The Hippocampus and Memory Integration:

Building Knowledge to Navigate Future

Decisions

Margaret L. Schlichting and Alison R. Preston

Abstract Everyday behaviors require a high degree of flexibility, in which prior

knowledge is applied to inform behavior in new situations. Such flexibility is
thought to be supported in part by memory integration, a process whereby related
memories become interconnected in the brain through recruitment of overlapping
neuronal populations. Mechanistically, integration is thought to occur through
specialized hippocampal encoding processes that integrate related events during
learning. By recalling past events during new experiences, connections can be
created between newly formed and existing memories. The resulting integrated
memory traces would extend beyond direct experience in anticipation of future
judgments that require consideration of multiple learned events. Recent advances in
cognitive and behavioral neuroscience have provided empirical evidence for the
existence of such a mechanism, with hippocampal encoding mechanisms—in
coordination with medial prefrontal cortex—supporting memory integration.
Emerging research suggests that abstracted representations in medial prefrontal
cortex guide reactivation of related memories during new encoding events, thus
promoting hippocampal integration of related experiences. Moreover, recent work
indicates that integrated memories can impact a host of behaviors, from promoting
spatial navigation and imagination to resulting in memory distortion and deletion.

M.L. Schlichting
Department of Psychology, University of Toronto, 100 St. George St., Sidney Smith Hall,
Toronto, ON M5S 3G3, Canada
e-mail: [email protected]
A.R. Preston (*)
Center for Learning and Memory, The University of Texas at Austin, 1 University Station
C7000, Austin, TX 78712, USA
Department of Psychology, The University of Texas at Austin, 1 University Station C7000,
Austin, TX 78712, USA
Department of Neuroscience, The University of Texas at Austin, 1 University Station C7000,
Austin, TX 78712, USA
e-mail: [email protected]

© Springer International Publishing AG 2017 405

D.E. Hannula, M.C. Duff (eds.), The Hippocampus from Cells to Systems,
DOI 10.1007/978-3-319-50406-3_13
406 M.L. Schlichting and A.R. Preston

Introduction

Decades’ worth of research documents the involvement of the hippocampus in

rapidly encoding new episodes, which are then transferred (i.e., consolidated) to
neocortex over time. However, memory is a dynamic phenomenon. The once
widely accepted view that such consolidated memories are immune to modification
has long since been refuted. Consolidated memories may be reactivated during new
experiences, at which point they are susceptible to distortion, deletion, or updating
(Nadel and Hardt 2011; McKenzie and Eichenbaum 2011; Nadel et al. 2012).
Conversely, reactivated memories may also influence how new content is encoded
(Zeithamova et al. 2012a; Gershman et al. 2013). Here, we review the recent work
in cognitive and behavioral neuroscience that investigates the complex ways in
which memories influence one another and change over time. One way by which
such mutual influence may occur is through memory integration.
Memory integration refers to the idea that memories for related experiences are
stored as overlapping representations in the brain, forming memory networks that
span events and support the flexible extraction of novel information. For example,
imagine you see a woman walking her dog in the park near your house (Fig. 1).

Fig. 1 Schematic depiction of related events that might lead to memory integration and their
associated neural codes. One day while walking in the park, you encounter a woman and her dog
(initial experience, left). Connections are formed among a group of simultaneously activated
neurons, coding the woman–dog association (blue network). A few days later, you encounter the
same dog in town, this time with a man (overlapping event, right). The dog (overlapping element)
triggers reactivation of your initial experience in the park (woman–dog association). Such
reactivation enables connections to be formed among neural representations of the woman, dog,
and man, linking the related events across time (overlapping blue and yellow networks). The
resulting integrated memories are hypothesized to support novel judgments that require consider-
ation of both events; here, for instance, you may infer a relationship between the woman and the
man despite never having seen them together. Figure adapted with permission from Schlichting
and Preston (2015)
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 407

During this experience, you form a memory for the event that represents the
relationship among the woman, the dog, and the park. The next day, you see the
same dog out for a walk in town with a man. The familiar element (the dog) during
this second experience may serve as a cue for hippocampal pattern completion,
triggering the reactivation of your prior experience with the woman and dog. The
new event (the man walking the dog in town) is then encoded in the presence of the
reactivated information about your first experience with the dog. In this way, a link
between the woman, the man, and the dog can be formed during encoding, despite
the fact that you have never seen the woman with the man.
The notion that new encoding and prior knowledge interact with one another is
by no means new (Bartlett 1932; Tolman 1948; Cohen and Eichenbaum 1993); yet,
the neural mechanisms and behavioral implications of memory integration have
only recently become the subject of empirical investigation. The field’s growing
interest in understanding these complex, real-world aspects of episodic memory has
been realized thanks to the advent of elegant behavioral paradigms and advanced
analysis methods for neural data. We first review evidence for the neural mecha-
nisms that underlie memory integration. We then turn to a discussion of the range of
behaviors that might be supported by integration, from flexible navigation to
imagination and creativity. Finally, we set forth questions and considerations for
future research.

Neural Mechanisms

Memory integration has been studied in both rodents and humans using highly
controlled experimental paradigms in which subjects make decisions that span
multiple learned experiences. In one example task, the associative inference para-
digm (Preston et al. 2004), participants encode a series of overlapping events: AB
followed later by BC, where ‘AB’ denotes a studied arbitrary association between
items A and B. Participants are later tested on their memory for directly experi-
enced information (AB, BC) as well as on their ability to make novel inferences
(AC) that require consideration of two events. In this task, performance on the AC
inference test serves as the critical behavioral index of memory integration. By
recalling past (AB) events during new (BC) experiences, knowledge structures are
formed that integrate the newly learned information into prior memories (Fig. 1).
The resulting integrated memories would allow for direct extraction of novel
inferences that cross event boundaries, thereby promoting performance on the AC
test. In addition to a behavioral index of integration, such experimental designs
allow researchers to index the neural processes specific to encoding of the second
(BC) overlapping event, during which there is a unique opportunity to integrate
across related memories. A number of similar paradigms have been used in the
literature (Eichenbaum et al. 1996; Shohamy and Wagner 2008), all of which
require participants to make novel decisions spanning learned pieces of informa-
tion. For simplicity, in this section we refer to behaviors thought to index memory
408 M.L. Schlichting and A.R. Preston

integration as integration behaviors; see section “Implications for Behavior” for a

detailed discussion of the diverse set of behaviors potentially impacted by memory
integration. In section “Mechanism Overview”, we provide an overview of the
neural mechanisms underlying memory integration. We then describe examples of
empirical evidence for these processes in sections “Evidence for a Hippocampal
Role in Integration” and “Hippocampal-Medial Prefrontal Interactions”.

Mechanism Overview

Human and animal lesion work highlights the critical role of the hippocampus and
an interconnected structure, the medial prefrontal cortex (Iordanova et al. 2007;
DeVito et al. 2010b; Koscik and Tranel 2012; Ghosh et al. 2014; Warren et al.
2014), in memory integration (Fig. 2). Damage to either of these structures impairs
the ability to combine information acquired during different episodes despite intact
memory for the previously learned individual events (Bunsey and Eichenbaum
1996; DeVito et al. 2010b; Koscik and Tranel 2012). Work in rodents also demon-
strates dynamic interactions between these structures during memory updating,
perhaps reflecting the flow of information from hippocampus to MPFC (Tse et al.

Fig. 2 Proposed roles of human hippocampus and MPFC in memory integration. Locations and
hypothesized functions of regions critical for memory integration in the human brain. Purple,
hippocampus; green, medial prefrontal cortex. Here, we intentionally provide a broad definition of
MPFC due to high variability in the precise location of effects reported across studies. For
instance, we include anterior cingulate cortex, which has been implicated in memory integration
(Wang et al. 2012) and the formation of memory models (Roy et al. 2012). Inset, cross section
through the hippocampus (purple) highlighting area CA1 (dark purple portion). Approximate
hippocampal subfield boundaries are indicated with thin dashed lines. Location of cross section
along hippocampal axis is indicated with a thick dashed line. MPFC medial prefrontal cortex, CA1
Cornu ammonis field 1, DG/CA2,3 dentate gyrus and Cornu ammonis fields 2 and 3, SUB
subiculum. Figure adapted with permission from Schlichting and Preston (2015)
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 409

2011). However, while these data underscore the importance of hippocampus and
MPFC in memory integration, the precise mechanisms by which these regions
contribute have only recently started to become clear.
One period during which memory integration may take place is when new
learning experiences share content (e.g., a person, place, or thing) with existing
memory traces (Fig. 1). Through this process, termed integrative encoding, mem-
ories are formed that integrate information across distinct experiences (Nadel and
Hardt 2011; McKenzie and Eichenbaum 2011; Nadel et al. 2012) in anticipation of
future use. This constructive, or prospective, nature of memory (Klein et al. 2002b;
Buckner 2010; Addis and Schacter 2012) dates back to Tolman’s concept of a
“cognitive map” (Tolman 1948) and is reflected in modern memory theories
including relational memory theory (Eichenbaum et al. 1999), multiple trace theory
(Nadel and Moscovitch 1997), and schema theory (Bartlett 1932; van Kesteren
et al. 2012). Memory integration has been proposed as a key mechanism underlying
a host of flexible behaviors, including inferring novel relationships (O’Reilly and
Rudy 2001; Shohamy and Wagner 2008; Zeithamova and Preston 2010;
Zeithamova et al. 2012a), determining new routes through the environment
(Gupta et al. 2010), and making adaptive decisions (Wimmer and Shohamy
2012). These ideas are also highly related to the influential temporal context
model (Kahana 1996; Howard et al. 2005), in which items are bound to the learning
context in which they occur. In this case, learning context may include related
content that has been reactivated.
When new event relates to prior experience, pattern completion mechanisms
supported by the hippocampus reactivate the previously stored, overlapping mem-
ory (Zeithamova et al. 2012b; Preston and Eichenbaum 2013). Empirical support
for reactivation of prior memories during overlapping learning experiences has
recently been garnered using neural decoding of fMRI data (Fig. 3) (Kuhl et al.
2012; Zeithamova et al. 2012a; Gershman et al. 2013). With the related content
reinstated in the brain, hippocampal area CA1 (Fig. 2) is thought to compare prior
memories with incoming information from the environment (Hasselmo and Schnell
1994). CA1 may signal the presence of associative novelty (i.e., when new experi-
ences violate memory-based predictions) and facilitate new encoding (Hasselmo
et al. 1996; Larkin et al. 2014). Models of hippocampal subfield function have
suggested that CA1 novelty signals may influence neural dynamics via feedback
connections to the medial septum, modulating acetylcholine levels and setting
appropriate dynamics for learning (i.e., encoding rather than retrieval; Hasselmo
and Schnell 1994; Hasselmo et al. 1996). The resulting integrated memories are
highly structured, with shared elements coded similarly across experiences
(McKenzie et al. 2013, 2014). One recent study (McKenzie et al. 2014) has
shown that hippocampal CA field firing patterns for overlapping events reflect a
hierarchy of features coded according to their behavioral relevance. This organiza-
tion scheme could then be exploited to extract commonalities across episodes and
support a host of behaviors, as discussed below.
Hippocampal mechanisms may be additionally influenced by operations in
MPFC. While its specific role in memory is only starting to be uncovered, at least
410 M.L. Schlichting and A.R. Preston

Fig. 3 Example use of neural decoding to quantify memory reactivation. Left panel, depiction of a
neural decoding approach quantifying the degree of memory reactivation during learning. The
neural pattern evoked during the overlapping event (man-dog from Fig. 1) is hypothesized to
reflect reinstatement of the related—but not presently viewed—element (the woman). The fMRI
signal is extracted for each voxel in a region of interest (here, ventral temporal cortex is used as an
example). This information is then input into a neural decoder trained to recognize activation
patterns associated with different kinds of stimuli (e.g., faces). Based on the weights for each voxel
learned during training, the decoder outputs a value reflecting the degree to which the neural
pattern reflects reactivation of the related versus unrelated content. These evidence scores can then
be used as an index of reactivation. Right panel, reactivation during encoding of overlapping
events predicts later flexible inference (woman-man association), a behavioral index of memory
integration. Data are adapted with permission from Zeithamova et al. (2012a). Figure adapted with
permission from Schlichting and Preston (2015)

two functions have been proposed for MPFC that are of relevance to the present
discussion. First, MPFC is thought to represent mental models that guide behavior
across a number of domains (Roy et al. 2012; Wilson et al. 2014). With regards to
memory, some suggest that MPFC encodes interconnected information to form
mental models based on mnemonic content (i.e., memory models) (Schacter et al.
2012; St. Jacques et al. 2013), which may include features such as behavioral
relevance and appropriate response associated with a particular context (Miller
and Cohen 2001; Euston et al. 2012; Kroes and Fernández 2012). This functionality
may explain the involvement of MPFC in reinforcement learning, which has been
hypothesized to reflect its coding of action-outcome associations. Anatomical
features of MPFC may make it especially well suited to form such complex
representations of goals or task rules, as it receives a broad range of input from
sensory and limbic regions (Price and Drevets 2009).
A second possible function of MPFC is in biasing learning-phase retrieval
toward the most behaviorally relevant memories, thereby influencing what will be
integrated (van Kesteren et al. 2012; Kroes and Fernández 2012; Preston and
Eichenbaum 2013). This may be conceptualized as the deployment of memory
models to resolve conflict among related experiences. Memory models are thought
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 411

to be activated when incoming information relates to existing knowledge. MPFC

may then select specific task-relevant memories for reactivation (van Kesteren et al.
2012; Kroes and Fernández 2012; Wilson et al. 2014), perhaps via white matter
projections to the medial temporal lobe (MTL) cortical structures that provide the
major input to hippocampus (Cavada et al. 2000). Hippocampus may then bind
reactivated content to current experience, resulting in an integrated trace. Following
integration in hippocampus, memory models may be updated with new content as
needed through direct hippocampal inputs to MPFC (van Kesteren et al. 2012).
Through this process, MPFC may come to represent integrated memories that have
been abstracted away from individual episodes (i.e., schema) over time (van
Kesteren et al. 2012; Richards et al. 2014).
Of course, the possibilities we describe here are neither exhaustive nor mutually
exclusive. Future research will be needed to fully understand the role of MPFC in
memory integration, and assess whether its functionality might differ across
subregions.

Evidence for a Hippocampal Role in Integration

Electrophysiological studies in rodents have shown hippocampal-mediated replay

of prior event sequences in new spatial contexts (Karlsson and Frank 2009) and
never-experienced spatial trajectories that represent a shortcut through a well-
learned environment (Gupta et al. 2010; although note that novel routes represented
a very small proportion of all replay events), consistent with the idea that memories
extend beyond direct experience. Furthermore, in environments with overlapping
elements, individual hippocampal neurons demonstrate experience-dependent gen-
eralized firing patterns that respond in multiple similar locations (Singer et al. 2010;
McKenzie et al. 2013) or to the overlapping features themselves (Wood et al. 1999).
Such generalized firing patterns suggest that hippocampal neurons develop repre-
sentations that code the similarities between events. By representing features
common to multiple events similarly, hippocampal codes can capture regularities
shared across different experiences and, in doing so, may act as “nodes” that link
distinct behavioral episodes (Fig. 1) (Eichenbaum et al. 1999).
Behavioral work in humans suggests that reactivating related memories imme-
diately prior to a new learning experience increases the likelihood that new content
will be integrated into existing memories (Hupbach et al. 2007). Using neuroim-
aging, researchers have also related the degree of reactivation of prior experience
during encoding of new overlapping events to evidence for integration (Kuhl et al.
2010; Zeithamova et al. 2012a). In one study, the evidence for hippocampus-
mediated reactivation of prior memories was associated with greater retention of
the reactivated information (Kuhl et al. 2010), demonstrating that reactivating
memories during new learning helps reduce forgetting of past events. In an asso-
ciative inference paradigm, another study demonstrated that reactivation of existing
knowledge during new learning of overlapping associations predicted superior
412 M.L. Schlichting and A.R. Preston

integration behavior, suggesting that combining related memories during learning

might underlie successful inferential reasoning (Fig. 3; Zeithamova et al. 2012a; see
also Richter et al. 2015).
With relevant prior experience reactivated in the brain, hippocampus is then
thought to bind or integrate current and prior experience (Shohamy and Wagner
2008; Zeithamova and Preston 2010; Zeithamova et al. 2012a). In one study
(Shohamy and Wagner 2008), increases in hippocampal activation across the
learning phase were associated with individual differences in integration behavior,
even when accounting for performance differences on trained associations
(Fig. 4a). Changes in hippocampal activation over learning in the associative
inference task were also related to integration behavior across participants, even
when accounting for differences in memory for single events (Zeithamova et al.
2012a). Moreover, interrogation of trial-by-trial neural engagement revealed that
hippocampal activation during encoding of overlapping associations (BC), but not
initially acquired associations (AB), differentiated between subsequently correct
and incorrect inference judgments (AC; Fig. 4b-i) (Zeithamova et al. 2012a).
Collectively, these findings highlight the importance of a hippocampal encoding
mechanism whereby overlapping experiences are integrated into a network of
related memories as they are learned.

Fig. 4 Learning-phase integration signatures in hippocampus. (a) Activation in both left (not
shown) and right hippocampus during encoding of overlapping associations was correlated with
individual differences in inference performance. Specifically, increases in hippocampal activation
from the early to late portion of the training phase were associated with superior performance on
inferential probe trials. (b-i) In an associative inference task, right hippocampal activation during
encoding of overlapping associations (BC) was greater for trials in which the corresponding
inference judgment (AC) was later correct relative to trials on which the inference judgment
was later incorrect. Hippocampal activation during initially acquired associations (AB) was not
related to subsequent inferential performance. (ii) Using high-resolution fMRI, this signature was
isolated to the CA1 subfield of the hippocampus. Data are adapted with permission from: Shohamy
and Wagner (2008) (panel a), Zeithamova and Preston (2010) (panel b-i), and Schlichting et al.
(2014) (panel b-ii)
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 413

Integration is thought to be triggered by hippocampal comparator processes,

with hippocampal area CA1 signaling deviations between current events and
reactivated content (i.e., associative novelty signaling; Lisman and Grace 2005).
Empirical work in humans has supported the notion that area CA1 signals devia-
tions from prior experience, with engagement of this region increasing as the
number of changes to a studied stimulus increase (i.e., with an increasing degree
of mismatch; Duncan et al. 2012a). Activation in human CA1 during the encoding
of events that overlap with prior experiences has been shown to relate to a
behavioral measure of memory integration (Fig. 4b-ii; Schlichting et al. 2014),
consistent with the notion that novelty signals triggers the formation of links
between new content and prior memories.

Hippocampal-Medial Prefrontal Interactions

Recent evidence suggests that hippocampus interacts with MPFC to support mem-
ory integration in many circumstances (Fig. 2). One possible explanation for this
region’s involvement in encoding-phase memory updating lies in its pattern of
anatomical connectivity: MPFC is directly connected to the hippocampus, receiv-
ing inputs primarily from the anterior portion of CA1 (Barbas and Blatt 1995;
Cavada et al. 2000). MPFC also has extensive connections with a diverse set of
sensory, limbic, and subcortical structures (Cavada et al. 2000), suggesting that it
might be important for combining across episodic memories, represented in the
brain across distributed cortical and subcortical networks. Consistent with this idea,
recent studies have observed encoding-phase engagement (Zeithamova et al.
2012a) and evidence for reactivation of prior memories in MPFC (Richter et al.
2015), demonstrating the importance of this region for memory integration during
encoding. Moreover, enhanced functional coupling of hippocampus and MPFC has
been shown when new learning can be integrated into prior knowledge (Schlichting
and Preston 2016), consistent with the notion that MPFC interacts with hippocam-
pus to promote integration. Integration behavior has also been linked to individual
differences in the intrinsic functional connectivity (Gerraty et al. 2014) and struc-
tural connectivity (Schlichting and Preston 2016) of hippocampus and MPFC,
highlighting that even static neural characteristics might render some individuals
better suited for combining across related events.

Learning Factors Promoting Integration

A number of studies have investigated the learning factors that influence integra-
tion. For instance, while there is evidence that integration can occur in the absence
of conscious awareness (Shohamy and Wagner 2008; Wimmer and Shohamy 2012;
Henke et al. 2013; Munnelly and Dymond 2014), studies have shown that
414 M.L. Schlichting and A.R. Preston

integration may be facilitated when subjects become aware of the task structure
(either via instructional manipulations or spontaneously) (Kumaran and Melo 2013;
Richter et al. 2015). In fact, one experiment (Kumaran and Melo 2013) demon-
strated that such knowledge specifically benefitted judgments that spanned episodes
with no effect on memory for the individual episodes themselves, suggesting that
integration does not necessarily emerge with learning of the underlying experi-
ences. One possibility is that awareness constrains MPFC control processes, which
in turn biases hippocampal reactivation during learning toward task-relevant mem-
ories, allowing for integration across events.
It has been hypothesized that being reminded of related memories prior to a new
learning experience also increases the likelihood of integration, as the reactivated
memories become labile and readily updated. Consistent with this idea, behavioral
work in humans (Hupbach et al. 2007) found more intrusions from a second learned
list (List 2) when recalling the initial list (List 1) if participants had been reminded
of List 1 before encoding List 2. This finding was recently replicated in rodents
using “lists” of ordered feeder locations (Jones et al. 2012), with animals who
learned two lists in the same relative to different spatial contexts producing more
intrusions. Another study manipulated the degree of retrieval on a trial-by-trial
basis within participants (Duncan et al. 2012b). That study similarly found superior
integration performance for learning experiences that followed an old item (i.e.,
when retrieval was possible) versus those that followed a new item (when retrieval
was not possible). These findings are consistent with the proposal that integration
occurs via reactivation of prior memories; this work further highlights that
reminding the learner of the prior related memory may encourage integration.
The strength of existing memories may be an additional factor mediating
integration. In particular, stronger memories might be more readily reactivated
during learning, thereby allowing for integration across memories. One neuroim-
aging study showed that offline processing of initial memories was associated with
more evidence for reactivation and superior integration behavior during a subse-
quent learning experience, suggesting that memory strengthening during rest facil-
itates integration (Schlichting and Preston 2014). Integration signatures have also
been preferentially observed when initial memories are well-learned at the time of
the first overlapping event, as is the case in blocked learning (i.e., multiple AB
learning opportunities occurring before any BC learning; Schlichting et al. 2015).
These results suggest that strong prior memories may promote reactivation during
learning, thereby allowing for integration across memories. This work underscores
that integration may be especially likely when initial memories are well established
prior to new learning.
Other factors hypothesized to impact integration include (1) the nature of the
underlying memory representations, with more distributed as opposed to localized
representations proposed to promote integration (Schiller and Phelps 2011); and
(2) the degree of competition between new content and prior memories (i.e.,
whether or not the two memories can coexist), with integration preferentially
occurring in cases when competition is minimal (Hupbach 2011).
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 415

Offline Processes Promoting Integration

Numerous empirical studies (Tambini et al. 2010; Jadhav et al. 2012; Deuker et al.
2013; Staresina et al. 2013) and theoretical accounts (Marr 1970; McClelland et al.
1995) highlight the importance of offline processes—such as reinstatement of
recent experience and enhanced interregional communication—for episodic mem-
ory. It has been proposed that through hippocampal-neocortical interactions
(McClelland et al. 1995; Nadel et al. 2000), memories are reactivated during
periods of sleep and awake rest. Such reactivation (or replay) is thought to support
the strengthening and transfer of memory traces from the hippocampus to neocor-
tical regions for long-term storage (i.e., consolidation).
These mechanisms may also support the integration of memories across expe-
riences (Kumaran and McClelland 2012). Recent theories suggest that
hippocampus-mediated replay of event sequences during sleep (Hoffman and
McNaughton 2002; Ji and Wilson 2007) provides a potential mechanism for
constructing networks of related memories that anticipate future decisions and
actions (Sara 2010; Diekelmann and Born 2010; Lewis and Durrant 2011)—a
process referred to as prospective consolidation (Buckner 2010). Such theories
propose that by reactivating memories during sleep, representations are recombined
and recoded, resulting in rich networks of related memories that extend beyond
initially encoded events (Kumaran and McClelland 2012). This process is thought
to promote both the integration of new information into existing memories and
abstraction across episodes in neocortical regions, particularly MPFC (Lewis and
Durrant 2011). According to this view, stored memories are not veridical represen-
tations of events, but rather derived representations formed in anticipation of future
use. Sleep-based replay of hippocampal memory traces, therefore, could enhance
integration behaviors that tap knowledge about the relationships among events
experienced at different times (Ellenbogen et al. 2007; Werchan and Gómez
2013; Coutanche et al. 2013). Consistent with this notion, one study (Ellenbogen
et al. 2007) demonstrated that participants who slept following learning showed
better integration behavior relative to a comparison group who remained awake.
In addition to sleep-based mechanisms that might promote integration, offline
processes occurring during periods of awake rest have also been suggested to be
important for memory. The mnemonic consequences of reactivation of recent
experience has been demonstrated during awake rest using neurophysiological
techniques in rodents (Jadhav et al. 2012) and, more recently, in humans using
pattern information analysis of fMRI data (Deuker et al. 2013; Staresina et al.
2013). For instance, more delay period reactivation was observed for stimuli that
were remembered relative to those that were forgotten in a subsequent test
(Staresina et al. 2013). Moreover, studies have shown that the degree of
hippocampal-neocortical functional coupling during rest periods following learning
relates to later memory for the learned content (Tambini et al. 2010).
Recent evidence suggests that similar rest-phase mechanisms may promote the
integration of memories that span related events (van Kesteren et al. 2010; Craig
416 M.L. Schlichting and A.R. Preston

et al. 2015; Schlichting and Preston 2016), with integration-related neural signa-
tures persisting into offline periods following encoding. For example, one study
showed increased hippocampal-MPFC functional coupling during encoding condi-
tions that necessitate schema reorganization and updating; interestingly, this pattern
persisted during the post-encoding rest period (van Kesteren et al. 2010). These
findings are consistent with the idea that neural patterns evoked during encoding are
reactivated during offline rest periods, potentially reflecting early-phase consolida-
tion mechanisms. Similar neural signatures have been reported following memory
updating in an associative inference task, with the degree of hippocampal-MPFC
connectivity enhancements during awake rest following an integration opportunity
predicting individual differences in behavior (Schlichting and Preston 2016). While
the precise effect of these rest-phase processes for memory integration is yet to be
determined, it may be the case that memory reactivation and increased interregional
coupling may strengthen connections among related memories, thereby further
promoting the formation of integrated memory representations.

Neural Representations

Initial research suggests that one way in which the hippocampus supports behav-
ioral flexibility is by integrating information across multiple experiences to estab-
lish links between related events, either during new learning or offline through
replay of related experiences during sleep and rest. However, questions remain
regarding the precise nature of the underlying hippocampal representations. Several
theoretical and computational frameworks have proposed alternate accounts of the
properties of memory representations that can support inference, which we
describe here.
One hypothesized representational structure supporting integration is one in
which new events are incorporated into existing memory traces to be parsimoni-
ously represented in a single, composite memory representation (Fig. 5a). For
instance, consider the simple example of two events that share a common element
(AB, BC) as in the associative inference paradigm. When a new event occurs that
contains an element overlapping with a previous event (e.g., BC after encoding
AB), the overlapping element (B) can trigger pattern completion of the previously
encoded memory (AB). According to this hypothesized representational structure,
elements from the new, overlapping event (in this case, C) would be encoded into
the existing, reactivated memory (AB) to form a single integrated representation
that combines the two experiences (ABC). Because these integrated representations
directly code the novel relationship between A and C along with the original
experiences, this representational format provides a basis for the inferential use of
memory, but has a notable cost in that details of the individual experiences may not
be preserved (e.g., the knowledge that A and C were presented in two different
temporal contexts).
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 417

Fig. 5 Schematic depiction of alternative accounts of hippocampal representation in memory

integration. Representations of overlapping events (here AB, BC in the associative inference
paradigm) are shown using a simplified two-layer architecture. The bottom layer contains units
for each event element; the top layer contains hypothesized patterns of hippocampal representa-
tion. (a) Single integrated representation for overlapping events. According to this hypothesized
structure, new, overlapping event elements (C) are encoded into an existing, reactivated memory
(AB) to form a single composite representation for the two related associations. (b) Pattern
separated representations of individual events. In this view, a new event (BC) with partial overlap
to a previous memory (AB) would recruit a distinct hippocampal representation that preserves the
details of each individual experience. Links between the common element (B) and each of the
individual experiences could be used to mediate inference at encoding or retrieval. (c) Relational
representation of overlapping events. In this framework, separate representations are maintained
for overlapping events (AB, BC) and direct links between those events (at the level of the
hippocampus) code their relationship to one another. (d) Evidence for dissociable coding schemes
for indirectly related (A, C) items in an associative inference task across the anterior-posterior axis
of hippocampus. While posterior hippocampus showed that A and C items became more dissimilar
following overlapping encoding (blue cluster), anterior hippocampus coded A and C items more
similarly, particularly when memories were strong (green cluster). These findings suggest simul-
taneous separated and integrated representation of overlapping memories in posterior and anterior
hippocampus, respectively. Panels (a–c) are as originally published in Zeithamova D, Schlichting
ML, Preston AR (2012) The hippocampus and inferential reasoning: building memories to
navigate future decisions. Frontiers in Human Neuroscience 6:70. doi:10.3389/
fnhum.2012.00070. Panel (d) is adapted from Schlichting et al. (2015)
418 M.L. Schlichting and A.R. Preston

The influential cognitive map theory (Tolman 1948; O’Keefe and Nadel 1978)—
which first sparked interest in the flexible functions of the hippocampus—implicitly
assumes such integrated representations. In the context of this theoretical frame-
work, memory traces for newly learned individual events (i.e., recently traveled
routes) are combined with memories of previously traveled routes to allow for the
creation of an integrated map of the environment, including information about
paths not traveled. As a cognitive map of an environment becomes established, it
can be reactivated when an animal enters the same environment at a later point and
updated with new experiences in that environment. When familiar routes to a goal
are blocked, the cognitive map will enable navigation to the goal via an alternate
route because information about this novel (i.e., never before traveled) route is
included in a single representational structure of the environment.
In the context of non-spatial integration tasks, there is some evidence to support
this hypothesized ABC representational structure. For example, one study showed
that successful participants perform as quickly on integration judgments as on
explicitly trained associations (Shohamy and Wagner 2008), suggesting similar
representations for both directly learned and inferential associations. Moreover,
informal assessment suggested that the majority of participants in this study failed
to recognize the inferential probes as novel combinations of items, perhaps indi-
cating that some contextual details of original experiences were lost. Returning to
the dog-walking example, you may remember that the woman and man are a couple
with a dog, but may not remember specific details about how you first encountered
them. Future studies may provide a more detailed account of the circumstances
under which memory for original experience may become degraded.
The loss of experiential detail is a significant downside to the single, composite
representational structure linking elements of discrete events. Other computational
perspectives propose a different representational structure for hippocampus, with
pattern separation processes preserving distinct individual experiences and recur-
rent connections between the element and event representations allowing inference
across experiences (Fig. 5b; Kumaran and McClelland 2012; McClelland et al.
1995). In our example, this representational structure would predict that a new
event partially overlapping with a previous event (i.e., BC) would recruit a different
hippocampal representation to make it distinct from the originally experienced
event (AB). The two events would be linked through their individual connections
to the shared event element (B). Because of the recurrent connections between
individual element and event representations (ascribed to entorhinal cortex and
hippocampus, respectively), such a hypothesized structure allows for preservation
of event details while also supporting inferential judgments about the relationship
between experiences. For example, when presented with a novel inferential probe
(AC), each individual element (A and C) may serve as a partial cue leading to the
reactivation of the originally experienced events (AB and BC). Activation of the
common item (B) in both cases would lead to successful inference.
Results showing unique hippocampal responses during integration behavior
itself (Preston et al. 2004; Zalesak and Heckers 2009; Zeithamova and Preston
2010) might reflect the use of such pattern separated inputs to support performance.
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 419

This representational structure can also explain recruitment of the hippocampus

during encoding of overlapping events (Shohamy and Wagner 2008; Zeithamova
and Preston 2010), which potentially reflects changes in the weights linking com-
mon elements to the individually experienced events. It is important to note that
even such pattern-separated representations would be expected to change over time
and become more generalized as follows (Kumaran and McClelland 2012).
Reactivation of these memory representations during the consolidation process or
during offline replay would result in more frequent reactivation of common ele-
ments and strengthening of their connections to event representations. In contrast,
idiosyncratic elements unique to individual events would be reactivated less fre-
quently and gradually lose their connections to event representations (Lewis and
Durrant 2011). This process would lead to the gradual loss of episodic details in
favor of abstracted representations that capture regularities across experiences
(McClelland et al. 1995).
An alternate view that combines elements of both of these frameworks stems
from relational memory theory (Cohen and Eichenbaum 1993). Relational memory
theory proposes that the hippocampus maintains representations of individual
events while also directly encoding relationships between separate experiences
(Eichenbaum et al. 1999). In our symbolic representation of this theory, different
hippocampal units are recruited to represent individual events, but a lateral con-
nection exists at the second level, linking the representations of overlapping events
together (Fig. 5c). Both pattern separation and pattern completion at the level of the
hippocampus would contribute to the formation of such networks of related mem-
ories. For example, a new overlapping event (BC) would recruit a hippocampal
representation distinct from the originally experienced event (AB). Simultaneously,
the overlapping element (B) serves as a partial cue that reactivates the prior event
(AB). Based on a Hebbian learning rule, the connection between the two hippo-
campal memory traces would be strengthened and an explicit link between the
overlapping events would be formed. Like the representational structure above,
such relational networks would support mnemonic inference while simultaneously
preserving memory for individual experiences.
Different coding strategies may be preferred across subregions of the hippocam-
pus. Prior work has implicated anterior hippocampus in processing relational
information (Schacter and Wagner 1999; Kirwan and Stark 2004; Chua et al.
2007) and combining information across episodes (Preston et al. 2004; Addis
et al. 2007; Barron et al. 2013), typically on the basis of activation enhancements
during tasks that require consideration of multiple episodes. Mechanistically,
anterior hippocampus might form generalized representations promoting behav-
ioral flexibility using its broad place fields (Poppenk et al. 2013; Preston and
Eichenbaum 2013; Strange et al. 2014). In contrast, posterior hippocampus, with
its more finely tuned place fields, is thought to code event specifics. Consistent with
this notion, rodent work has shown that while anterior hippocampal neurons
respond similarly across related episodes, posterior hippocampal firing patterns
are event-specific (Komorowski et al. 2013). Moreover, the ability to retrieve
details has been differentially related to hippocampal volumes across the long
420 M.L. Schlichting and A.R. Preston

axis, with smaller anterior and larger posterior regions being associated with
superior recollection across individuals (Poppenk et al. 2013). These findings and
others (Demaster et al. 2013) suggest dissociable functions along the hippocampal
anterior-posterior axis in humans, with anterior generalizing across events and
posterior representing event details (Poppenk et al. 2013). Anterior hippocampus
also shares the strongest anatomical connections with MPFC (Barbas and Blatt
1995), making it a good candidate region for integrating across related experiences.
Despite the prominence of these theories, empirical evidence as to how elements
of overlapping events are coded in human hippocampus has been demonstrated
only recently (Collin et al. 2015; Schlichting et al. 2015). One study (Fig. 5d;
Schlichting et al. 2015) scanned participants during viewing of individual items
both before and after encoding of overlapping (AB, BC) associations to quantify
how the representations of individual memory elements shift as a function of
learning. In anterior hippocampus, indirectly related (A and C) items became
more similar to one another following learning, consistent with integration across
the related AB and BC events. In contrast, indirectly related items became more
dissimilar in posterior hippocampus, suggesting separation of the overlapping
events in this region. Similar findings were reported in another fMRI study (Collin
et al. 2015) using a paradigm involving related events that could be combined to
form narratives. Results revealed a gradient in the granularity of memory represen-
tations across the anterior-posterior axis of hippocampus, with individual events
(small scale network) coded in posterior hippocampus and indirect relationships
among related events (large scale network) represented only in anterior hippocam-
pus. Neural codes also related to behavior, with only participants showing behav-
ioral evidence of integration demonstrating a gradient in memory representation
granularity. The results of both studies demonstrate that there are important
regional differences in neural codes that allow for the simultaneous representation
of integrated and separated memories within the hippocampus.

Implications for Behavior

Forming memories that integrate across related episodes is thought to confer a

degree of mnemonic flexibility. For instance, by coding the relationships that span
events, memories may be formed in anticipation of future decisions. In this section,
we discuss the behavioral implications of memory integration across a number of
cognitive domains. In section “Behavioral Benefits”, we focus on the various
benefits conferred by integrated memories on behavior. However, memory integra-
tion may also yield undesirable mnemonic consequences, which we describe in
section “Behavioral Consequences”.
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 421

Behavioral Benefits

Inferring Relationships

Inference is typically conceptualized as a logical, effortful process by which

multiple memories are recombined to make a novel decision. In line with this
intuition, initial studies of inference focused on hippocampal contributions to
successful inference at the time of retrieval (Heckers et al. 2004; Preston et al.
2004; Zalesak and Heckers 2009; DeVito et al. 2010a). More recently, however,
research attention has turned to the specialized hippocampal encoding mechanisms
supporting the formation of integrated memories well suited to later decisions.
Integrated memories may facilitate a host of novel judgments that require knowl-
edge of the relationships among events, such as in associative inference, transitive
inference, and acquired equivalence paradigms (Fig. 6; Zeithamova et al. 2012b;
c.f. Kumaran 2012). These judgments tap memory flexibility, requiring participants
to make novel inferences on the basis of trained associations; for simplicity, we
group these behaviors under the term “inference.” Because integrated memories
code for the relationships among learned associations (Fig. 1), they may be rein-
stated and the new information directly extracted during an inference judgment
itself (Shohamy and Wagner 2008).
Empirical evidence using neural decoding of cognitive states has demonstrated
that an integration state can be differentiated from both pure retrieval and pure
encoding states (Richter et al. 2015), suggesting that integration is neurally distinct
from its underlying components. Evidence for an integration state in that study also
predicted performance on the subsequent inference test both within and across
participants, demonstrating that fluctuations in learning-phase integration impact
subsequent behavioral flexibility. Recent work has also directly linked the degree of
neural evidence for learning-phase reactivation of related memories to subsequent
behavior (Kuhl et al. 2010; Zeithamova et al. 2012a; Richter et al. 2015). For
instance, the degree to which previously encoded content is reactivated during new
events has been shown to predict both subsequent memory for the reactivated
content itself (Kuhl et al. 2010) and later inference (Fig. 3; Zeithamova et al.
2012a; Richter et al. 2015), consistent with the notion that reactivation supports
memory strengthening and flexibility via integration. One study (Zeithamova et al.
2012a) also demonstrated that engagement of hippocampus and ventral MPFC
related to later inference performance. Moreover, that study observed functional
connectivity enhancements across learning repetitions, suggesting that memories
bound in hippocampus may come to depend on MPFC as they are integrated and
strengthened (Zeithamova et al. 2012a). Within the hippocampus, CA1 engagement
during overlapping events has been shown to predict subsequent inference
(Schlichting et al. 2014). The degree to which learning-phase CA1 patterns are
reinstated during inference has also been shown to relate to speed and accuracy,
consistent with ideas regarding this region’s role in integration (Schlichting et al.
2014).
422 M.L. Schlichting and A.R. Preston

Fig. 6 Inference tasks. (a) Transitive inference task with six elements. A set of overlapping
training pairs forms an ordered hierarchy of relationships. Participants learn each individual
training pair via feedback (e.g., A > B) and are then tested on novel inference and novel
non-inference judgments. Items in inferential probe trials may be separated by one element in
the hierarchy (e.g., B ? D, indicated as 1! ) or two elements (e.g., B ? E, indicated as 2! ). Novel
non-inferential probes test knowledge of the relationship between the end items of the hierarchy
(A ? F). (b) Acquired equivalence task. In stage one of training, participants are trained via
feedback to associate two faces (F1 and F2) with a particular scene (S1). In stage two, participants
learn to select a second scene (S2) when cued with one of the faces (F1). Inference is then measured
as the proportion of trials on which participants choose S2 when cued with F2. The schematic
depicts trained stimulus–response relationships (solid black arrows) and inferential relationships
(dashed black arrows). (c) Associative inference task. Participants learn an overlapping set of
associations (here, face–house associations), in which two stimuli (a man and a woman) are
associated with a common third item (a house). Novel inference trials evaluate knowledge for
the indirect relationship between items (who lives together in the same house). Figure as originally
published in Zeithamova D, Schlichting ML, Preston AR (2012) The hippocampus and inferential
reasoning: building memories to navigate future decisions. Frontiers in Human Neuroscience
6:70. doi:10.3389/fnhum.2012.00070

Recent work has also shown that inference is impaired in patients with lesions to
ventral MPFC (Koscik and Tranel 2012). Furthermore, novel inference judgments
are selectively facilitated following sleep (Ellenbogen et al. 2007; Werchan and
Gómez 2013), emphasizing the importance of offline processes in integration.
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 423

Spatial Navigation

Perhaps the most familiar and widely studied form of memory integration stems
from Tolman’s seminal work on cognitive maps (Tolman 1948). Tolman proposed
that navigation relies on the coherent representation of spatial layouts, which can
flexibly give rise to new inferences about the relative locations of landmarks in the
environment (Tolman 1948). One mechanism by which cognitive maps may be
formed is by representing both recent past and future experience in the hippocam-
pus at the same time. One rodent study demonstrated such simultaneous coding of
retrospective and prospective paths leading up to a choice point in a continuous
T-maze (Catanese et al. 2014), consistent with the notion that the hippocampus
forms an ongoing representation of space including both past and future routes (see
also e.g., Johnson and Redish 2007). Another recent report (Wu and Foster 2014)
suggests that rather than separately representing the temporal structure of multiple
traversed paths, the hippocampus integrates across overlapping routes to accurately
codes the overall spatial topology of the environment. This type of representational
scheme might support the ability to generate novel paths when, for instance, there is
an obstacle blocking a learned route.
Recent work in humans has demonstrated a relationship between hippocampal
volumes and the ability to infer novel spatial relationships among a set of trained
landmarks (Schinazi et al. 2013), consistent with the idea that the hippocampus
constructs integrated spatial maps. Behavioral studies have further found sleep-
(Coutanche et al. 2013) and rest- (Craig et al. 2015) related increases in spatial
relational inference performance. For instance, participants who passively rested
for 10 min following route learning through a virtual environment had better
memory for the spatial layout relative to participants who engaged in a 10-min
distractor task (Craig et al. 2015). Importantly, the memory test tapped the forma-
tion of a cognitive map by assessing knowledge of routes that had never been
directly experienced. Similar behavioral benefits have been reported in a group of
participants who slept relative to a group who remained awake (Coutanche et al.
2013), indicating that early phase consolidation processes engaged during offline
periods may facilitate the construction of cognitive maps.
Work in rodents demonstrates that the firing patterns of hippocampal CA1
neurons predict future routes (Pfeiffer and Foster 2013). In one study, hippocampal
trajectory events predicted rats’ immediate future behavior as they navigated to a
previously learned goal location in a familiar open arena. Trajectory events were
more consistent with future than with previously traveled routes (Pfeiffer and Foster
2013), suggesting a role for hippocampal processing in planning future navigation
through a familiar environment. Interestingly, trajectories can represent even novel
future paths (although this is rare; Gupta et al. 2010; Pfeiffer and Foster 2013),
suggesting that the hippocampus—perhaps guided by MPFC (de Bruin et al.
1994)—may support flexible navigation by simulating and evaluating possible
trajectories in the context of current goals.
424 M.L. Schlichting and A.R. Preston

It is of note that uncertainty remains in the literature about precisely how the
hippocampus encodes cognitive maps. For instance, it has been proposed that CA
cells code the transition among locations (“transition cells”). Thus, the relationship
between the memory integration mechanisms described here and the emergence of
a cognitive map remain unclear at this point, and will be an important avenue of
future investigations.

Mapping Social “Space”

The role of the hippocampus in integration is thought to be domain-general, with

recent work extending this idea into studying social relationships. For example, one
study (Kumaran et al. 2012) taught participants both social and non-social hierar-
chies in a transitive inference paradigm. Results showed that while fMRI activation
and volume of the amygdala was specifically related to performance on the social
hierarchy, the hippocampus represented the hierarchical structure for both social
and non-social scenarios. In another recent experiment (Tavares et al. 2015),
participants performed a role-playing task comprising a series of interactions with
fictional characters. Over the course of the experiment, characters moved across
social space due to changes in their power over and affiliation with the participant.
Hippocampal engagement was modulated by the position of the character in social
space, suggesting that the hippocampus codes for characters’ relative positions as a
function of their social attributes (i.e., power and affiliation). Hippocampal activa-
tion was also correlated with behavioral measures of social skills across partici-
pants, consistent with the notion that hippocampal representations of social space
may explain some variability in real-world social behaviors. Taken together, these
studies suggest that hippocampal integration mechanisms may aid us in forming a
cognitive map of social space.

Decision Making

Integrated memories may also influence non-mnemonic decision making. For

example, one recent fMRI study (Wimmer and Shohamy 2012) suggests that the
hippocampus supports the transfer of monetary value across related experiences
through additional recruitment of reward regions. Participants first learned a series
of arbitrary S1S2 associations. They then learned that half of the S2 stimuli predicted
a monetary reward (S2þ). During the critical decision phase, participants chose
between two S1 stimuli, only one of which was indirectly associated with a
monetary reward (S1þ) through its association with a rewarded S2. Value transfer
was operationalized as the tendency to choose S1þ over S1#; importantly, though
neither S1 stimulus had been directly associated with a reward, one was indirectly
predictive of monetary gain via S2þ. The researchers showed greater reactivation
of prior related knowledge during encoding of new reward information for stimuli
that showed more evidence of subsequent preference shifts toward S1þ.
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 425

Hippocampal-striatal functional coupling was also associated with value-related

preference changes (Wimmer and Shohamy 2012), suggesting that hippocampus
may interact with domain-specific regions (e.g., striatum in value learning tasks) in
service of integration.
Consistent with a domain-general role for hippocampus in memory integration,
rodent work (Blanquat et al. 2013) found that the hippocampus is necessary for
updating a known goal location with new value information. These updated mem-
ories may then be transferred to neocortex, as MPFC was necessary for retaining the
updated knowledge to support performance on the next day (Blanquat et al. 2013).
Thus, integrated memories incorporating value information may be maintained as
memory models in MPFC that will later bias behavior. We note that this role for
MPFC is likely also domain-general given its documented involvement in a number
of tasks lacking an explicit value component.

Schemas

Schemas are knowledge frameworks that capture regular patterns in the environ-
ment by abstracting information across experiences (Bartlett 1932) and represent
features common to multiple different events while discarding idiosyncratic details.
For example, a “restaurant schema” may contain commonly experienced elements
such as sitting at a table, ordering from a menu, and paying the bill, but not one-time
elements such as the waiter spilling water on you. We suggest that while the specific
paradigms typically used to study memory schema are quite different from the
associative learning tasks that are the focus of this chapter, these bodies of work
share important features and the behaviors may be supported by a common neural
mechanism. Like memory integration, building upon an existing knowledge struc-
ture (schema) to incorporate new information in particular has been shown to
involve both hippocampus and MPFC.
Schemas guide behavior by providing a set of expectations for a given experi-
ence. Like integrated memory representations, schemas also contain information
derived from multiple events that may support inferential decisions. Specifically,
schemas represent relationships among elements commonly associated with certain
types of situations, despite the fact that these elements have not necessarily been
experienced together. Moreover, encoding new events in the context of a
reactivated schema may provide an additional mechanism for inferential reasoning.
For example, a person may come to your table at the end of your meal and inquire
about the quality of the food and service. In the absence of an introduction, you may
infer that this person is the owner or manager of the restaurant because your
restaurant schema contains information about who is likely to ask for feedback
about your dining experience.
Recent attention has focused on the behavioral benefits conferred by memory
schema. For instance, research in rodents has demonstrated that reactivation of an
existing task schema (in this case, a well-learned spatial layout) allowed for rapid
acquisition of new flavor–place associations in a single trial (Tse et al. 2007, 2011).
426 M.L. Schlichting and A.R. Preston

Without an existing schema, such associative learning required repeated training

across multiple weeks. Importantly, rats with hippocampal lesions failed to show
facilitated learning of new information in the presence of reactivated schemas,
highlighting a critical role for this region in the rapid incorporation of new infor-
mation into existing knowledge frameworks. Echoing these results, a number of
human studies have reported behavioral benefits in learning and memory when new
information can be incorporated into an existing schema (Kumaran 2013; van
Kesteren et al. 2013, 2014).
Rodent (Tse et al. 2011) and human (van Kesteren et al. 2010, 2013, 2014) work
suggests that both MPFC and hippocampus are engaged during learning of schema-
related information (i.e., schema updating). Recent empirical data indicate that one
factor that may influence the relative engagement of MTL and MPFC is the degree
of consistency between new information and existing schema. Specifically, one
study (van Kesteren et al. 2013) demonstrated that MPFC engagement was more
predictive of subsequent memory for information congruent with existing schema,
perhaps reflecting direct encoding of new content into prior knowledge. Note that
this idea contrasts with standard views of consolidation, which propose that hippo-
campal memories are transferred to neocortex after long time periods; however,
recent work suggests the possibility of neocortical encoding of new information
independent of the hippocampus (Sharon et al. 2011; see however Smith et al. 2014;
Warren and Duff 2014). Conversely, MTL engagement was more predictive of
successful encoding of incongruent information. Application of a schema to a new
scenario has also been shown to primarily recruit hippocampus (Kumaran et al.
2009; de Hoz and Martin 2014). For example, one fMRI study (Kumaran et al.
2009) found that while engagement and connectivity of hippocampus and ventral
MPFC was enhanced during generation of a task schema, the application of schema
to guide behavior in a novel but similarly structured task selectively recruited
hippocampus.
One theory (van Kesteren et al. 2012) of schema-dependent learning suggests
that with increasing congruency, MPFC becomes increasingly able to bias
reactivation toward related memories. Increasing congruency would also be asso-
ciated with decreasing novelty, which may result in diminished reliance on hippo-
campal integration triggered by area CA1. In such cases, MPFC memory models
may guide reactivation and be updated directly, thus bypassing hippocampal
involvement. In contrast, when an existing memory model is weak or nonexistent,
MPFC would play no role in guiding memory retrieval. In this case, new content
would be encoded by hippocampus. Across multiple related experiences (i.e., when
forming a new schema), MPFC may come online (Zeithamova et al. 2012a),
reflecting the emergence of guided reactivation and the abstraction across experi-
ences. However, in many cases, new events are likely to be neither entirely novel
nor identical replications of prior experience. These events will instead share a
moderate level of congruency with existing memory models, and would thus be
expected to involve both MPFC and hippocampus.
While one important characteristic typically ascribed to schemas is the loss of
idiosyncratic details that code the differences among events, it remains unknown
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 427

whether the same is true of integrated memory representations. Anecdotal evidence

from the acquired equivalence paradigm suggests that some event details may also
be lost during integration, as participants failed to recognize inferential probe trials
as novel pairings of stimuli (Shohamy and Wagner 2008). This finding suggests that
details about directly experienced events may sometimes be lost in favor of an
abstracted, generalized framework that codes consistencies among distinct
stimulus-response relationships. However, whether a similar loss of detailed
event information is typical in other inference paradigms, especially those that
utilize rapid acquisition procedures (e.g., single-trial learning), is not known. More
research is needed to understand how the processes supporting inference are related
to those implicated in the formation and use of schemas. Consideration of how task
dynamics influence the type of representational structure formed may provide
important insights into how the hippocampus codes overlapping event information
and interacts with MPFC to support mnemonic flexibility. Moreover, it is notewor-
thy that the operational definition of “schema” varies across species (e.g., spatial
layouts in rodents versus movie knowledge in humans) and across studies within a
species (e.g., movie knowledge versus semantic knowledge in humans). Future
work should seek to bridge the gap between animal and human work to better
specify the conditions and mechanisms that support the building, updating, and use
of memory schemas.

Learning and Associative Facilitation

Recent work suggests that new learning can be promoted by integrating new
information into existing knowledge structures. This phenomenon is highly related
to findings in the schema literature showing a behavioral benefit to encoding
schema-congruent information (described in section “Schemas”). However, here
we make no assumptions about the level of detail retained in the existing knowledge
structure; prior memories need not be generalized.
The observation that prior knowledge can boost learning is by no means new;
classic studies have shown that prior knowledge is beneficial to new learning under
some circumstances (Bransford and Johnson 1972). For example, one such classic
study showed a memory advantage for new responses paired with well-learned old
stimuli (i.e., stimuli previously learned with a different response), a phenomenon
known as associative facilitation (Underwood 1949). These observations appear
robust across species, with existing knowledge of a spatial layout shown to facil-
itate acquisition of new related associations in rodents (Tse et al. 2007), for
example. Such facilitation may also extend to novel judgments that require the
simultaneous consideration of multiple memories (e.g., inferences).
Behaviorally, memory integration has been shown to have a protective effect on
memory; instructing participants to integrate is associated with better memory for
both the initial and newly encoded content (Anderson and McCulloch 1999;
Forcato et al. 2010; see however Richter et al. 2015). Neuroimaging studies using
the associative inference paradigm have shown that memory integration
428 M.L. Schlichting and A.R. Preston

mechanisms may underlie associative facilitation (Schlichting and Preston 2014,

2016). Participants first formed strong memories for (AB) face-object pairs across
four study-test iterations in a pre-training phase. They then encoded new object-
object associations in a single exposure, half of which overlapped with (BC) and
half of which did not overlap with (XY) prior knowledge. Importantly, overlapping
and non-overlapping pairs were matched in terms of content type (two objects) and
number of exposures (one per pair); thus, any differences in neural or behavioral
signatures are attributable to the presence or absence of prior related knowledge.
Results showed that the degree of evidence for memory reactivation during a rest
period following AB pre-training predicted individual differences in the ability to
later encode the new overlapping associations. Moreover, neural signatures during
rest predicted engagement of face-sensitive regions at task, suggesting that offline
memory processing promotes reactivation during the new learning phase
(Schlichting and Preston 2014). Successful overlapping pair encoding was also
associated with engagement of the hippocampal-MPFC circuit (Schlichting and
Preston 2016). These findings suggest that the same memory integration mecha-
nisms that support the ability to make novel inferences spanning events may also
facilitate the encoding of new, related information.

Creativity and Imagination

Memory integration may also underlie the ability to recombine prior memories to
construct new ideas and imagine future scenarios (Schacter et al. 2012). Consistent
with this notion, recent work (Duff et al. 2013) has demonstrated that hippocampal
damage results in impaired performance on creativity tasks in which participants
generate novel responses on the basis of existing knowledge. MPFC may also
support performance in such tasks; one fMRI study (Takeuchi et al. 2012) showed
that individual differences in resting state functional connectivity of MPFC with
posterior cingulate cortex predicted creativity.
Hippocampus and MPFC are also engaged during imagination (Martin et al.
2011; Barron et al. 2013), particularly when imagined scenarios are rich in episodic
detail. One human fMRI study showed enhanced connectivity between hippocam-
pus and MPFC during imagination of future scenarios that were later remembered
(Martin et al. 2011), consistent with the notion that these regions are important for
creating and maintaining integrated memories—even those representing imagined
events. Another study (Barron et al. 2013) required participants to construct mental
representations of novel foods from two familiar ingredients. Using an fMRI
adaptation paradigm, researchers found that imagining novel foods engaged the
same neuronal populations as did the ingredients in both hippocampus and MPFC,
reflecting retrieval and recombination of prior memories during mental construc-
tion. The ingredient items themselves also came to recruit overlapping neuronal
populations, perhaps reflecting integration of the simultaneously reactivated mem-
ories (Fig. 5). Interestingly, the degree of representational overlap of the ingredients
in hippocampus and MPFC tracked across participants with subjective value of the
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 429

imagined foods, suggesting that integration may be enhanced according to behav-

ioral relevance (here, for high value items).

Behavioral Consequences

While we focus primarily on the positive outcomes associated with memory

integration, a few noteworthy studies have highlighted its negative behavioral
consequences. For example, integration may lead to the formation of false memo-
ries (i.e., through overgeneralization) (Cabeza et al. 2001; Warren et al. 2014),
memory misattributions (Hupbach et al. 2007; Jones et al. 2012; Gershman et al.
2013; St. Jacques et al. 2013), and interference (Chan and LaPaglia 2013).
Both MTL and MPFC have been implicated in the formation of false memories.
Neuroimaging studies have reported similar MTL engagement during recognition
of both studied items (“true” memories) and unstudied lures (“false” memories)
(Cabeza et al. 2001; Slotnick and Schacter 2004; Abe et al. 2008), suggesting that
integrated hippocampal representations might underlie the tendency to incorrectly
identify conceptually similar items as having been studied. Interestingly, these
effects appear somewhat specific to anterior aspects of both hippocampus and
MTL cortex (Cabeza et al. 2001; Abe et al. 2008; McTighe et al. 2010), while
more posterior MTL regions (e.g., parahippocampal cortex) typically differentiate
true from false memories based on activation (Cabeza et al. 2001; Okado and Stark
2003; Kim and Cabeza 2007a, b). These results are broadly consistent with the
notion that anterior hippocampus in particular is well suited to integrate across
related memories, perhaps at the cost of memory specificity. Ventral MPFC has also
been implicated in constructing generalized memory representations; patients with
ventral MPFC lesions show reduced false memories relative to healthy control
participants for words that were never seen but are thematically related to a studied
word list (Warren et al. 2014).
Integration may also explain the phenomenon of memory misattribution, in
which an episodic experience is incorrectly attributed to a different encoding
context than the one in which it occurred (e.g., as measured by intrusions). Mis-
attributions may result when prior knowledge is reactivated and updated with the
current experience to the detriment of memory accuracy. One fMRI study
(Gershman et al. 2013) used neural decoding to quantify the reinstatement of the
context associated with prior memories (List 1) during new learning (List 2).
Results showed that greater evidence for reactivation of the List 1 context was
associated with more misattributions of List 2 words to List 1. Another study
(St. Jacques et al. 2013) showed that when participants reactivated a prior experi-
ence during new encoding, engagement of both hippocampus and ventral MPFC
was associated with later memory misattributions, consistent with a role for these
regions in linking experiences across time.
Memory integration mechanisms may also lead to interference or forgetting.
When a memory is retrieved during a new learning experience, that memory
430 M.L. Schlichting and A.R. Preston

becomes malleable and susceptible to change as a function of the current experi-

ence. One possible outcome of learning-phase reactivation is integration—that is,
prior memories are updated to incorporate the new information. However, learning-
phase reactivation can also lead to forgetting of the initial memory under some
conditions (Walker et al. 2003; Forcato et al. 2007; Chan and LaPaglia 2013). For
example, one behavioral study had participants watch a movie of a crime, which
served as the initial memory. Later, participants listened to a narrative describing
the crime that included misinformation: the crime was committed with a different
weapon than the one depicted in the movie. Critically, reactivation of the initial
memory prior to hearing the narrative resulted in forgetting of the initial, “true”
memory of the crime (Chan and LaPaglia 2013). By design, the newly learned
information in that study directly competed with or replaced the prior knowledge.
Thus, whether memories for the original events are “overwritten” or simply updated
to incorporate the new information may depend largely on the degree to which the
two memories are compatible (Hupbach 2011).
It is notable that in the hippocampus, forgetting has typically been attributed to
passive decay rather than interference due to the strong hippocampal tendency to
pattern separate (Hardt et al. 2013). However, recent work suggesting that hippo-
campus—particularly its anterior portion—can form integrated codes that span
related memories (Collin et al. 2015; Schlichting et al. 2015) calls this view into
question. That is, memory integration predicts that even hippocampal memories
may be forgotten when related content is incorporated into existing memory traces
(i.e., through interference).

Conclusions

In summary, extensive evidence indicates that the hippocampus and its interactions
with MPFC promote memory integration processes that support flexible cognition.
Hippocampus does so by building memory representations that code not only
associations within individual events, but also relationships spanning multiple
episodes. In this way, the function of the hippocampus is not merely to enable the
retrospective use of memory; rather, hippocampal function is “intrinsically pro-
spective” (Klein et al. 2002a), aimed at constructing representations that can be
used to successfully negotiate future judgments and actions. Integration tasks thus
provide a powerful tool for studying the adaptive nature of memory and how the
computational properties of the hippocampus allow memories to be reconstructed
into prospectively useful formats.
The findings described here collectively suggest the importance of hippocampal
encoding processes in linking related experiences. Integrated memories may sup-
port a host of flexible behaviors, from navigating our environment to imagining our
future. Importantly, hippocampus does not work in isolation; rather, it communi-
cates with other cortical regions to facilitate reactivation of memories, encoding of
new memories, and updating of existing representations to incorporate new
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 431

information. In doing so, it plays a key role in the extraction of knowledge across
learning events.

Acknowledgments The authors were supported by the National Institute of Mental Health of the
National Institutes of Health under award number R01MH100121 and by the National Science
Foundation under CAREER award number 1056019 to Alison R. Preston during the writing of this
chapter.

References

Abe N, Okuda J, Suzuki M et al (2008) Neural correlates of true memory, false memory, and
deception. Cereb Cortex 18:2811–2819. doi:10.1093/cercor/bhn037
Addis DR, Schacter DL (2012) The hippocampus and imagining the future: where do we stand?
Front Hum Neurosci 5:173. doi:10.3389/fnhum.2011.00173
Addis DR, Wong AT, Schacter DL (2007) Remembering the past and imagining the future:
common and distinct neural substrates during event construction and elaboration.
Neuropsychologia 45:1363–1377. doi:10.1016/j.neuropsychologia.2006.10.016
Anderson MC, McCulloch KC (1999) Integration as a general boundary condition on retrieval-
induced forgetting. J Exp Psychol Learn Mem Cogn 25:608–629. doi:10.1037/0278-7393.25.3.
608
Barbas H, Blatt GJ (1995) Topographically specific hippocampal projections target functionally
distinct prefrontal areas in the rhesus monkey. Hippocampus 5:511–533
Barron HC, Dolan RJ, Behrens TEJ (2013) Online evaluation of novel choices by simultaneous
representation of multiple memories. Nat Neurosci 16:1492–1498. doi:10.1038/nn.3515
Bartlett F (1932) Remembering: a study in experimental and social psychology. Cambridge
University Press, Cambridge
Blanquat PDS, Hok V, Save E et al (2013) Differential role of the dorsal hippocampus, ventro-
intermediate hippocampus, and medial prefrontal cortex in updating the value of a spatial goal.
Hippocampus 23:342–351. doi:10.1002/hipo.22094
Bransford JD, Johnson MK (1972) Contextual prerequisites for understanding: some investiga-
tions of comprehension and recall. J Verbal Learn Verbal Behav 11:717–726
Buckner RL (2010) The role of the hippocampus in prediction and imagination. Annu Rev Psychol
61:27–48. doi:10.1146/annurev.psych.60.110707.163508
Bunsey M, Eichenbaum H (1996) Conservation of hippocampal memory function in rats and
humans. Nature 379:255–257. doi:10.1038/379255a0
Cabeza R, Rao SM, Wagner AD et al (2001) Can medial temporal lobe regions distinguish true
from false? An event-related functional MRI study of veridical and illusory recognition
memory. Proc Natl Acad Sci U S A 98:4805–4810. doi:10.1073/pnas.081082698
Catanese J, Viggiano A, Cerasti E et al (2014) Retrospectively and prospectively modulated
hippocampal place responses are differentially distributed along a common path in a contin-
uous T-maze. J Neurosci 34:13163–13169. doi:10.1523/JNEUROSCI.0819-14.2014
Cavada C, Compa~ny T, Tejedor J et al (2000) The anatomical connections of the macaque monkey
orbitofrontal cortex. A review. Cereb Cortex 10:220–242
Chan JCK, LaPaglia JA (2013) Impairing existing declarative memory in humans by disrupting
reconsolidation. Proc Natl Acad Sci U S A 110:9309–9313. doi:10.1073/pnas.1218472110
Chua EF, Schacter DL, Rand-Giovannetti E, Sperling RA (2007) Evidence for a specific role of the
anterior hippocampal region in successful associative encoding. Hippocampus 17:1071–1080.
doi:10.1002/hipo.20340
Cohen NJ, Eichenbaum H (1993) Memory, amnesia, and the hippocampal system. The MIT Press,
Cambridge, MA
432 M.L. Schlichting and A.R. Preston

Collin SHP, Milivojevic B, Doeller CF (2015) Memory hierarchies map onto the hippocampal
long axis in humans. Nat Neurosci 18:1–5. doi:10.1038/nn.4138
Coutanche MN, Gianessi CA, Chanales AJH et al (2013) The role of sleep in forming a memory
representation of a two-dimensional space. Hippocampus 23:1189–1197. doi:10.1002/hipo.
22157
Craig M, Dewar M, Harris MA, Della Sala S (2015) Wakeful rest promotes the integration of
spatial memories into accurate cognitive maps. Hippocampus. doi:10.1002/hipo.22502
de Bruin JPC, S!anchez-Santed F, Heinsbroek RPW et al (1994) A behavioural analysis of rats with
damage to the medial prefrontal cortex using the morris water maze: evidence for behavioural
flexibility, but not for impaired spatial navigation. Brain Res 652:323–333
de Hoz L, Martin SJ (2014) Double dissociation between the contributions of the septal and
temporal hippocampus to spatial learning: the role of prior experience. Hippocampus
24:990–1005. doi:10.1002/hipo.22285
Demaster DM, Pathman T, Lee JK, Ghetti S (2013) Structural development of the hippocampus
and episodic memory: developmental differences along the anterior/posterior axis. Cereb
Cortex 24:3036–3045
Deuker L, Olligs J, Fell J et al (2013) Memory consolidation by replay of stimulus-specific neural
activity. J Neurosci 33:19373–19383. doi:10.1523/JNEUROSCI.0414-13.2013
DeVito LM, Kanter BR, Eichenbaum H (2010a) The hippocampus contributes to memory
expression during transitive inference in mice. Hippocampus 20:208–217. doi:10.1002/hipo.
20610
DeVito LM, Lykken C, Kanter BR, Eichenbaum H (2010b) Prefrontal cortex: role in acquisition of
overlapping associations and transitive inference. Learn Mem 17:161–167
Diekelmann S, Born J (2010) The memory function of sleep. Nat Rev Neurosci 11:114–126.
doi:10.1038/nrn2762
Duff MC, Kurczek J, Rubin R et al (2013) Hippocampal amnesia disrupts creative thinking.
Hippocampus 23:1143–1149. doi:10.1002/hipo.22208
Duncan K, Ketz N, Inati SJ, Davachi L (2012a) Evidence for area CA1 as a match/mismatch
detector: a high-resolution fMRI study of the human hippocampus. Hippocampus 22:389–398.
doi:10.1002/hipo.20933
Duncan K, Sadanand A, Davachi L (2012b) Memory’s penumbra: episodic memory decisions
induce lingering mnemonic biases. Science 337:485–487. doi:10.1126/science.1221936
Eichenbaum H, Schoenbaum G, Young B, Bunsey M (1996) Functional organization of the
hippocampal memory system. Proc Natl Acad Sci U S A 93:13500–13507
Eichenbaum H, Dudchenko PA, Wood E et al (1999) The hippocampus, memory, and place cells:
is it spatial memory or a memory space? Neuron 23:209–226
Ellenbogen JM, Hu PT, Payne JD et al (2007) Human relational memory requires time and sleep.
Proc Natl Acad Sci U S A 104:7723–7728. doi:10.1073/pnas.0700094104
Euston DR, Gruber AJ, McNaughton BL (2012) The role of medial prefrontal cortex in memory
and decision making. Neuron 76:1057–1070. doi:S0896-6273(12)01108-7
Forcato C, Burgos VL, Argibay PF et al (2007) Reconsolidation of declarative memory in humans.
Learn Mem 14:295–303. doi:10.1101/lm.486107
Forcato C, Rodrı́guez MLC, Pedreira ME, Maldonado H (2010) Reconsolidation in humans opens
up declarative memory to the entrance of new information. Neurobiol Learn Mem 93:77–84.
doi:10.1016/j.nlm.2009.08.006
Gerraty RT, Davidow JY, Wimmer GE et al (2014) Transfer of learning relates to intrinsic
connectivity between hippocampus, ventromedial prefrontal cortex, and large-scale networks.
J Neurosci 34:11297–11303. doi:10.1523/JNEUROSCI.0185-14.2014
Gershman SJ, Schapiro AC, Hupbach A, Norman KA (2013) Neural context reinstatement predicts
memory misattribution. J Neurosci 33:8590–8595. doi:10.1523/JNEUROSCI.0096-13.2013
Ghosh VE, Moscovitch M, Melo Colella B, Gilboa A (2014) Schema representation in patients
with ventromedial PFC lesions. J Neurosci 34:12057–12070. doi:10.1523/JNEUROSCI.0740-
14.2014
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 433

Gupta AS, van der Meer MAA, Touretzky DS, Redish AD (2010) Hippocampal replay is not a
simple function of experience. Neuron 65:695–705. doi:10.1016/j.neuron.2010.01.034
Hardt O, Nader K, Nadel L (2013) Decay happens: the role of active forgetting in memory. Trends
Cogn Sci 17:109–118. doi:10.1016/j.tics.2013.01.001
Hasselmo ME, Schnell E (1994) Laminar selectivity of the cholinergic suppression of synaptic
transmission in rat hippocampal region CA1: computational modeling and brain slice physi-
ology. J Neurosci 14:3898–3914
Hasselmo ME, Wyble BP, Wallenstein GV (1996) Encoding and retrieval of episodic memories:
role of cholinergic and GABAergic modulation in the hippocampus. Hippocampus 6:693–708.
doi:10.1002/(SICI)1098-1063(1996)6:6<693::AID-HIPO12>3.0.CO;2-W
Heckers S, Zalesak M, Weiss AP et al (2004) Hippocampal activation during transitive inference
in humans. Hippocampus 14:153–162. doi:10.1002/hipo.10189
Henke K, Reber TP, Duss SB (2013) Integrating events across levels of consciousness. Front
Behav Neurosci 7:68. doi:10.3389/fnbeh.2013.00068
Hoffman KL, Mcnaughton BL (2002) Sleep on it: cortical reorganization after-the fact. Trends
Neurosci 25:1–2
Howard MW, Fotedar MS, Datey AV, Hasselmo ME (2005) The temporal context model in spatial
navigation and relational learning: toward a common explanation of medial temporal lobe
function across domains. Psychol Rev 112:75–116. doi:10.1037/0033-295X.112.1.75
Hupbach A (2011) The specific outcomes of reactivation-induced memory changes depend on the
degree of competition between old and new information. Front Behav Neurosci 5:33. doi:10.
3389/fnbeh.2011.00033
Hupbach A, Gomez R, Hardt O, Nadel L (2007) Reconsolidation of episodic memories: a subtle
reminder triggers integration of new information. Learn Mem 14:47–53. doi:10.1101/lm.
365707
Iordanova MD, Killcross AS, Honey RC (2007) Role of the medial prefrontal cortex in acquired
distinctiveness and equivalence of cues. Behav Neurosci 121:1431–1436. doi:10.1037/0735-
7044.121.6.1431
Jadhav SP, Kemere C, German PW, Frank LM (2012) Awake hippocampal sharp-wave ripples
support spatial memory. Science 336:1454–1457. doi:10.1126/science.1217230
Ji D, Wilson MA (2007) Coordinated memory replay in the visual cortex and hippocampus during
sleep. Nat Neurosci 10:100–107
Johnson A, Redish AD (2007) Neural ensembles in CA3 transiently encode paths forward of the
animal at a decision point. J Neurosci 27:12176–12189. doi:10.1523/JNEUROSCI.3761-07.
2007
Jones B, Bukoski E, Nadel L, Fellous J-M (2012) Remaking memories: reconsolidation updates
positively motivated spatial memory in rats. Learn Mem 19:91–98. doi:10.1101/lm.023408.
111
Kahana M (1996) Associative retrieval processes in free recall. Mem Cognit 24:103–109. doi:10.
3758/BF03197276
Karlsson MP, Frank LM (2009) Awake replay of remote experiences in the hippocampus. Nat
Neurosci 12:913–918. doi:10.1038/nn.2344
Kim H, Cabeza R (2007a) Differential contributions of prefrontal, medial temporal, and sensory-
perceptual regions to true and false memory formation. Cereb Cortex 17:2143–2150. doi:10.
1093/cercor/bhl122
Kim H, Cabeza R (2007b) Trusting our memories: dissociating the neural correlates of confidence
in veridical versus illusory memories. J Neurosci 27:12190–12197. doi:10.1523/JNEUROSCI.
3408-07.2007
Kirwan CB, Stark CEL (2004) Medial temporal lobe activation during encoding and retrieval of
novel face-name pairs. Hippocampus 14:919–930. doi:10.1002/hipo.20014
Klein SB, Cosmides L, Tooby J, Chance S (2002a) Decisions and the evolution of memory:
multiple systems, multiple functions. Psychol Rev 109:306–329. doi:10.1037//0033-295X.109.
2.306
434 M.L. Schlichting and A.R. Preston

Klein SB, Loftus J, Kihlstrom JF (2002b) Memory and temporal experience: the effects of episodic
memory loss on an amnesic patient’s ability to remember the past and imagine the future. Soc
Cogn 20:353–379
Komorowski RW, Garcia CG, Wilson A et al (2013) Ventral hippocampal neurons are shaped by
experience to represent behaviorally relevant contexts. J Neurosci 33:8079–8087. doi:10.1523/
JNEUROSCI.5458-12.2013
Koscik TR, Tranel D (2012) The human ventromedial prefrontal cortex is critical for transitive
inference. J Cogn Neurosci 24:1191–1204. doi:10.1162/jocn_a_00203
Kroes MCW, Fernández G (2012) Dynamic neural systems enable adaptive, flexible memories.
Neurosci Biobehav Rev 36:1646–1666. doi:10.1016/j.neubiorev.2012.02.014
Kuhl BA, Shah AT, DuBrow S, Wagner AD (2010) Resistance to forgetting associated with
hippocampus-mediated reactivation during new learning. Nat Neurosci 13:501–506. doi:10.
1038/nn.2498
Kuhl BA, Bainbridge WA, Chun MM (2012) Neural reactivation reveals mechanisms for updating
memory. J Neurosci 32:3453–3461. doi:10.1523/JNEUROSCI.5846-11.2012
Kumaran D (2012) What representations and computations underpin the contribution of the
hippocampus to generalization and inference? Front Hum Neurosci 6:157. doi:10.3389/
fnhum.2012.00157
Kumaran D (2013) Schema-driven facilitation of new hierarchy learning in the transitive inference
paradigm. Learn Mem 20:388–394. doi:10.1101/lm.030296.113
Kumaran D, McClelland JL (2012) Generalization through the recurrent interaction of episodic
memories: a model of the hippocampal system. Psychol Rev 119:573–616. doi:10.1037/
a0028681
Kumaran D, Melo HL (2013) Transitivity performance, relational hierarchy knowledge and
awareness: results of an instructional framing manipulation. Hippocampus 23:1259–1268.
doi:10.1002/hipo.22163
Kumaran D, Summerfield JJ, Hassabis D, Maguire EA (2009) Tracking the emergence of
conceptual knowledge during human decision making. Neuron 63:889–901. doi:10.1016/j.
neuron.2009.07.030
Kumaran D, Melo H, Duzel E (2012) The emergence and representation of knowledge about social
and nonsocial hierarchies. Neuron 76:653–666. doi:10.1016/j.neuron.2012.09.035
Larkin MC, Lykken C, Tye LD et al (2014) Hippocampal output area CA1 broadcasts a general-
ized novelty signal during an object-place recognition task. Hippocampus 24:773–783. doi:10.
1002/hipo.22268
Lewis PA, Durrant SJ (2011) Overlapping memory replay during sleep builds cognitive schemata.
Trends Cogn Sci 15:343–351. doi:10.1016/j.tics.2011.06.004
Lisman JE, Grace AA (2005) The hippocampal-VTA loop: controlling the entry of information
into long-term memory. Neuron 46:703–713. doi:10.1016/j.neuron.2005.05.002
Marr D (1970) A theory for cerebral neocortex. Proc R Soc Lond B Biol Sci 176:161–234
Martin VC, Schacter DL, Corballis MC, Addis DR (2011) A role for the hippocampus in encoding
simulations of future events. Proc Natl Acad Sci U S A 108:13858–13863. doi:10.1073/pnas.
1105816108
McClelland JL, McNaughton BL, O’Reilly RC (1995) Why there are complementary learning
systems in the hippocampus and neocortex: insights from the successes and failures of
connectionist models of learning and memory. Psychol Rev 102:419–457. doi:10.1037//
0033-295X.102.3.419
McKenzie S, Eichenbaum H (2011) Consolidation and reconsolidation: two lives of memories?
Neuron 71:224–233. doi:10.1016/j.neuron.2011.06.037
McKenzie S, Robinson NTM, Herrera L et al (2013) Learning causes reorganization of neuronal
firing patterns to represent related experiences within a hippocampal schema. J Neurosci
33:10243–10256. doi:10.1523/JNEUROSCI.0879-13.2013
McKenzie S, Frank AJ, Kinsky NR et al (2014) Hippocampal representation of related and
opposing memories develop within distinct, hierarchically organized neural schemas. Neuron
83:202–215. doi:10.1016/j.neuron.2014.05.019
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 435

McTighe SM, Cowell RA, Winters BD et al (2010) Paradoxical false memory for objects after
brain damage. Science 330:1408–1410. doi:10.1126/science.1194780
Miller EK, Cohen JD (2001) An integrative theory of prefrontal cortex function. Annu Rev
Neurosci 24:167–202. doi:10.1146/annurev.neuro.24.1.167
Munnelly A, Dymond S (2014) Relational memory generalization and integration in a transitive
inference task with and without instructed awareness. Neurobiol Learn Mem 109:169–177.
doi:10.1016/j.nlm.2014.01.004
Nadel L, Hardt O (2011) Update on memory systems and processes. Neuropsychopharmacology
36:251–273. doi:10.1038/npp.2010.169
Nadel L, Moscovitch M (1997) Memory consolidation, retrograde amnesia, and the hippocampal
complex. Curr Opin Neurobiol 7:217–227
Nadel L, Samsonovich A, Ryan L, Moscovitch M (2000) Multiple trace theory of human memory:
computational, neuroimaging, and neuropsychological results. Hippocampus 10:352–368
Nadel L, Hupbach A, Gomez R, Newman-Smith K (2012) Memory formation, consolidation and
transformation. Neurosci Biobehav Rev 36:1640–1645. doi:10.1016/j.neubiorev.2012.03.001
O’Keefe J, Nadel L (1978) The hippocampus as a cognitive map. Clarendon, London
O’Reilly RC, Rudy JW (2001) Conjunctive representations in learning and memory: principles of
cortical and hippocampal function. Psychol Rev 108:311–345. doi:10.1037//0033-295X.108.2.
311
Okado Y, Stark C (2003) Neural processing associated with true and false memory retrieval. Cogn
Affect Behav Neurosci 3:323–334. doi:10.3758/CABN.3.4.323
Pfeiffer BE, Foster DJ (2013) Hippocampal place-cell sequences depict future paths to remem-
bered goals. Nature 497:74–81. doi:10.1038/nature12112
Poppenk J, Evensmoen HR, Moscovitch M, Nadel L (2013) Long-axis specialization of the human
hippocampus. Trends Cogn Sci 17:230–240. doi:10.1016/j.tics.2013.03.005
Preston AR, Eichenbaum H (2013) Interplay of hippocampus and prefrontal cortex in memory.
Curr Biol 23:R764–R773. doi:10.1016/j.cub.2013.05.041
Preston AR, Shrager Y, Dudukovic N, Gabrieli JDE (2004) Hippocampal contribution to the novel
use of relational information in declarative memory. Hippocampus 14:148–152. doi:10.1002/
hipo.20009
Price JL, Drevets WC (2009) Neurocircuitry of mood disorders. Neuropsychopharmacology
35:192–216. doi:10.1038/npp.2009.104
Richards BA, Xia F, Santoro A et al (2014) Patterns across multiple memories are identified over
time. Nat Neurosci 17:981–986. doi:10.1038/nn.3736
Richter FR, Chanales AJH, Kuhl BA (2015) Predicting the integration of overlapping memories by
decoding mnemonic processing states during learning. Neuroimage. doi:10.1016/j.
neuroimage.2015.08.051
Roy M, Shohamy D, Wager TD (2012) Ventromedial prefrontal-subcortical systems and the
generation of affective meaning. Trends Cogn Sci 16:147–156. doi:10.1016/j.tics.2012.01.005
Sara SJ (2010) Reactivation, retrieval, replay and reconsolidation in and out of sleep: connecting
the dots. Front Behav Neurosci 4:185. doi:10.3389/fnbeh.2010.00185
Schacter DL, Wagner AD (1999) Medial temporal lobe activations in fMRI and PET studies of
episodic encoding and retrieval. Hippocampus 9:7–24
Schacter DL, Addis DR, Hassabis D et al (2012) The future of memory: remembering, imagining,
and the brain. Neuron 76:677–694
Schiller D, Phelps EA (2011) Does reconsolidation occur in humans? Front Behav Neurosci 5:24.
doi:10.3389/fnbeh.2011.00024
Schinazi VR, Nardi D, Newcombe NS et al (2013) Hippocampal size predicts rapid learning of a
cognitive map in humans. Hippocampus 23:515–528. doi:10.1002/hipo.22111
Schlichting ML, Preston AR (2014) Memory reactivation during rest supports upcoming learning
of related content. Proc Natl Acad Sci U S A 111:15845–15850. doi:10.1073/pnas.1404396111
Schlichting ML, Preston AR (2015) Memory integration: neural mechanisms and implications for
behavior. Curr Opin Behav Sci 1:1–8. doi:10.1016/j.cobeha.2014.07.005
436 M.L. Schlichting and A.R. Preston

Schlichting ML, Preston AR (2016) Hippocampal-medial prefrontal circuit supports memory

updating during learning and post-encoding rest. Neurobiol Learn Mem. doi:10.1016/j.nlm.
2015.11.005
Schlichting ML, Zeithamova D, Preston AR (2014) CA1 subfield contributions to memory
integration and inference. Hippocampus 24:1248–1260. doi:10.1002/hipo.22310
Schlichting ML, Mumford JA, Preston AR (2015) Learning-related representational changes
reveal dissociable integration and separation signatures in the hippocampus and prefrontal
cortex. Nat Commun 6:8151. doi:10.1038/ncomms9151
Sharon T, Moscovitch M, Gilboa A (2011) Rapid neocortical acquisition of long-term arbitrary
associations independent of the hippocampus. Proc Natl Acad Sci U S A 108:1146–1151.
doi:10.1073/pnas.1005238108/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.
1005238108
Shohamy D, Wagner AD (2008) Integrating memories in the human brain: hippocampal-midbrain
encoding of overlapping events. Neuron 60:378–389. doi:10.1016/j.neuron.2008.09.023
Singer AC, Karlsson MP, Nathe AR et al (2010) Experience-dependent development of coordi-
nated hippocampal spatial activity representing the similarity of related locations. J Neurosci
30:11586–11604. doi:10.1523/JNEUROSCI.0926-10.2010
Slotnick SD, Schacter DL (2004) A sensory signature that distinguishes true from false memories.
Nat Neurosci 7:664–672. doi:10.1038/nn1252
Smith CN, Urgolites ZJ, Hopkins RO, Squire LR (2014) Comparison of explicit and incidental
learning strategies in memory-impaired patients. Proc Natl Acad Sci U S A 111:475–479.
doi:10.1073/pnas.1322263111
St. Jacques PL, Olm C, Schacter DL (2013) Neural mechanisms of reactivation-induced updating
that enhance and distort memory. Proc Natl Acad Sci U S A 110:19671–19678. doi:10.1073/
pnas.1319630110
Staresina BP, Alink A, Kriegeskorte N, Henson RN (2013) Awake reactivation predicts memory in
humans. Proc Natl Acad Sci U S A 110:21159–21164
Strange BA, Witter MP, Lein ES, Moser EI (2014) Functional organization of the hippocampal
longitudinal axis. Nat Rev Neurosci 15:655–669. doi:10.1038/nrn3785
Takeuchi H, Taki Y, Hashizume H et al (2012) The association between resting functional
connectivity and creativity. Cereb Cortex 22:2921–2929. doi:10.1093/cercor/bhr371
Tambini A, Ketz N, Davachi L (2010) Enhanced brain correlations during rest are related to
memory for recent experiences. Neuron 65:280–290. doi:10.1016/j.neuron.2010.01.001
Tavares RM, Mendelsohn A, Grossman Y et al (2015) A map for social navigation in the human
brain. Neuron 87:231–243. doi:10.1016/j.neuron.2015.06.011
Tolman EC (1948) Cognitive maps in rats and men. Psychol Rev 55:189–208. doi:10.1037/
h0061626
Tse D, Langston RF, Kakeyama M et al (2007) Schemas and memory consolidation. Science
316:76–82. doi:10.1126/science.1135935
Tse D, Takeuchi T, Kakeyama M et al (2011) Schema-dependent gene activation and memory
encoding in neocortex. Science 333:891–895. doi:10.1126/science.1205274
Underwood BJ (1949) Proactive inhibition as a function of time and degree of prior learning. J Exp
Psychol 39:24–34
van Kesteren MTR, Fernández G, Norris DG, Hermans EJ (2010) Persistent schema-dependent
hippocampal-neocortical connectivity during memory encoding and postencoding rest in
humans. Proc Natl Acad Sci U S A 107:7550–7555. doi:10.1073/pnas.0914892107
van Kesteren MTR, Ruiter DJ, Fernández G, Henson RN (2012) How schema and novelty
augment memory formation. Trends Neurosci 35:211–219. doi:10.1016/j.tins.2012.02.001
van Kesteren MTR, Beul SF, Takashima A et al (2013) Differential roles for medial prefrontal and
medial temporal cortices in schema-dependent encoding: from congruent to incongruent.
Neuropsychologia 51:2352–2359. doi:10.1016/j.neuropsychologia.2013.05.027
van Kesteren MTR, Rijpkema M, Ruiter DJ et al (2014) Building on prior knowledge: schema-
dependent encoding processes relate to academic performance. J Cogn Neurosci
26:2250–2261. doi:10.1162/jocn
The Hippocampus and Memory Integration: Building Knowledge to Navigate. . . 437

Walker MP, Brakefield T, Hobson JA (2003) Dissociable stages of human memory consolidation
and reconsolidation. Nature 425:616–620. doi:10.1038/nature01951.1
Wang S-H, Tse D, Morris RGM (2012) Anterior cingulate cortex in schema assimilation and
expression. Learn Mem 19:315–318. doi:10.1101/lm.026336.112
Warren DE, Duff MC (2014) Not so fast: hippocampal amnesia slows word learning despite
successful fast mapping. Hippocampus 24:920–933. doi:10.1002/hipo.22279
Warren DE, Jones SH, Duff MC, Tranel D (2014) False recall is reduced by damage to the
ventromedial prefrontal cortex: implications for understanding the neural correlates of sche-
matic memory. J Neurosci 34:7677–7682. doi:10.1523/JNEUROSCI.0119-14.2014
Werchan DM, Gómez RL (2013) Generalizing memories over time: sleep and reinforcement
facilitate transitive inference. Neurobiol Learn Mem 100:70–76. doi:10.1016/j.nlm.2012.12.
006
Wilson RC, Takahashi YK, Schoenbaum G, Niv Y (2014) Orbitofrontal cortex as a cognitive map
of task space. Neuron 81:267–279. doi:10.1016/j.neuron.2013.11.005
Wimmer GE, Shohamy D (2012) Preference by association: how memory mechanisms in the
hippocampus bias decisions. Science 338:270–273. doi:10.1126/science.1223252
Wood ER, Dudchenko PA, Eichenbaum H (1999) The global record of memory in hippocampal
neuronal activity. Nature 397:613–616. doi:10.1038/17605
Wu X, Foster DJ (2014) Hippocampal replay captures the unique topological structure of a novel
environment. J Neurosci 34:6459–6469. doi:10.1523/JNEUROSCI.3414-13.2014
Zalesak M, Heckers S (2009) The role of the hippocampus in transitive inference. Psychiatry Res
Neuroimaging 172:24–30. doi:10.1016/j.pscychresns.2008.09.008
Zeithamova D, Preston AR (2010) Flexible memories: differential roles for medial temporal lobe
and prefrontal cortex in cross-episode binding. J Neurosci 30:14676–14684. doi:10.1523/
JNEUROSCI.3250-10.2010
Zeithamova D, Dominick AL, Preston AR (2012a) Hippocampal and ventral medial prefrontal
activation during retrieval-mediated learning supports novel inference. Neuron 75:168–179.
doi:10.1016/j.neuron.2012.05.010
Zeithamova D, Schlichting ML, Preston AR (2012b) The hippocampus and inferential reasoning:
building memories to navigate future decisions. Front Hum Neurosci 6:70. doi:10.3389/fnhum.
2012.00070