learning and memory may be at [66] used a task in which rats learn how to locate a platform in a pool
least partially attributable to their effects on synaptic plasticity full of opaque liquid and showed that spacing the training trials led to
[56]. Spacing the electrical stimulations that induce LTP seems to better acquisition performance than massed trials and that better
elicit a greater effect than massing stimulations. For instance, learning was associated with a higher rate of neurons surviving in the
Scharf et al. [57], working in vitro, delivered tetra burst stimula- hippocampus. In another study they showed that memory retention
tions into hippocampal slices of mice, either in a massed (20 s of two weeks later was best in the spaced condition and that the level of
lag between two successive stimulations) or in spaced fashion performance was correlated with cell survival.
(5 min). Spaced stimulations led to a larger LTP effect than massed In summary, spacing the occurrences of repeated stimulations or
stimulations but this advantage was attenuated by the application learning trials has been shown to trigger a cascade of cellular processes
of anisomycin, a protein-synthesis inhibitor. These results were that massed stimulations do not. These mechanisms are thought to
interpreted to indicate that spacing produces greater L-LTP by contribute to the behavioral changes that are regularly observed.
facilitating recruitment of protein-synthesis-dependent processes. Recent reviews of the detailed cellular and biochemical mechanisms
Additional support for this hypothesis has been obtained from associated with the spacing effect can be found in [62,67].
learning studies that incorporated genetic alteration techniques (e.g.,
[58–60]). In one study, Josselyn et al. [59] found that, when viral 2.3. The effect of longer inter-study intervals
vector-mediated gene transfer was used to increase levels in the rat
amygdala of an important transcription factor involved in protein The research we have considered so far has, for the most part,
synthesis (CREB), the degree of fear conditioning produced by massed focused on the effect of spacing when inter-study intervals are short,
training was greatly increased relative to that brought about in typically varying within a single day. These processes have been
control animals. In another study investigating object recognition in related to encoding processes and synaptic consolidation that seem
mutant mice and control animals, Genoux et al. [58] found that to be complete in a matter of minutes to several hours (e.g., [48]).
memory was poorer in control mice following training with relatively However, spacing can be effective over much longer intervals. For
short, as opposed to long, spacing intervals. Practice involving shorter example, Cepeda et al. [68], (Exp. 2) had participants study unfamiliar
spacing intervals was characterized in these animals by higher levels facts and object names, and varied inter-study intervals in terms of
of an important protein phosphatase (PP1) that negatively regulates days. Large and significant benefits to subsequent memory were found
(reduces) protein synthesis. However, memory following training as the inter-study interval increased from one day to 28 days. The fact
with short spacing intervals was much improved in mutant mice that the benefit of spacing applies to longer time scales has practical
who were altered genetically so that PP1 was inhibited during importance because, in educational settings, practice and reviews
learning. In contrast, performance was similar in mutant and normal often take place over relatively long periods (e.g., several days).
mice in the long spacing condition, suggesting that only the short It seems likely that the processes responsible for the effects of
spacing condition was associated with a deficit in protein synthesis distributing practice over one or more days are different from those
due to high levels of PP1. that mediate the effects when the degree of spacing is measured in
The findings reported by Genoux et al. [58] suggest a general minutes or hours. First, the influence of deficient processing will be
mechanism that may partially underpin the deficient-processing much weaker (or even non-existent) after 24 h in comparison to
account of the spacing effect that was discussed earlier. That is, several minutes. Second, the degree of forgetting between initial
short inter-repetition intervals may tilt the balance between posi- learning and the occurrence of a repetition will be substantially
tive and negative regulatory factors in favor of the latter, inhibiting greater after a period measured in days. Third, synaptic consolida-
protein synthesis and synaptic plasticity. Longer inter-repetition tion should be largely complete with inter-study intervals of a day
intervals may allow the balance to shift in favor of positive factors, or more, but the degree to which systems consolidation can take
facilitating long-term memory formation (e.g., [61,62]). place between repetitions will vary with the length of the inter-
A potentially complementary deficient-processing-like mechan- repetition interval. Fourth, inter-study intervals of 24 h or more are
ism was identified by Kramar et al. [63] (see also [64]). These likely to include one or more nights of sleep between repetitions,
investigators found that greater LTP with spaced stimulations was and sleep has been shown to facilitate the consolidation of
associated with a set of processes in which an initial stimulus memories (for review, see [69]). Systems consolidation is thought
application triggers events that lead to LTP in some synapses and to involve the reorganization of memories. At one level, this may
also sensitize other high-threshold synapses that were resistant to involve shifting the burden of memory from one brain system to
the first stimulus application. These processes unfold over time. another (e.g., from the hippocampus to the neocortex). At another
Thus, when a second stimulus application is sufficiently spaced, it is level, systems consolidation is thought to involve the integration of
able to trigger LTP in a subpopulation of previously unaffected newly learned information with previous knowledge, and, in the
synapses that have become sensitized, increasing the total number process, the abstraction of structure in bodies of knowledge
of synapses for which LTP has been induced. In contrast, when the that are continually being updated (e.g., [49,70]). This implies that
second stimulus application is massed, it may have little impact for the nature of the original memory may change as consolidation
lack of responsive synapses either because LTP has already been progresses during the spacing interval between repetitions. Thus,
 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59 53

A further complication is that there seems to be no unique inter-

repetition interval that leads to optimal memory performance,
because the duration of the retention interval (i.e., the time between
the last study trial and the memory task) has an influence on which
degree of spacing is optimal (e.g., [5,68]). For instance, in Cepeda
et al.'s [68] study, after a retention interval of 10 days, the best lag was
only one day between repetitions (Experiment 1). In contrast, after a
retention interval of 168 days, the best lag was 28 days between
learning sessions (Experiment 2). Glenberg and Lehmann [76]
observed that, after a retention interval of a few minutes, a lag of
several minutes was better than a lag of one or seven days. In contrast,
after retention interval of seven days, the one-day lag was the best.
Cepeda et al. [77] replicated the pattern over longer time scales
by combining multiple lags (from 0 to 105 days) and retention
intervals (from 7 to 350 days). The best lag for a 7-day retention
interval was one day, while the best lag for 350 days of retention
Fig. 2. Percentage of correct recall as a function of lag and retention interval, both
interval was 21 days (see also [73]). The results are presented in
measured in days. Based on Fig. 3a of Cepeda et al. [77].
Fig. 2. Thus, the general picture is that if you want to retain your
history lesson longer, you will have to space the repetitions more.
spacing repetitions over a day or more may lead to better future However, if you target short-term success and not long-term
memory because it is likely to be based more on relatively durable retention, short repetition intervals may be more effective.
semantic information (e.g., gist) related to pre-existing knowledge
and less on shallow, ephemeral information (e.g, irrelevant, situation-
specific stimulus cues). Returning to our opening example of learning 4. The role of retrieval or reactivation
a history lesson, studying the lesson (or a part of the lesson) again one
or more days later may foster the encoding of meaningful, relevant Although deficient processing and synaptic plasticity (see above)
information because this semantic information is the most likely to are potential candidates to account at least partially for the advan-
have been retained over a long spacing interval and to have been tage of spaced over massed items, their applicability seems limited.
consolidated during sleep. In contrast, when a repetition occurs after A deficient-processing mechanism is not able to account for the
several minutes or hours, encoding may be less selective because it decline in performance observed when lags are longer than optimal
can be based on a larger number of different cues, some of which [77], because it predicts either that performance will be a mono-
might be shallow, irrelevant features (e.g., font of the letters, location tonically increasing function of lag or that performance will increase
on the page, or some recent episodic memory irrelevant to the monotonically until it reaches an asymptote. There also is nothing
knowledge itself) that are unlikely to support later retrieval (see [2]). in the hypothesis to explain why the optimal degree of lag depends
on the retention interval. Thus, deficient processing may contribute
to the effect of spacing repetitions as suggested by evidence we
reviewed earlier, but it does not seem sufficient to account for
3. More on the effect of spacing: complications the full range of effects associated with spacing. In this section, we
consider alternative mechanisms that also may contribute to the
Generally speaking, it is better to space study reviews widely effect of spacing.
than to space them just a little. However, there is an important
limitation to this generalization. Recent studies have revealed that
increasing the level of spacing between repetitions of to-be-learned 4.1. Study-phase retrieval
material leads to improvements in subsequent memory perfor-
mance up to a certain point, but increasing the spacing beyond this A leading cognitive account of the more complicated effects of
point leads to a decline in subsequent memory. That is, final spaced practice is called the study-phase-retrieval hypothesis. Its
memory performance is an inverted-U-shaped function of the name derives from the critical assumption that, in order for
degree of spacing between the study episodes (e.g., [8,34]). repeated practice to be effective, studying repeated information
This inverted-U-shaped function has been observed when spa- must remind the learner of his or her previous experience with the
cing is varied in the range of seconds and minutes (e.g. [5,71]) information (e.g., [78,79]). More difficult study-phase retrieval
although it may be seen more readily when study episodes and (provided it ultimately is successful) is assumed to have a greater
memory tasks are distributed over longer, more educationally beneficial effect on subsequent memory, in part because it is
relevant intervals of several days or more. For example, in the hypothesized to slow the rate of forgetting (e.g., [2,74]).
Cepeda et al.'s [68] (Exp. 2) study mentioned previously, the two This hypothesis predicts an inverted-U-shaped spacing function
learning sessions were separated by 0, 1, 7, 28, 84, or 168 days. Final because increasing the spacing between repeated occurrences of
recall occurred 168 days after the second study session. Memory information makes study-phase retrieval more difficult but also less
performance was best for the participants with 28 days of lag. The likely to be successful. Thus, as long as study-phase retrieval tends
inverted-U-shaped function also has been shown with educational to be successful, increased spacing will lead to better subsequent
materials like texts (e.g., [72]) and foreign vocabulary learning in memory. However, there must be a lag at which the negative effect
secondary school (e.g., [73]). This inverted-U-shaped function of increasingly frequent failures of study-phase retrieval begins to
appears with infants [74] and in classical conditioning in rats [75] outweigh the beneficial effect of increasingly difficult successes
as well. The following pattern seems to emerge: very short lags lead [34,71,80]. The fact that the optimal lag becomes longer as a
to low levels of performance compared to the optimal lag, while function of the length of the retention interval has been related
longer-than-optimal lags are less detrimental to performance. In by modeling studies to the assumption that greater spacing results
other words, the detrimental effect of spacing too much is not as in information being forgotten at a slower rate (e.g., [80]). Essen-
critical as that of not spacing enough. tially, if long inter-repetition lags slow the rate of forgetting relative
54 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59

to shorter lags, their beneficial effect will be most apparent after a a memory-altering effect. The basis of this effect is still a matter of
longer retention interval. some discussion, but there is evidence that more difficult retrieval
An alternative theoretical perspective, the encoding variability is more beneficial for later memory. Thus, subsequent memory is
hypothesis, arguably achieved its most sophisticated and elegant better to the extent that study-phase retrieval occurs after a longer
manifestation in the form of Glenberg's Component-Levels Theory interval of time [89], is based on recall rather than recognition
[81]. In this view, as the spacing between repetitions increases, (e.g., [90]), and occurs with less cue support (e.g., [91]).
successive occurrences of the same information become more likely Modigliani [92] varied the spacing interval between original
to be encoded differently due to a changing environment. Variable encoding of small sets of words and a practice test without feedback
encoding is hypothesized to improve subsequent memory because (retrieval practice), and he subsequently assessed memory with a free
it provides more ways in which the target information can be recall test. The results were consistent with the study-phase-retrieval
accessed on a later test. Performance should be best when succes- hypothesis. Overall final free recall performance was an inverted-U-
sive encodings are independent of one another, that is, when there shaped function of spacing. These findings, in turn, were attributable
is no overlap in encoded information, a condition that should to the combined influence of three other effects that also were
preclude study-phase retrieval which presumably requires some consistent with expectations derived from the study-phase-retrieval
overlap between encoded events. hypothesis. First, recall on the practice tests declined as a function of
Although the encoding variability hypothesis can explain many spacing, consistent with the assumption that increased spacing makes
effects associated with distributed practice and has enjoyed a high study-phase retrieval more difficult. Second, final free recall for items
level of popularity among memory researchers, it is inconsistent with that were not retrieved on the practice test was poor and unrelated to
several lines of evidence that favor the study-phase-retrieval hypoth- spacing. Third, final recall of items that were successfully recalled on
esis instead. First, the spacing effect is observed only for those items the practice tests increased monotonically as a function of spacing.
that are recognized when they are repeated [4]. Second, when
context is varied to control encoding, repeating items after long lags 4.3. Reconsolidation
in the same context (yielding similar encodings and fostering
successful study-phase retrieval) yields better subsequent memory In view of the evidence that retrieval or reactivation of a
performance than repeating items at the equivalent lags in a different memory may play a critical role in strengthening memory and in
context (yielding relatively different encodings) [2]. Third, recall contributing to the benefit of spaced practice, it is important to
performance following spaced practice often exceeds the level that consider research on reconsolidation which provides a potential link
would be expected if repetitions were encoded independently between reactivation of a prior memory and the synaptic plasticity
[80,82]. That is, performance significantly exceeds the maximum underlying learning and memory. Unlike consolidation which has
that is possible according to the encoding variability hypothesis, long been a central component of memory theory, reconsolidation
suggesting that repetitions are not encoded independently. has a relatively short history. After its initial discovery [93],
Related evidence also was obtained in a recent imaging study reconsolidation was largely ignored until a paper by Nader et al.
reported by Xue et al. [83]. In three experiments, faces or words were [94] rekindled interest and initiated a period of intense research
presented three or four times, then recall or recognition was measured activity that has yet to subside. (For reviews, see [48,49,95].)
after one or several hours of retention interval. Patterns of voxel Nader et al. [94] conditioned rats to fear a tone. A day later, some
activation were recorded in fMRI scanning during item presentations animals were reminded of the prior conditioning by the presentation
in specific brain areas known to be involved in encoding and of the tone alone, whereas others were not reminded. Then a protein
perceptual processing. The voxel patterns triggered by successive synthesis inhibitor (anisomycin) was administered either immediately
presentations of the same item were compared. It was shown that or several hours later. In a subsequent test, the memory of rats who
the voxel patterns were more similar from presentation to presenta- had received no reminder remained intact, unaffected by the aniso-
tion for those items that were subsequently recalled or recognized mycin. This indicated that the originally-learned memory was fully
than for those subsequently forgotten. The opposite pattern was never consolidated after one day. Yet, animals who received anisomycin
observed. This result is not consistent with the encoding variability immediately after a reminder subsequently exhibited impaired mem-
hypothesis but is consistent with the study-phase-retrieval hypothesis ory of the tone conditioning, whereas those who received anisomycin
because an item that is processed in the same way during successive several hours after the reminder did not. Together, these findings
presentations gains an advantage for subsequent retention. indicate that reactivating a consolidated memory can return it to a
labile state in which it is susceptible to modification. It must be re-
4.2. Testing effect stabilized or reconsolidated through a time-dependent process that
resembles consolidation. However, more recent research has indicated
The study-phase-retrieval hypothesis is based on the assump- that, although consolidation and reconsolidation both depend on a
tion that retrieval is integrally involved in strengthening memory. similar cascade of molecular and cellular events, there are identifiable
Therefore, it is interesting to consider research on the testing effect differences, suggesting that they should be considered to be distinct
which demonstrates that retrieval per se does, in fact, improve processes (e.g., [96–98]).
memory. A number of studies (e.g., [84,85]) have compared final The reconsolidation effect has been extended to a variety of
memory performance as a function of whether repetition involving behavioral paradigms (see [90,93]) and species, including humans. In
long lags entailed restudy opportunities (typical repetition) or the latter case, reactivation is usually followed by potentially interfering
practice tests without feedback (retrieval practice). Overall final learning events (e.g., [99,100]) or emotion-inducing stimuli (e.g., [101]).
memory performance is usually better following practice tests. For example, Hupbach et al. [100] had subjects learn a different
These studies also have varied the retention interval, and they find list of objects on each of two days and subsequently tested free
that the advantage conferred by retrieval practice is apparent recall. If subjects were reminded of list 1 before learning list 2,
primarily after relatively long retention intervals. This suggests subsequent recall of list 1 on the third day included significantly
(but does not prove) that retrieval practice may slow the rate of more intrusions from list 2 than if subjects had not been reminded
forgetting. (See [86,87] for an alternative interpretation.) of list 1 prior to learning list 2. The reminder presumably reacti-
Although the beneficial effect of retrieval practice is limited to vated the memory of list 1, rendering it modifiable so that elements
items that are successfully retrieved [88], the findings demon- of list 2 could intrude. Interestingly, this effect was eliminated if the
strate that the retrieval of previously encoded information has recall test was administered immediately after learning list 2,
 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59 55

suggesting that the intrusion effect was critically due to reconsoli- 5. Implications for education
dation which occurs over time. In the study by Strange et al. [101]
using emotion-inducing stimuli, subjects studied a list of words on The cognitive and neuroscientific research we have reviewed in
Day 1. On Day 2, reminder trials consisted of presenting only the this article indicates that spaced practice leads to better learning
first three letters to the subjects who had to recall the word. and longer-lasting memories than relatively massed practice. The
Pictures of negative emotional faces, known to disrupt consolida- fact that this phenomenon has been demonstrated in a very large
tion, were interspersed between words. On Day 3, a cued-recall task number of behavioral experiments involving a wide variety of
similar to that of Day 2 was performed. Those words that had species suggests that it is produced by very fundamental learning
immediately preceded a negative face on Day 2 were recalled mechanisms that are undoubtedly the products of our evolution-
less often on Day 3, suggesting that reactivated memories did not ary history. Brain imaging studies have related the spaced-practice
properly reconsolidate because of the presentation of the emotional effect to neural suppression, consistent with the hypothesis that if
stimulus. As in the Hupbach et al. study, if the final recall test took a repetition occurs too soon, it receives deficient processing. Other
place on Day 2, immediately after the reminders and exposure to studies, especially those involving LTP and consolidation have
the emotional faces, the deleterious effect of the emotional stimuli related the phenomenon to synaptic plasticity, the neuronal basis
was eliminated, implicating the involvement of a time-dependent of learning and memory. Finally, findings indicating that memory
reconsolidation process. The critical point of these studies, however, is an inverted-U-shaped function of spacing and that the optimal
is that, as in other species, reactivating a memory in humans also spacing is longer for longer retention intervals suggest the effect of
seems to return it to a labile state in which it can be modified. repeated practice is importantly affected by retrieval (or remind-
Furthermore, the process by which the memory is modified seems ing) through which repetition triggers the reactivation of prior
to unfold over time, although little is known about the temporal memories. Retrieval is now recognized by cognitive scientists as a
parameters in humans or how they might vary, depending on the critical and powerful process in learning. From the neuroscience
particular learning conditions involved. perspective, it seems likely that the effect of retrieval on learning
A growing number of scientists have begun to hypothesize is related to synaptic plasticity through the process of reconsolida-
that reconsolidation requires a fundamental change in our view of tion which may be triggered when a prior memory is reactivated.
consolidation-like processes in memory and that it offers the The most general implication of these findings for educational
promise of integrating consolidation-based theories with mechan- practice is that the basic mechanisms underlying the effectiveness of
isms hypothesized by cognitive scientists (e.g., [48,49,102,103]). spaced practice and retrieval should be embraced in order to max-
Thus, instead of the standard conceptualization of a memory under- imize learning and memory. Educators should exploit these funda-
going consolidation once, the truth may be that memories typically mental processes in devising instructional and learning strategies. Yet,
undergo repeated reconsolidation [48]. This raises the question of over 100 years after the benefits of spaced practice were discovered
what the adaptive function of reconsolidation may be. In this regard, [112], Dempster [1] lamented that spaced practice had not been
the labile state to which reactivation returns a memory may provide implemented frequently or widely in education. There is little evi-
an opportunity to update the memory with new and relevant dence that anything has changed in the last 25 years. In many cases,
information (e.g., [104]). Also, it has been proposed that reactivation students study and practice material over only a short time period
and subsequent reconsolidation may provide a mechanism for before being tested on it (e.g., a foreign vocabulary list or a given topic
strengthening memories (e.g., [105,106]). In line with these hypoth- in science per week) and most of them cram in order to prepare for
eses, reconsolidation has been proposed to underlie the effect of the tests. These represent suboptimal learning strategies. Although
retrieval practice that yields the testing effect [107]. there may be much we still do not fully understand at both the
Given that reactivation of a memory during learning must rely behavioral and neurophysiological levels, researchers are in increasing
on study-phase retrieval, it does not seem like too great a stretch to agreement that we know more than enough to apply spaced practice
suggest that reconsolidation processes may contribute to the effect and retrieval advantageously in educational settings (e.g., [3,113]).
of spacing practice. From a reconsolidation perspective, however,
why would memory improve with increased spacing between
original learning and a repetition? One possible scenario is based 5.1. The promise and limits of educational application
on a variation of the study-phase-retrieval explanation of spaced-
repetition effects that was proposed by Toppino and Gerbier [2]. In Despite the overwhelming scientific support for the efficacy of
this view, successful study-phase retrieval leading to activation of distributed practice and testing, much of this research has used
the original memory will be based on information that is shared relatively simple materials and tasks in laboratory settings. It is
between the original memory and the encoding of the information encouraging, therefore, that both the benefit of spacing (e.g., [8,114])
when it is repeated. As the lag between presentations increases, the and the benefit of testing (e.g., [115,116]) have been demonstrated
shared information will tend to be more stable information, capable with textual materials.2 Furthermore, an increasing number of
of bridging longer temporal intervals. In a sense, successful study- studies have demonstrated the facilitative effects of spaced practice
phase retrieval reveals what information in the situation is relevant and/or spaced tests in real educational settings involving students of
for sustaining long-term memory. Reconsolidation may update the various ages and a variety of content areas including, for example,
memory, emphasizing this relevant information or, perhaps equiva-
lently, may preferentially strengthen this relevant information. 2
Dempster [1] noted that some studies failed to obtain an effect of spaced
Despite the promise of reconsolidation, we must close this section practice when the repetition of textual materials was paraphrased (e.g., [117,118]),
with a cautionary note. Much more research is needed. Not all studies while others have shown that the benefits of spacing can be obtained even when
repetitions are not verbatim (e.g., [119,120]). This seems to be an example of a
have obtained evidence of reconsolidation, raising the possibility of
broader issue of the effect of identical versus varied repetitions. See [2] for a review.
boundary conditions that remain to be explicated (e.g., [108,109]). Varied repetitions can lead to superior performance with massed practice, often
Furthermore, although a great deal of research has been conducted on attenuating the spacing effect, most likely by reducing the effects of deficient
animal models of reconsolidation, relatively little research has been processing. Varied repetition can also enhance the effect of longer degrees of
conducted on human subjects [110]. Consequently, there has been spacing, provided that study-phase retrieval is successful. However, in cases for
which varied repetition impedes study-phase retrieval, it may reduce the benefit of
insufficient opportunity to rule out the possibility that reconsolidation- longer spacing intervals (e.g., [121]). For the purpose of educational applications,
like effects in humans can be explained by other memory mechanisms varied repetition is likely to be beneficial unless it is presented in a way that
(e.g., [111]). discourages or prevents study-phase retrieval.
56 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59

vocabulary learning [16,122], statistics [123], history facts [22], be beneficial in limited circumstances, especially early in learning
introductory psychology [124], phonics-based reading skills [125], with complex materials, in order to achieve correct initial encod-
middle-school science [126], and a college web-based course on the ing [133]. Also, memory performance is an inverted-U-shaped
brain and behavior [127]. Finally, recent research in cognitive science function of the spacing between repetitions. Therefore, although
has indicated that these effects are not limited to rote learning and long spacing intervals are generally recommended, they can be
memory but extend to more complex, educationally relevant indi- too long. If information is forgotten (i.e., cannot be retrieved or
cators of knowledge. Spaced testing has been shown to produce reactivated) when it is repeated, the repetition may have little
learning that generalizes to related information (e.g., [128]), and both benefit on subsequent memory (e.g., [5,68]). What can be done?
testing and spaced practice support superior performance on criter- In an attempt to maximize learning effectiveness, one can use
ion tests requiring inferences or applications that go beyond the an expanding spacing schedule that counteracts short-term for-
specifics of prior training (e.g., [129,130]). getting while still reaping the benefit of long lags. This kind of
Dempster [1] identified the lack of applied research as one of the schedule consists of initial short delays between repetitions
biggest obstacles to widespread, systematic use of distributed followed by increasingly longer delays between subsequent repe-
practice and testing in education. Teachers and curriculum devel- titions. The initial short spacing intervals allow efficient encoding
opers operate within a complex environment in which many factors and success in retrieving or reactivating memories of earlier
are at play. Basic scientific research, such as that reviewed in this occurrences. Memory for information strengthened in this fashion
article, identifies critical variables that affect learning and memory, can then bridge longer intervals that are presented later, thus
but it does not answer questions of how these variables should be gaining the benefits of spaced practice.
specifically applied for maximum benefit in particular educational Although there has been some controversy about the efficacy of
contexts. Applied research will be necessary to answer these expanding schedules, recent research suggests that expanding
questions, and increased emphasis on such applied research seems schedules are efficient and effective when forgetting is too rapid
to be more than justified by impressive body of basic scientific to bridge long spacing intervals at the outset of learning (e.g., [134]
findings. Until the needed applied research has been conducted, but see [12]). Thus, expanding schedules are likely to be especially
however, considerable educational benefits may be reaped to the useful in high interference conditions that produce rapid forgetting.
extent that teachers – and students – are aware of the power of They also may be useful when the learning and memory ability of
distributed practice and testing and apply their own ingenuity in an the students is limited by characteristics such as age (e.g., [135]) and
attempt to use these fundamental principles systematically in their working memory capacity (e.g., [136]). Expanding schedules also
particular educational domains. With this more limited goal in are indicated as a bridge to very long spacing intervals measured
mind, it may be useful to consider general motivations, guidelines across days or weeks (e.g., [137,138]). In all cases, optimal perfor-
and examples related to why and how spaced practice and testing mance will be achieved when the initial spacing intervals are the
could be used in educational settings. longest that can be tolerated without producing high levels of
forgetting before a repetition occurs. In some cases, this could
5.2. Learning for the long term require initial massing. However, after satisfactory encoding and
initial short intervals between repetitions, one should try to multi-
The goal of education is not merely to learn new information or ply the number of encounters with the to-be-learned information
skills but to learn them and retain them for a long time – possibly and to distribute these encounters over a long time period (weeks,
for a lifetime. Learning in school without long-term retention is a months, and even years) in order to yield long-term retention and
waste of time and effort. The goal of learning for the long term leads efficient use of the information in new contexts (i.e., transfer).
to three practical recommendations based on research indicating Within-day spacing seems less relevant for long-term retention
that durable memories require initial learning to be supplemented than between-day spacing, but is more useful than massed practice
by relearning and retrieval practice that is distributed over time. (e.g., [125]).
First, massed practice generally should be avoided, based on a
wealth of evidence that it usually leads to deficient processing and 5.3. Implementation of distributed practice in the classroom
poor retention. Second, the intervals between distributed practice
opportunities should be as long as possible because research The principles we are advocating are easily implementable and
generally indicates that longer spacing intervals produce more very powerful in virtually all contexts and with all materials [3].
durable memories. Third, spaced practice should include formal or Following these principles does not necessarily require using
informal practice tests (i.e., situations that require effortful retrieval) different or unusual instructional techniques. It may suggest a
because they afford challenging retrieval practice that research has different mix of common methods, but it primarily calls for a
linked to memories that seem more resistant to forgetting. different way of scheduling the material.
If these recommendations are to be followed, two common Instructors frequently summarize at the end of a class session
traps must be avoided. First, one should not believe that the ease the material that was covered that day. That is a good practice.
of initial learning is a good predictor of retention. Although it may However, it might be beneficial to summarize the previous day's
be counterintuitive, easy, rapid learning and long-term retention lessons at the beginning of the next class, repeating information
are not necessarily correlated [131]. For example, learning through after a long enough spacing interval that the information has
massed practice often seems easy (e.g., [131,132]), but it is an undergone a significant amount of consolidation.
inefficient use of time that leads to poor long-term retention. In Instructors also sometimes plan review sessions into their
contrast, learning through spaced practice may be more difficult, courses. However, the optimal scheduling of these reviews is
but it is long lasting. Second, one should not believe that, once challenging within the limited time frame of a course. If reviews
knowledge is acquired, it need not be reviewed. Even perfect cover long periods of time (e.g., only at the end of a semester), as
performance is not an inoculation against the inexorable effect of usually is the case, they will have the advantage that some of the
forgetting. Good retention requires spaced practice and review material from early in the period being covered will be reviewed
even after initial learning seems to be complete. after a relatively long spacing interval, but they will have the
Following the three general recommendations listed above disadvantage that some recently covered material will be reviewed
would constitute an effective educational strategy, but it might after only a short spacing interval and that all of the material will be
not be optimal because there are exceptions. Massed practice may reviewed infrequently. The approach we are advocating would
 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59 57

place less emphasis on the first instructional period with a set of Unfortunately, there are reasons why getting students to sponta-
information than is currently the norm and would favor multiple neously distribute their study time may be not be easy. Aside from
extended review or reinstruction periods instead. Because the the challenge of being planful and disciplined, students may find
burden of learning is shifted from a single learning session to it more difficult to restudy a piece of information after a long delay
multiple repeated sessions, all of the sessions can be relatively (e.g., one week) than after a short delay (e.g., a few hours). The
short. The follow-up sessions should be scheduled at the longest impression of having forgotten much of the information can be
intervals that are compatible with multiple opportunities for review discouraging. In research studies, students prefer to use a massed
or reinstruction or, if an expanding schedule is employed, at strategy (e.g., [132]) and, even after experiencing the benefit of
increasingly long intervals. Given the relationship between spacing spaced practice over massed practice, they still may predict that
and the durability of memory, it would be advantageous to reduce massed practice will produce better long-term memory (e.g., [140]).
the rigid temporal structure of education systems (i.e., discrete Even if students understood the benefit of spacing for long-term
classes, years in school, etc.) allowing the curriculum to be arranged retention, many might not care. They might be more interested in
so that some reviews could take place after extremely long periods their immediate test performance and the resulting grade than on
(e.g., after a year or more). Appropriate spacing is made essentially whether they remember much later. In that case, cramming (massed
impossible when very compressed learning schedules are required, practice right before a test) actually may be a reasonable strategy
for example, in summer classes or immersion programs in a foreign because massed practice is most beneficial when the retention
language. To foster long-term retention in these cases, it may be interval is very short.
necessary to include distributed follow-up reviews after weeks or Students also are advised to maximize their learning by supple-
months to reactivate the material. menting their spaced-study regimen with self-testing. As noted
The importance of reactivating knowledge in establishing durable earlier, effortful retrieval has a strong beneficial effect on learning.
memories implies that instructional strategies should encourage So, combining spaced practice with self-testing may be an optimal
retrieval by the students. Repeating information may have that effect, study strategy. However, students largely are unaware that testing is
but placing the repetition in the context of questions is likely to be a powerful learning activity (e.g., [141–143]). When they self-test,
more effective because questions encourage more effortful retrieval. they do so to evaluate their level of learning, and, even then, they
Thus, asking questions should be an important part of an instructor's tend to do it ineffectively. That is, students tend to study and then
repertoire. For many instructors it is, although it is not clear how often test themselves soon after. The consequence is that they get many
the technique is applied optimally. Likewise, formal testing should be test items correct that they would not have gotten correct after a
considered to be an important part of an effective instructional longer retention interval. So, they mistakenly believe that they will
strategy. Although tests typically are considered to be merely an remember more on the instructor's test than they actually will. Self-
evaluation tool, research has made it clear that they are a powerful testing whether as an evaluative technique or as a learning strategy
learning tool in and of themselves because they require the reactiva- should be conducted after an extended period of time has passed.
tion of previously-learned knowledge. Thus, if an instructor begins A convenient way to find support in establishing an effective
class with a summary of material covered in the last class period, it spacing and testing schedule for at least some kinds of material may
might be helpful to incorporate a quiz prior to summarizing. It should be to use certain computer programs that have been designed for
be noted that questions may be beneficial even if the reactivated this purpose. Free programs can be found on the Internet in which
information is insufficient to yield a correct answer. There is evidence one's own material can be introduced (e.g., [144]). These programs
that retrieval attempts that are followed by an opportunity to study present the items for test trials to the learner and provide feedback
the correct answer potentiate learning. That is, learning appears to be (correct answers) to allow restudying. The delay before the next
better following a retrieval attempt than when no retrieval attempt review of any given item is calculated on an item-by-item basis,
was made (e.g., [139]). The basis of this effect remains unclear but it depending on performance during tests. The logic behind the
may reflect the updating function of reconsolidation. computation is that a successful response will increase the duration
Instructors can use tests and homework as means of scheduling of the next interval, whereas a failure will shorten it. More
student studying. Left to their own devices, many students will not sophisticated algorithms are being developed experimentally and
study until right before a test. Thus, scheduling infrequent tests have proven to be very efficient (e.g., [145]).
encourages a form of massed practice known as “cramming.” More So, how should you schedule your learning of the history
frequent tests encourage a steadier stream of studying, but the tests lesson?
then are being administered after shorter intervals which may yield
non-optimal spacing. Therefore, it seems advisable to use cumulative
quizzes and tests that cover recent material as well as material covered
References
longer ago. Homework is another staple of educational practice.
However, instead of devoting homework to material that was covered
[1] Dempster FN. The spacing effect: a case study in the failure to apply the
most recently in class, some of each homework assignment could be results of psychological research. Am Psychol 1988;43(8):627–34.
devoted to material that was covered at an earlier date. Manipulating [2] Toppino TC, Gerbier E. About practice: repetition, spacing, and abstraction.
homework assignments in this way seems like a painless way (at least In: Ross Brian H, editor. Psychology of learning and motivation, 60. Burling-
ton: Academic Press; 2014. p. 113–89.
for the instructor) to schedule spaced practice. [3] Dunlosky J, Rawson KA, Marsh EJ, Nathan MJ, Willingham DT. Improving
students' learning with effective learning techniques promising directions
5.4. Practical pointers for learners from cognitive and educational psychology. Psychol Sci Public Interest
2013;14(1):4–58.
[4] Melton AW. Repetition and retrieval from memory. Science 1967;158
Spaced practice works regardless of who controls the study (3800):532.
schedule. Thus, the most successful students in a course will [5] Glenberg AM. Monotonic and nonmonotonic lag effects in paired-associate
and recognition memory paradigms. J Verbal Learn Verbal Behav 1976;15
incorporate distributed practice into their study regimen. That is,
(1):1–16.
they will study the same material multiple times, separated by [6] Russo R, Parkin AJ, Taylor SR, Wilks J. Revising current two-process accounts of
relatively long inter-study intervals. This requires the students to spacing effects in memory. J Exp Psychol Learn Mem Cogn 1998;24(1):161–72.
be planful and disciplined enough to be able to map out an [7] Paivio A. Spacing of repetitions in the incidental and intentional free recall of
pictures and words. J Verbal Learn Verbal Behav 1974;13(5):497–511.
appropriate study schedule and then stick to it. The likely result will [8] Rawson KA, Kintsch W. Rereading effects depend on time of test. J Educ
be better grades and better retention, long after the course is over. Psychol 2005;97(1):70–80.
58 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59

[9] Simon DA, Bjork RA. Metacognition in motor learning. J Exp Psychol Learn [46] Stark SM, Gordon B, Stark CEL. Does the presence of priming hinder
Mem Cogn 2001;27(4):907–12. subsequent recognition or recall performance. Memory 2008;16(2):157–73.
[10] Bird S. Effects of distributed practice on the acquisition of second language [47] Dudai Y. The neurobiology of consolidations, or, how stable is the engram.
English syntax. Appl Psycholinguist 2010;31(04):635–50. Annu Rev Psychol 2004;55:51–86.
[11] Rohrer D, Taylor K. The effects of overlearning and distributed practise on the [48] Dudai Y. The restless engram: consolidations never end. Annu Rev Neurosci
retention of mathematics knowledge. Appl Cogn Psychol 2006;20(9):1209–24. 2012;35(1):227–47.
[12] Kornmeier J, Spitzer M, Sosic-Vasic Z. Very similar spacing-effect patterns in [49] Hardt O, Einarsson EÖ, Nader K. A bridge over troubled water: reconsolida-
very different learning/practice domains. PLoS One 2014;9(3):e90656. tion as a link between cognitive and neuroscientific memory research
[13] Cepeda NJ, Pashler H, Vul E, Wixted JT, Rohrer D. Distributed practice in traditions. Annu Rev Psychol 2010;61(1):141–67.
verbal recall tasks: a review and quantitative synthesis. Psychol Bull [50] Bliss TV, Collingridge GL. A synaptic model of memory: long-term potentia-
2006;132(3):354–80. tion in the hippocampus. Nature 1993;361(6407):31–9.
[14] Delaney PF, Verkoeijen PPJL, Spirgel A. Chapter 3 – spacing and testing [51] Izquierdo I, Bevilaqua LM, Rossato JI, Silva WCD, Bonini J, Medina JH, et al.
effects: a deeply critical, lengthy, and at times discursive review of the The molecular cascades of long-term potentiation underlie memory con-
literature. In: Ross Brian H, editor. Psychology of learning and motivation, solidation of one-trial avoidance in the CA1 region of the dorsal hippocam-
vol. 53. San Diego: Academic Press; 2010. p. 63–147. pus, but not in the basolateral amygdala or the neocortex. Neurotox Res
[15] Donovan JJ, Radosevich DJ. A meta-analytic review of the distribution of practice 2008;14(2–3):273–94.
effect: now you see it, now you don't. J Appl Psychol 1999;84(5):795–805. [52] Lynch MA. Long-term potentiation and memory. Physiol. Rev. 2004;
[16] Bloom KC, Shuell TJ. Effects of massed and distributed practice on the learning 84(1):87–136.
and retention of second-language vocabulary. J Educ Res 1981;74(4):245–8. [53] Martin SJ, Grimwood PD, Morris RGM. Synaptic plasticity and memory: an
[17] Bahrick HP, Bahrick LE, Bahrick AS, Bahrick PE. Maintenance of foreign evaluation of the hypothesis. Annu Rev Neurosci 2000;23(1):649–711.
language vocabulary and the spacing effect. Psychol Sci 1993;4(5):316–21. [54] Whitlock JR, Heynen AJ, Shuler MG, Bear MF. Learning induces long-term
[18] Bahrick HP, Hall LK. The importance of retrieval failures to long-term potentiation in the hippocampus. Science 2006;313(5790):1093–7.
retention: a metacognitive explanation of the spacing effect. J Mem Lang [55] Cooke SF, Bliss TVP. Plasticity in the human central nervous system. Brain
2005;52(4):566–77. 2006;129(7):1659–73.
[19] Cornell EH. Distributed study facilitates infants' delayed recognition mem- [56] Kornmeier J, Sosic-Vasic Z. Parallels between spacing effects during beha-
ory. Mem Cognit 1980;8(6):539–42. vioral and cellular learning. Front Hum Neurosci 2012;6.
[20] Toppino TC. The spacing effect in young children's free recall: support for [57] Scharf MT, Woo NH, Lattal KM, Young JZ, Nguyen PV, Abel T. Protein
automatic-process explanations. Mem Cognit 1991;19(2):159–67. synthesis is required for the enhancement of long-term potentiation and
[21] Toppino TC, DeMesquita M. Effects of spacing repetitions on children's long-term memory by spaced training. J Neurophysiol 2002;87(6):2770–7.
memory. J Exp Child Psychol 1984;37(3):637–48. [58] Genoux D, Haditsch U, Knobloch M, Michalon A, Storm D, Mansuy IM.
[22] Carpenter SK, Pashler H, Cepeda NJ. Using tests to enhance 8th grade students' Protein phosphatase 1 is a molecular constraint on learning and memory.
retention of U.S. history facts. Appl Cogn Psychol 2009;23(6):760–71. Nature 2002;418(6901):970–5.
[23] Balota DA, Duchek JM, Paullin R. Age-related differences in the impact of [59] Josselyn SA, Shi C, Carlezon WA, Neve RL, Nestler EJ, Davis M. Long-term
spacing, lag, and retention interval. Psychol Aging 1989;4(1):3–9. memory is facilitated by camp response element-binding protein over-
[24] Simone PM, Bell MC, Cepeda NJ. Diminished but not forgotten: effects of
expression in the Amygdala. J Neurosci 2001;21(7):2404–12.
aging on magnitude of spacing effect Benefits. J Gerontol B Psychol Sci Soc
[60] Pagani MR, Oishi K, Gelb BD, Zhong Y. The phosphatase SHP2 regulates the
Sci 2013;68(5):674–80.
spacing effect for long-term memory induction. Cell 2009;139(1):186–98.
[25] Tully T, Preat T, Boynton SC, Del Vecchio M. Genetic dissection of con-
[61] Abel T, Martin KC, Bartsch D, Kandel ER. Memory suppressor genes:
solidated memory in Drosophila. Cell 1994;79(1):35–47.
inhibitory constraints on the storage of long-term memory. Science
[26] Carew TJ, Pinsker HM, Kandel ER. Long-term habituation of a defensive
1998;279(5349):338–41.
withdrawal reflex in aplysia. Science 1972;175(4020):451–4.
[62] Philips GT, Kopec AM, Carew TJ. Pattern and predictability in memory
[27] Deisig N, Sandoz J-C, Giurfa M, Lachnit H. The trial-spacing effect in olfactory
formation: from molecular mechanisms to clinical relevance. Neurobiol
patterning discriminations in honeybees. Behav Brain Res 2007;176(2):314–22.
Learn Mem 2013;105:117–24.
[28] Lattal MK. Trial and intertrial durations in Pavlovian conditioning: issues of
[63] Kramar EA, Babayan AH, Gavin CF, Cox CD, Jafari M, Gall CM, et al. Synaptic
learning and performance. J Exp Psychol Anim Behav Process 1999;25(4):433–50.
evidence for the efficacy of spaced learning. Proc Natl Acad Sci 2012;
[29] Anderson JR, Schooler LJ. Reflections of the environment in memory. Psychol
109(13):5121–6.
Sci 1991;2(6):396–408.
[64] Lynch G, Kramár EA, Babayan AH, Rumbaugh G, Gall CM. Differences
[30] Magliero A. Pupil dilations following pairs of identical and related to-be-
between synaptic plasticity thresholds result in new timing rules for
remembered words. Mem Cognit 1983;11(6):609–15.
[31] Johnston WA, Uhl CN. The contributions of encoding effort and variability to maximizing long-term potentiation. Neuropharmacology 2013;64:27–36.
[65] Wu GY, Deisseroth K, Tsien RW. Spaced stimuli stabilize MAPK pathway
the spacing effect on free recall. J Exp Psychol Hum Learn. 1976;2(2):153–60.
[32] Shaughnessy JJ, Zimmerman J, Underwood BJ. Further evidence on the MP-DP activation and its effects on dendritic morphology. Nat Neurosci 2001;4(2):151–8.
effect in free-recall learning. J Verbal Learn Verbal Behav 1972;11(1):1–12. [66] Sisti HM, Glass AL, Shors TJ. Neurogenesis and the spacing effect: learning
[33] Hintzman DL, Summers JJ, Eki NT, Moore MD. Voluntary attention and the over time enhances memory and the survival of new neurons. Learn Mem
spacing effect. Mem Cognit 1975;3(5):576–80. 2007;14(5):368–75.
[34] Toppino TC, Bloom LC. The spacing effect, free recall, and two-process [67] Naqib F, Sossin WS, Farah CA. Molecular determinants of the spacing effect.
theory: a closer look. J Exp Psychol Learn Mem Cogn 2002;28(3):437–44. Neural Plast 2012;2012.
[35] Challis BH. Spacing effects on cued-memory tests depend on level of [68] Cepeda NJ, Coburn N, Rohrer D, Wixted JT, Mozer MC, Pashler H. Optimizing
processing. J Exp Psychol Learn Mem Cogn 1993;19(2):389–96. distributed practice: theoretical analysis and practical implications. Exp
[36] Henson RN. Neuroimaging studies of priming. Prog Neurobiol 2003; Psychol Former Z Für Exp Psychol 2009;56(4):236–46.
70(1):53–81. [69] Rasch B, Born J. About sleep's role in memory. Physiol Rev 2013;
[37] Henson R. Neuroimaging evidence for dissociable forms of repetition 93(2):681–766.
priming. Science 2000;287(5456):1269–72. [70] Wang SH, Morris G. Hippocampal-neocortical interactions in memory formation,
[38] Henson RN, Rylands A, Ross E, Vuilleumeir P, Rugg MD. The effect of consolidation, and reconsolidation. Annu Rev Psychol 2010;61:49–79.
repetition lag on electrophysiological and haemodynamic correlates of visual [71] Verkoeijen PPJL, Rikers RMJP, Schmidt HG. Limitations to the spacing effect:
object priming. NeuroImage 2004;21(4):1674–89. demonstration of an inverted u-shaped relationship between interrepetition
[39] van Turennout M. Modulation of neural activity during object naming: spacing and free recall. Exp Psychol Former Z Für Exp Psychol 2005;
effects of time and practice. Cereb Cortex 2003;13(4):381–91. 52(4):257–63.
[40] Xue G, Mei L, Chen C, Lu Z-L, Poldrack R, Dong Q. Spaced learning enhances [72] Verkoeijen PPJL, Rikers RMJP, Özsoy B. Distributed rereading can hurt the
subsequent recognition memory by reducing neural repetition suppression. spacing effect in text memory. Appl Cogn Psychol 2008;22(5):685–95.
J Cogn Neurosci 2010;23(7):1624–33. [73] Küpper-Tetzel CE, Erdfelder E, Dickhäuser O. The lag effect in secondary
[41] Kanwisher N, Yovel G. The fusiform face area: a cortical region specialized for the school classrooms: enhancing students' memory for vocabulary. Instr Sci
perception of faces. Philos Trans R Soc B Biol Sci 2006;361(1476):2109–28. 2014;42(3):373–88.
[42] Xue G, Mei L, Chen C, Lu Z-L, Poldrack RA, Dong Q. Facilitating memory for [74] Rovee-Collier C. Time windows in cognitive development. Dev Psychol
novel characters by reducing neural repetition suppression in the left 1995;31(2):147–69.
fusiform cortex. PLoS One 2010;5(10):e13204. [75] Parsons RG, Davis M. A metaplasticity-like mechanism supports the selection
[43] Callan DE, Schweighofer N. Neural correlates of the spacing effect in explicit of fear memories: role of protein kinase a in the amygdala. J Neurosci Off
verbal semantic encoding support the deficient-processing theory. Hum J Soc Neurosci 2012;32(23):7843–51.
Brain Mapp 2010;31(4):645–59. [76] Glenberg AM, Lehmann TS. Spacing repetitions over 1 week. Mem Cognit
[44] Van Strien JW, Verkoeijen PPJL, Van der Meer N, Franken IHA. Electrophy- 1980;8(6):528–38.
siological correlates of word repetition spacing: ERP and induced band [77] Cepeda NJ, Vul E, Rohrer D, Wixted JT, Pashler H. Spacing effects in learning a
power old/new effects with massed and spaced repetitions. Int J Psycho- temporal ridgeline of optimal retention. Psychol Sci 2008;19(11):1095–102.
physiol 2007;66(3):205–14. [78] Raaijmakers JGW. Spacing and repetition effects in human memory: applica-
[45] Wagner AD, Maril A, Schacter DL. Interactions between forms of memory: tion of the SAM model. Cogn Sci 2003;27(3):431–52.
when priming hinders new episodic learning. J Cogn Neurosci 2000;12 [79] Thios SJ, D'Agostino PR. Effects of repetition as a function of study-phase
(Suppl. 2):S52–60. retrieval. J Verbal Learn Verbal Behav 1976;15(5):529–36.
 E. Gerbier, T.C. Toppino / Trends in Neuroscience and Education 4 (2015) 49–59 59

[80] Benjamin AS, Tullis J. What makes distributed practice effective. Cognit [113] Carpenter SK, Cepeda NJ, Rohrer D, Kang SHK, Pashler H. Using spacing to
Psychol 2010;61(3):228–47. enhance diverse forms of learning: review of recent research and implica-
[81] Glenberg AM. Component-levels theory of the effects of spacing of repeti- tions for instruction. Educ Psychol Rev 2012;24(3):369–78.
tions on recall and recognition. Mem Cognit 1979;7(2):95–112. [114] Rawson KA. Why do rereading lag effects depend on test delay. J Mem Lang
[82] Begg I, Green C. Repetition and trace interaction: superadditivity. Mem 2012;66(4):870–84.
Cognit 1988;16(3):232–42. [115] Agarwal PK, Karpicke JD, Kang SHK, Roediger HL, McDermott KB. Examining
[83] Xue G, Dong Q, Chen C, Lu Z, Mumford JA, Poldrack RA. Greater neural the testing effect with open- and closed-book tests. Appl Cogn Psychol
pattern similarity across repetitions is associated with better memory. 2008;22(7):861–76.
Science 2010;330(6000):97–101. [116] Nungester RJ, Duchastel PC. Testing versus review: effects on retention. J Educ
[84] Roediger HL, Karpicke JD. The power of testing memory: basic research and Psychol 1982;74(1):18–22.
implications for educational practice. Perspect Psychol Sci 2006;1(3):181–210. [117] Dellarosa D, Bourne LE. Surface form and the spacing effect. Mem Cognit
[85] Toppino TC, Cohen MS. The testing effect and the retention interval: questions 1985;13(6):529–37.
and answers. Exp Psychol Former Z Für Exp Psychol 2009;56(4):252–7. [118] Glover JA, Corkill AJ. Influence of paraphrased repetitions on the spacing
[86] Halamish V, Bjork RA. When does testing enhance retention? A distribution- effect. J Educ Psychol 1987;79(2):198–9.
based interpretation of retrieval as a memory modifier J Exp Psychol Learn [119] Reder LM, Anderson JR. Effects of spacing and embellishment on memory for
Mem Cogn 2011;37(4):801–12. the main points of a text. Mem Cognit 1982;10(2):97–102.
[87] Kornell N, Bjork RA, Garcia MA. Why tests appear to prevent forgetting: a [120] Smith SM, Rothkopf EZ. Contextual enrichment and distribution of practice
in the classroom. Cogn Instr 1984;1(3):341–58.
distribution-based bifurcation model. J Mem Lang 2011;65(2):85–97.
[121] Appleton‐Knapp SL, Bjork RA, Wickens TD. Examining the spacing effect in
[88] Jang Y, Wixted JT, Pecher D, Zeelenberg R, Huber DE. Decomposing the
advertising: encoding variability, retrieval processes, and their interaction.
interaction between retention interval and study/test practice: the role of
J Consum Res, 32; 2005. p. 266–76.
retrievability. Q J Exp Psychol 2012;65(5):962–75.
[122] Sobel HS, Cepeda NJ, Kapler IV. Spacing effects in real-world classroom
[89] Pyc MA, Rawson KA. Testing the retrieval effort hypothesis: does greater
vocabulary learning. Appl Cogn Psychol 2011;25(5):763–7.
difficulty correctly recalling information lead to higher levels of memory. J
[123] Budé L, Imbos T, van de Wiel MW, Berger MP. The effect of distributed practice on
Mem Lang 2009;60(4):437–47.
students' conceptual understanding of statistics. High Educ 2011;62(1):69–79.
[90] Glover JA. The “testing” phenomenon: not gone but nearly forgotten. J Educ
[124] Cranney J, Ahn M, McKinnon R, Morris S, Watts K. The testing effect,
Psychol 1989;81(3):392–9. collaborative learning, and retrieval-induced facilitation in a classroom
[91] Carpenter SK, Delosh EL. Impoverished cue support enhances subsequent setting. Eur J Cogn Psychol 2009;21(6):919–40.
retention: support for the elaborative retrieval explanation of the testing [125] Seabrook R, Brown GDA, Solity JE. Distributed and massed practice: from
effect. Mem Cognit 2006;34(2):268–76. laboratory to classroom. Appl Cogn Psychol 2005;19(1):107–22.
[92] Modigliani V. Effects on a later recall by delaying initial recall. J Exp Psychol [126] McDaniel MA, Agarwal PK, Huelser BJ, McDermott KB, Roediger III HL. Test-
Hum Learn 1976;2(5):609–22. enhanced learning in a middle school science classroom: the effects of quiz
[93] Misanin JR, Miller RR, Lewis DJ. Retrograde amnesia produced by electro- frequency and placement. J Educ Psychol 2011;103(2):399–414.
convulsive shock after reactivation of a consolidated memory trace. Science [127] McDaniel MA, Wildman KM, Anderson JL. Using quizzes to enhance
1968;160(3827):554–5. summative-assessment performance in a web-based class: an experimental
[94] Nader K, Schafe GE, Le Doux JE. Fear memories require protein synthesis in the study. J Appl Res Mem Cogn 2012;1(1):18–26.
amygdala for reconsolidation after retrieval. Nature 2000;406(6797):722–6. [128] Chan JCK, McDermott KB, Roediger HL. Retrieval-induced facilitation: initially
[95] Nader K, Hardt O. A single standard for memory: the case for reconsolida- nontested material can benefit from prior testing of related material. J Exp
tion. Nat Rev Neurosci 2009;10(3):224–34. Psychol Gen 2006;135(4):553–71.
[96] Lee JLC, Everitt BJ, Thomas KL. Independent cellular processes for hippocampal [129] Karpicke JD, Blunt JR. Retrieval practice produces more learning than
memory consolidation and reconsolidation. Science 2004;304(5672):839–43. elaborative studying with concept mapping. Science 2011;331(6018):772–5.
[97] Li Y, Meloni EG, Carlezon WA, Milad MR, Pitman RK, Nader K, et al. Learning [130] Moulton CAE, Dubrowski A, MacRae H, Graham B, Grober E, Reznick R.
and reconsolidation implicate different synaptic mechanisms. Proc Natl Acad Teaching surgical skills: what kind of practice makes perfect: a randomized,
Sci 2013;110(12):4798–803. controlled trial. Trans Meet Am Surg Assoc 2006;124:66–75.
[98] Tronson NC, Taylor JR. Molecular mechanisms of memory reconsolidation. [131] Bjork RA, Bjork EL. Making things hard on yourself, but in a good way:
Nat Rev Neurosci 2007;8(4):262–75. creating desirable difficulties to enhance learning. Psychology and the real
[99] Forcato C, Burgos VL, Argibay PF, Molina VA, Pedreira ME, Maldonado H. world: essays illustrating fundamental contributions to society. First ed.. New
Reconsolidation of declarative memory in humans. Learn Mem 2007;14 York, NY: FABBS Foundation; 2011. p. 56–64.
(4):295–303. [132] Baddeley AD, Longman DJA. The influence of length and frequency of training
[100] Hupbach A, Gomez R, Hardt O, Nadel L. Reconsolidation of episodic session on the rate of learning to type. Ergonomics 1978;21(8):627–35.
memories: a subtle reminder triggers integration of new information. Learn [133] Roediger HL. Applying cognitive psychology to education translational
Mem 2007;14(1–2):47–53. educational science. Psychol Sci Public Interest 2013;14(1):1–3.
[101] Strange BA, Kroes MCW, Fan JE, Dolan RJ. Emotion causes targeted forgetting [134] Storm BC, Bjork RA, Storm JC. Optimizing retrieval as a learning event: when
of established memories. Front Behav Neurosci 2010;4(175). and why expanding retrieval practice enhances long-term retention. Mem
[102] Alberini CM. The role of reconsolidation and the dynamic process of long Cognit 2010;38(2):244–53.
term memory formation and storage. Front Behav Neurosci 2011;5:12. [135] Maddox GB, Balota DA, Coane JH, Duchek JM. The role of forgetting rate in
[103] Lee JLC. Reconsolidation: maintaining memory relevance. Trends Neurosci producing a benefit of expanded over equal spaced retrieval in young and
older adults. Psychol Aging 2011;26(3):661–70.
2009;32(8):413–20.
[136] Bui DC, Maddox GB, Balota DA. The roles of working memory and intervening
[104] Hupbach A, Hardt O, Gomez R, Nadel L. The dynamics of memory: context-
task difficulty in determining the benefits of repetition. Psychon Bull Rev
dependent updating. Learn Mem 2008;15(8):574–9.
2013;20(2):341–7.
[105] Inda MC, Muravieva EV, Alberini CM. Memory retrieval and the passage of
[137] Gerbier E, Koenig O. Influence of multiple-day temporal distribution of
time: from reconsolidation and strengthening to extinction. J Neurosci
repetitions on memory: a comparison of uniform, expanding, and contract-
2011;31(5):1635–43.
ing schedules,. Q J Exp Psychol 2012;65(3):514–25.
[106] Lee JLC. Memory reconsolidation mediates the strengthening of memories by
[138] Gerbier E. Toppino T.C. Koenig O. Optimizing retention through multiple
additional learning. Nat Neurosci 2008;11(11):1264–6.
study opportunities over days: the benefit of an expanding schedule of
[107] Finn B, Roediger HL. Enhancing retention through reconsolidation negative
repetitions, Memory, in press.
emotional arousal following retrieval enhances later recall. Psychol Sci [139] Kornell N, Hays MJ, Bjork RA. Unsuccessful retrieval attempts enhance
2011;22(6):781–6. subsequent learning. J Exp Psychol Learn Mem Cogn 2009;35(4):989–98.
[108] Forcato C, Argibay PF, Pedreira ME, Maldonado H. Human reconsolidation [140] Kornell N, Bjork RA. Learning concepts and categories: is spacing the “enemy
does not always occur when a memory is retrieved: the relevance of the of induction”. Psychol Sci 2008;19(6):585–92.
reminder structure. Neurobiol Learn Mem 2009;91(1):50–7. [141] Karpicke JD, Butler AC, Roediger III HL. Metacognitive strategies in student
[109] Squire LR, Slater PC, Chace PM. Reactivation of recent or remote memory learning: do students practise retrieval when they study on their own?
before electroconvulsive therapy does not produce retrograde amnesia. Memory 2009;17(4):471–9.
Behav Biol 1976;18(3):335–43. [142] Kornell N, Bjork RA. The promise and perils of self-regulated study. Psychon
[110] Phelps EA, Schiller D. Chapter 9 – reconsolidation in humans. In: Alberini CM, Bull Rev 2007;14(2):219–24.
editor. Memory reconsolidation. San Diego: Academic Press; 2013. p. 185–211. [143] Kornell N, Son LK. Learners' choices and beliefs about self-testing. Memory
[111] Sederberg PB, Gershman SJ, Polyn SM, Norman KA. Human memory recon- 2009;17(5):493–501.
solidation can be explained using the temporal context model. Psychon Bull [144] Anki. http://ankisrs.net.
Rev 2011;18(3):455–68. [145] Lindsey RV, Shroyer JD, Pashler H, Mozer MC. Improving students' long-term
[112] Ebbinghaus H. Memory: a contribution to experimental psychology. Oxford, knowledge retention through personalized review. Psychol Sci 2014;25
England: Dover; 1964. (3):639–47.