Download Memory Consolidation - Chapter 21 and more Papers Psychology in PDF only on Docsity!
C H A P T E R
Abstract Memory consolidation is a multifaceted concept. At a minimum, it refers to both cellular consolidation and systems consolidation. Cellular consolidation takes place in the hours after learning, stabilizing the memory trace—a process that may involve structural changes in hippocampal neurons. Systems consolidation refers to a more protracted process by which memories become independent of the hippocampus as they are established in cortical neurons—a process that may involve neural replay. Both forms of consolidation may preferentially unfold whenever the hippocampus is not encoding new information, although some theories hold that consolidation occurs exclusively during sleep. In recent years, the notion of reconsolidation has been added to the mix. According to this idea, previously consolidated memories, when later retrieved, undergo consolidation all over again. With new findings coming to light seemingly every day, the concept of consolidation will likely evolve in interesting and unpredictable ways in the years to come. Key Words: cellular consolidation, systems consolidation, reconsolidation, sleep and consolidation
Memory Consolidation
John T. Wixted and Denise J. Cai
Th e idea that memories require time to consoli- date has a long history, but the understanding of what consolidation means has evolved over time. In 1900, the German experimental psychologists Georg Müller and Alfons Pilzecker published a monograph in which a new theory of memory and forgetting was proposed, one that included—for the fi rst time—a role for consolidation. Th eir basic method involved asking subjects to study a list of paired-associate nonsense syllables and then testing their memory using cued recall after a delay of several minutes. Typically, some of the list items were for- gotten, and to investigate why that occurred, Müller and Pilzecker (1900) presented subjects with a sec- ond, interfering list of items to study before mem- ory for the target list was tested. They found that this interpolated list reduced memory for the target list compared with a control group that was not exposed to any intervening activity. Critically, the
position of the interfering list within the retention interval mattered such that interference occurring soon after learning had a more disruptive effect than interference occurring later in the retention inter- val. This led them to propose that memories require time to consolidate and that retroactive interference is a force that compromises the integrity of recently formed (and not-yet-consolidated) memories. In this chapter, we review the major theories of consol- idation—beginning with the still-relevant account proposed by Müller and Pilzecker (1900)—and we consider a variety of recent developments in what has become a rapidly evolving field.
Th e Early View: Consolidation and
Resistance to Interference
According to Müller and Pilzecker’s (1900) view, consolidation consists of “trace hardening” (cf. Wickelgren, 1974) in the sense that some
wixted, cai 437
become stronger, so it is important to keep in mind which meaning of a “stronger memory trace” applies in any discussion of consolidation. One way that a trace might become stronger is that it comes to more accurately reflect past experience than it did when it was first formed, much like a snapshot taken from a Polaroid camera comes into sharper focus over time. A trace that consolidated in this manner would yield an ever-clearer memory of the encoding event in response to the same retrieval cue. Another way that a trace might become stronger is that it becomes ever more likely to spring to mind in response to a retrieval cue (even more likely than it was when the memory was first formed). A memory trace that consolidated in either of these two ways would sup- port a higher level of performance than it did at the end of training, as if additional learning occurred despite the absence of additional training. Still another way that a trace can become stron- ger is that it becomes hardened against the destruc- tive forces of interference. A trace that hardens over time (i.e., a trace that consolidates in that sense) may simultaneously become degraded over time due to the interfering force of new learning or to some other force of decay. As an analogy, a clay replica of the Statue of Liberty will be at its finest when it has just been completed and the clay is still wet, but it will also be at its most vulnerable. With the passage of time, however, the statue dries and becomes more resistant to damage even though it may now be a less accurate replica than it once was (because of the damage that occurred before the clay dried). Müller and Pilzecker’s (1900) original view of consolidation, which was later elaborated by Wickelgren (1974), was analogous to this. That is, the consolidation process was not thought to render the trace more representative of past experience or to render it more likely to come to mind than it was at the time of formation; instead, consolidation was assumed to render the trace (or its association with a retrieval cue) more resistant to interference even while the integrity of the trace was gradually being compromised by interference. Th ese considerations suggest a relationship between Müller and Pilzecker’s (1900) view of con- solidation and the time course of forgetting. More specifi cally, the fact that a memory trace hardens in such a way as to become increasingly resistant to interference even as the trace fades may help to explain the general shape of the forgetting func- tion (Wixted, 2004b). Since the seminal work of Ebbinghaus (1885), a consistent body of evi- dence has indicated that the proportional rate of
physiological process perseverates and eventually renders the memory trace less vulnerable to inter- ference caused by new learning. The kind of inter- ference that a consolidated trace theoretically resists differs from the kind of interference that most experimental psychologists have in mind when they study forgetting. In the field of experimental psychology, new learning has long been thought to generate interference by creating competing asso- ciations linked to a retrieval cue, not by aff ecting the integrity of a fragile memory trace (e.g., Keppel, 1968; Underwood, 1957; Watkins & Watkins, 1975). Traditionally, this kind of interference has been investigated using an A-B, A-C paired-associ- ates paradigm in which the same cue words (the A items) are paired with different to-be-remembered target words across two lists (the B and C items, respectively). In a standard retroactive interference paradigm, for example, the memory test consists of presenting the A items and asking participants to recall the B items. Having learned the A-C associa- tions after learning the A-B associations, the abil- ity of participants to recall the B items is typically impaired, and this impairment is usually assumed to reflect retrieval competition from the C items. Th e powerful effect of this kind of “cue overload” interference on retention has been well established by decades of psychological research, but it is almost certainly not the only kind of interference that causes forgetting. The kind of interference envisioned by Müller and Pilzecker (1900) does not involve overloading a retrieval cue but instead involves directly com- promising the integrity of a partially consolidated memory trace. In what even today seems like a radical notion to many experimental psycholo- gists, Müller and Pilzecker (1900) assumed that the interference was nonspecific in the sense that the interfering material did not have to be similar to the originally memorized material for interference to occur. Instead, mental exertion of any kind was thought to be the interfering force (Lechner et al., 1999). “Mental exertion” is fairly vague concept, and Wixted (2004a) suggested that the kind of inter- vening mental exertion that Müller and Pilzecker (1900) probably had in mind consists specifically of new learning. Th e basic idea is that new learning, per se, serves as an interfering force that degrades recently formed and still fragile memory traces. Loosely speaking, it can be said that Müller and Pilzecker (1900) believed that the memory trace becomes strengthened by the process of consolidation. However, there is more than one way that a trace can
wixted, cai 439
performed on unimpaired subjects, though the rel- evant literature is somewhat mixed in this regard. Although some studies found more activity in the MTL during the recollection of recent semantic memories compared with remote semantic memo- ries (Douville et al., 2005; Haist et al., 2001; Smith & Squire, 2009), other studies found no difference (e.g., Bernard et al., 2004; Maguire et al., 2001; Maguire & Frith, 2003). For example, using func- tional magnetic resonance imaging (fMRI), Bernard et al. (2004) identified brain regions associated with recognizing famous faces from two different periods: people who became famous in the 1960s to 1970s and people who became famous in the 1990s. They found that the hippocampus was similarly active during the recognition of faces from both periods (i.e., no temporal gradient was observed). It is not clear why studies vary in this regard, but one pos- sibility is that the detection of a temporal gradient is more likely when multiple time points are assessed, especially during the several years immediately preceding memory impairment, than when only two time points are assessed (as in Bernard et al., 2004). In an imaging study that was patterned after the lesion study reported by Manns et al. (2003), Smith and Squire (2009) measured brain activity while subjects recalled news events from multiple time points over the past 30 years. In agreement with the lesion study, they found that regions in the MTL exhibited a decrease in brain activity as a function of the age of the memory over a 12-year period (whereas activity was constant for memories from 13 to 30 years ago). In addition, they found that regions in the frontal lobe, temporal lobe, and parietal lobe exhibited an increase in activity as a function of the age of the trace. Thus, it seems that the (systems) consolidation of semantic memories is a slow process that may require years to complete.
Temporal Gradient of Autobiographical
(Episodic) Memory
Although lesion studies and neuroimaging studies point to a temporal gradient for semantic memory lasting years, there is some debate about whether episodic memory—in particular autobio- graphical memory—exhibits any temporal gradient at all. For example, some recent studies performed on H.M. that were conducted not long before he passed away in 2008 showed that his memory for remote personal experiences, unlike his memory for remote factual knowledge, was not preserved (Steinvorth, Levine, & Corkin, 2005). In addition,
become independent of the hippocampus in a mat- ter of weeks. Controlled studies in humans sometimes suggest that the time course of systems consolidation often plays out over a much longer period of time period, a finding that is consistent with the time window of retrograde amnesia observed for H.M. However, the time course is rather variable, and the basis for the variability is not known. Both semantic and epi- sodic memory have been assessed in studies investi- gating the temporal gradient of retrograde amnesia. Semantic memory refers to memory for factual knowledge (e.g., what is the capital of Texas?), whereas episodic memory refers to memory for spe- cific events (e.g., memory for a recently presented list of words or memory for an autobiographical event, such as a trip to the Bahamas).
Temporal Gradient of Semantic Memory
Semantic knowledge is generally acquired gradu- ally across multiple episodes of learning and is forgot- ten slowly, so it seems reasonable to suppose that the systems consolidation of such knowledge would be extended in time. In one recent study of this issue, Manns et al. (2003) measured factual knowledge in six amnesic patients with damage limited to the hippocampal region. Participants were asked ques- tions about news events that had occurred from 1950 to early 2002 (e.g., Which tire manufacturer recalled thousands of tires? [Firestone] What soft- ware company was accused of running a monopoly? [Microsoft]). The data for a particular patient (and for several controls matched to that patient) were analyzed according to the year in which the patient became amnesic. As might be expected, memory for factual knowledge was reduced for the period of time following the onset of memory impairment. Thus, for example, if a patient became amnesic in 1985, then memory for news events that occurred after 1985 was impaired (i.e., anterograde amnesia was observed). In addition, and more to the point, factual knowledge for news events that occurred during the several years immediately before the onset of memory impairment (e.g., 1980 to 1985) was also impaired, particularly when memory was measured by free recall rather than by recognition. However, memory for events that occurred 11 to 30 years before the onset of memory impairment (e.g., 1955 to 1974) was intact. These older memories, it seems, had become fully consoli- dated in the neocortex and were no longer dependent on the structures of the MTL. Th e fi ndings from lesion studies have some- times been corroborated by neuroimaging studies
440 memory consolidation
the hippocampus and temporal neocortex associated with memory for two sets of paired-associate figures that subjects had previously memorized. One set was studied 8 weeks before the memory test (old memories), and the other was studied immediately before the memory test (new memories). Overall accuracy at the time of test was equated for the two conditions by providing extra study time for the items that were studied 8 weeks before. Thus, any differences in activity associated with old and new memories could not be attributed to differences in memory strength. The results showed that a region in right hippocampus was associated with greater activity during retrieval of new memories than old memories, whereas in left temporal neocortex, the opposite activation pattern (i.e., old > new) was observed. These results are consistent with a decreas- ing role of the hippocampus and increasing role of the neocortex as memories age over a period as short as 50 days (cf. Takashima et al., 2006), a time scale of consolidation that is similar to that observed in experimental animals (e.g., Anagnostaras et al., 1999). An even shorter time scale for systems consolida- tion was evident in a recent fMRI study reported by Takashima et al. (2009). Subjects in that study memorized two sets of face–location stimuli, one studied 24 hours before the memory test (old mem- ories) and the other studied 15 minutes before the memory test (new memories). To control for differ- ences in memory strength, they compared activity for high-confidence hits associated with the old and new memories and found that hippocampal activ- ity decreased and neocortical activity increased over the course of 24 hours. In addition, the connectiv- ity between the hippocampus and the neocortical regions decreased, whereas cortico-cortical con- nectivity increased (all over the course of only 24 hours). Results like these suggest that the process of systems consolidation can occur very quickly. What determines whether the temporal gradi- ent is short or long? The answer is not known, but Frankland and Bontempi (2006) suggested that the critical variable may be the richness of the memo- rized material. To-be-remembered stimuli presented in a laboratory are largely unrelated to a subject’s personal history and thus might be integrated with prior knowledge represented in the cortex in a rather sparse (yet rapid) manner. Autobiographical memories, by contrast, are generally related to a large preexisting knowledge base. The integra- tion of such memories into an intricate knowledge base may require more extended dialogue between
a number of case studies of memory-impaired patients have reported impairments of childhood memories (Cipolotti et al., 2001; Eslinger, 1998; Hirano & Noguchi, 1998; Kitchener et al., 1998; Maguire et al., 2006; Rosenbaum et al., 2004). Th ese findings appear to suggest that the MTL plays a role in recalling personal episodes even if they happened long ago, and the apparent impli- cation is that autobiographical memories do not undergo systems consolidation. However, by the time H.M.’s remote autobiographical memory impairment was documented, he was an elderly patient, and his brain was exhibiting signs of corti- cal thinning, abnormal white matter, and subcorti- cal infarcts (Squire, 2009). Thus, these late-life brain abnormalities could account for the loss of remote memories. In addition, in most of the case stud- ies that have documented remote autobiographical memory impairment, damage was not restricted to the MTL. To compare the remote memory effects of limited MTL damage and damage that also involved areas of the neocortex, Bayley et al. (2005) measured the ability of eight amnesic patients to recollect detailed autobiographical memories from their early life. Five of the patients had damage lim- ited to the MTL, whereas three had damage to the neocortex in addition to MTL damage. They found that the remote autobiographical memories of the fi ve MTL patients were quantitatively and quali- tatively similar to the recollections of the control group, whereas the autobiographical memories of the three patients with additional neocortical dam- age were severely impaired. Th is result suggests that semantic memory and episodic memory both even- tually become independent of the MTL through a process of systems consolidation, but the tempo- ral gradient of retroactive amnesia associated with that process can be obscured if damage extends to the neocortex. MacKinnon and Squire (1989) also found that the temporal gradient of autobiographi- cal memories for five MTL patients was similar in duration to the multiyear gradient associated with semantic memory.
Temporal Gradients Involving a Shorter
Time Scale
Recent neuroimaging studies have documented a temporal gradient of activity for memory of sim- ple laboratory stimuli on a time scale that is vastly shorter than the multiyear process of consolidation suggested by lesion studies of semantic memory and autobiographical memory. For example, using fMRI, Yamashita et al. (2009) measured activity in
442 memory consolidation
What does LTP have to do with the story of consolidation? The induction of LTP unleashes a molecular cascade in postsynaptic neurons that con- tinues for hours and results in structural changes to those neurons. The postsynaptic changes are pro- tein-synthesis dependent and involve morphologi- cal changes in dendritic spines (Yuste & Bonhoeffer,
- and the insertion of additional AMPA recep- tors into dendritic membranes (Lu et al., 2001). Th ese changes are generally thought to stabilize LTP because LTP degrades rapidly if they do not occur (or are prevented from occurring by the use of a protein synthesis inhibitor). LTP exhibits all of the characteristics of consoli- dation envisioned by Müller and Pilzecker (1900). In their own work, Müller and Pilzecker (1900) used an original learning phase (L1), followed by an interfering learning phase (L2), followed by a memory test for the original list (T1). Holding the retention interval between L1 and T1 constant, they essentially showed that L1-L2-----T1 yields greater interference than L1---L2---T (where the dashes represent units of time). In experimental animals, memories formed in the hippocampus and LTP induced in the hippocampus both exhibit a similar temporal gradient with respect to retroac- tive interference (Izquierdo et al., 1999; Xu et al., 1998). Whether L1 and L2 both involve hippocam- pus-dependent learning tasks (e.g., L1 = one-trial inhibitory avoidance learning, L2 = exploration of a novel environment), as reported by Izquierdo et al. (1999), or one involves the induction of LTP (L1) while the other involves exposure to a learning task (L2), as reported by Xu et al. (1998), the same pattern emerges. Specifically, L2 interferes with L when the time between them is relatively short (e.g., 1 hour), but not when the time between them is relatively long (e.g., 6 or more hours). Moreover, if an NMDA antagonist is infused into the hippocam- pus before L2 (thereby blocking the induction of interfering LTP that might be associated with the learning of a potentially interfering task), no inter- ference effect is observed even when the L1-L2 tem- poral interval is short. Th e point is that hippocampus-dependent memories and hippocampal LTP both appear to be vulnerable to interference early on and then become more resistant to interference with the passage of time. Moreover, the interfering force is the formation of new memories (or, analogously, the induction of LTP). Newly induced LTP, like a newly encoded memory, begins life in a fragile state. Over time, as the process of cellular consolidation
LTP is a relatively long-lasting enhancement of synaptic efficacy that is induced by a brief burst of high-frequency electrical stimulation (a tetanus) deliv- ered to presynaptic neurons (Bliss & Collingridge, 1993). Before the tetanus, a single (weak) test pulse of electrical stimulation applied to the presynaptic neuron elicits a certain baseline response in the post- synaptic neuron, but after the tetanus, that same test pulse elicits a greater response. The enhanced reac- tivity typically lasts hours or days (and sometimes weeks), so it presumably does not represent the way in which memories are permanently coded. Still, LTP is readily induced in hippocampal neurons, and it is, by far, the leading approach to modeling the neural basis of initial memory formation (Bliss, Collingridge, & Morris, 2003; Martin, Grimwood, & Morris, 2000). In this model, tetanic stimulation is analogous to the effect of a behavioral experience, and the enhanced efficacy of the synapse is analo- gous to the memory of that experience. Although LTP looks like neural memory for an experience (albeit an artificial experience consisting of a train of electrical impulses), what reason is there to believe that a similar process plays a role in real memories? The induction of LTP in hippocampal neurons involves the opening of calcium channels in postsynaptic N -methyl- D -aspartate (NMDA) receptors (Bliss & Collingridge, 1993). When those receptors are blocked by an NMDA antagonist, high-frequency stimulation fails to induce LTP. Perhaps not coincidentally, NMDA antagonists have often been shown to impair the learning of hippocampus-dependent tasks in animals (e.g., Morris et al., 1986; Morris, 1989), as if an LTP-like process in the hippocampus plays an important role in the formation of new episodic memories. One study suggests that the encoding of actual memo- ries (not just an artificial train of electrical pulses) also gives rise to LTP in the hippocampus. Whitlock et al. (2006) trained rats on an inhibitory avoidance task (a task known to be dependent on the hip- pocampus), and they were able to find neurons in the hippocampus that exhibited sustained LTP after training (not after an artificial tetanus). In addition, tetanic stimulation applied to these neurons after training now had a lesser effect (as if those neurons were already close to ceiling levels of LTP) than tetanic stimulation applied to the neurons of ani- mals who had not received training. These findings suggest that LTP may be more than just a model for memory formation; it may, in fact, be part of the mechanism that underlies the initial encoding of memory.
wixted, cai 443
again formed. If so, then less forgetting should be observed than would otherwise be the case. All of these findings are easily understood in terms of cellular consolidation (not systems consoli- dation), but a recent explosion of research on the role of sleep and consolidation has begun to suggest that the distinction between cellular consolidation and systems consolidation may not be as sharp as previously thought.
Sleep and Consolidation
In recent years, the idea that sleep plays a spe- cial role in the consolidation of both declarative and nondeclarative memory has received a great deal of attention. Declarative memory consists of the con- scious remembrance of either factual information (i.e., semantic memory) or past experience (i.e., episodic memory), and it is the kind of memory that we have discussed thus far in connection with systems consolidation and cellular consolidation. Nondeclarative memory, on the other hand, refers to the acquisition and retention of nonconscious skills and abilities, with the prototypical example being the ability to ride a bike. With practice, one’s riding ability improves, but the memory of how to balance on two wheels is not realized by consciously remembering anything about the past (as in the case of declarative memory). Instead, that memory is realized by climbing on the bike and discovering that you can ride it without falling off. Whereas declarative memory depends on the structures of the MTL, nondeclarative memories do not (Squire, 1992; Squire & Zola, 1996). As a result, amnesic patients with MTL damage have an impairment of declarative memory (both anterograde amnesia and temporally graded retrograde amnesia), but they are generally unimpaired at learning and retaining pro- cedural skills (Squire, 1992). An amnesic could, for example, learn to ride a bike as easily as you could, but, unlike you, the amnesic would have no con- scious declarative memory of the practice sessions. Recent research suggests that sleep plays a role in the consolidation of both declarative and nondeclara- tive memories. Because sleep is not an undifferentiated state, one focus of this line of research has been to iden- tify the specific stage of sleep that is important for consolidation. Sleep is divided into five stages that occur in a regular sequence within 90-minute cycles throughout the night. Stages 1 through 4 refer to ever-deeper levels of sleep, with stages 3 and 4 often being referred to as slow-wave sleep. Rapid eye movement (REM) sleep is a lighter stage of sleep
unfolds, recently formed LTP and recently encoded memories become more stable, which is to say that they become more resistant to interference caused by the induction of new LTP or by the encoding of new memories. The use of an NMDA antagonist in rats is not the only way to induce a temporary period of anterograde amnesia (thereby protecting recently induced LTP or recently formed memories). In sufficient quantities, alcohol and benzodiazepines have been shown to do the same in humans. Moreover, like NMDA antagonists, these drugs not only induce anterograde amnesia but also inhibit the induction of LTP in the hippocampus (Del Cerro et al., 1992; Evans & Viola-McCabe, 1996; Givens & McMahon, 1995; Roberto et al., 2002, Sinclair & Lo, 1986). Interestingly, they also result in a phenomenon known as retrograde facili- tation. That is, numerous studies have reported that even though alcohol induces amnesia for informa- tion studied under the influence of the drug, it actually results in improved memory for material studied just before consumption (e.g., Bruce & Pihl, 1997; Lamberty, Beckwith, & Petros, 1990; Mann, Cho-Young, & Vogel-Sprott, 1984; Parker et al., 1980, 1981). Similar findings have been frequently reported for benzodiazepines such as diazepam and triazolam (Coenen & Van Luijtelaar, 1997; Fillmore et al., 2001; Ghoneim, Hinrichs, & Mewaldt, 1984; Hinrichs, Ghoneim, & Mewaldt, 1984; Weingartner et al., 1995). Predrug memo- ries, it seems, are protected from interference that would have been created during the postdrug amnesic state. It is important to emphasize that postlearning amnesia-inducing agents (such as NMDA antago- nists used in rats or alcohol and benzodiazepines used in humans) do not enhance predrug memo- ries in an absolute sense. That is, in response to these drugs, the memories do not more accurately represent past experience and are not more likely to be retrieved than they were at the end of learn- ing. Instead, memories formed before drug intake are forgotten to a lesser degree than memories formed before placebo. By limiting the formation of new memories, alcohol and benzodiazepines (like NMDA antagonists) may protect memories that were formed just before drug intake. While protected from the trace-degrading force of new memory formation, these memories may be allowed to consolidate (via cellular consolidation) in a way that hardens them against the interference they will later encounter when new memories are once
wixted, cai 445
phenomenon of neural replay was initially observed in hippocampal cells of sleeping rats after they had run along a familiar track, and its discovery was tied to the earlier discovery of place cells in the hippocampus. Long ago, it was discovered that the firing of par- ticular hippocampal cells in awake rats is coupled to specific points in the rat’s environment (O’Keefe & Dostrovsky, 1971). These cells are known as “place cells” because they fire only when the rat traverses a particular place in the environment. Usually, hip- pocampal place cells fire in relation to the rat’s posi- tion on a running track. That is, as the rat traverses point A along the track, place cell 1 will reliably fire. As it traverses point B, place cell 2 will fire (and so on). An intriguing finding that may be relevant to the mechanism that underlies systems consolida- tion is that cells that fire in sequence in the hip- pocampus during a behavioral task tend to become sequentially coactive again during sleep (Wilson & McNaughton, 1994). This is the phenomenon of neural replay. Neural reply has most often been observed in rats during slow-wave sleep. It has also occasion- ally been observed during REM sleep, but, in that case, it occurs at a rate that is similar to the neu- ron firing that occurred during learning (Louie & Wilson, 2001) and thus may simply reflect dream- ing. The neural replay that occurs during slow-wave sleep occurs at a rate five to ten times faster than it did during the waking state (e.g., Ji & Wilson,
- and may therefore reflect a biological con- solidation process separate from mental activity like dreaming. It is as if the hippocampus is replaying the earlier behavioral experience, perhaps as a way to reorganize the representation of that experience in the neocortex. Th e fact that replay of sequential place cell activ- ity in the hippocampus occurs during slow-wave sleep does not, by itself, suggest anything about communication between the hippocampus and the neocortex (the kind of communication that is pre- sumably required for systems consolidation to take place). However, Ji and Wilson (2007) reported that hippocampal replay during slow-wave sleep in rats was coordinated with firing patterns in the visual cortex, which is consistent with the idea that this process underlies the reorganization of mem- ories in the neocortex. In addition, Lansink et al. (2009) performed multineuron recordings from the hippocampus and ventral striatum during wak- ing and sleeping states. While the rats were awake, the hippocampal cells fired when the rat traversed a
prediction would be that L1-S-----T1 will confer greater protection than L1---S---T1. If a temporal gradient is observed (i.e., if memory performance at T1 is greater in the first condition than the second), it would suggest that sleep does more than simply subtract out a period of retroactive interference that would otherwise occur. Instead, it would suggest that sleep (presumably slow-wave sleep) also allows the process of cellular consolidation to proceed in the absence of interference. Once again, Ekstrand (1972) performed the pioneering experiment on this issue. In that experi- ment, memory was tested for paired-associate words following a 24-hour retention interval in which sub- jects slept either during the 8 hours that followed list presentation or during the 8 hours that preceded the recall test. In the immediate sleep condition (in which L1 occurred at night, just before sleep), he found that 81 percent of the items were recalled 24 hours later; in the delayed sleep condition (in which L1 occurred in the morning), only 66 per- cent were recalled. In other words, a clear temporal gradient associated with the subtraction of retroac- tive interference was observed, one that is the mir- ror image of the temporal gradient associated with the addition of retroactive interference reported by Müller and Pilzecker (1900). More recent sleep studies have reinforced the idea that the temporal gradient of retrograde facilitation is a real phenom- enon, and they have addressed various confounds that could have accounted for the results that Ekstrand (1972) obtained (Gais, Lucas, & Born, 2006; Talamini et al., 2008). The temporal gradient associated with sleep, like the LTP and animal learn- ing research described earlier, is consistent with the notion that when memory formation is temporarily halted, recently formed and still-fragile memories are protected from interference. As a result, they are given a chance to become hardened against the forces of retroactive interference that they will later encounter (perhaps through a process of cellular consolidation).
Slow-Wave Sleep and Systems
Consolidation
Recent sleep studies have also shed light on the mechanism that may account for systems consoli- dation, which presumably involves some relatively long-lasting form of communication between the hippocampus and the neocortex (Marr, 1971). Th e mechanism of communication is not known, but a leading candidate is neural replay , and most of the work on this topic comes from sleep studies. Th e
446 memory consolidation
In both these studies, the hippocampal reactiva- tion (perhaps reflective of hippocampo-neocortical dialogue) occurred within hours of the learning episode, a time course of consolidation ordinarily associated with cellular consolidation. The timing observed in these studies is not unlike that observed in a neuroimaging study discussed earlier in which hippocampal activity decreased, and neocorti- cal activity increased, over a period as short as 24 hours (Takashima et al., 2009). Moreover, the tim- ing fits with studies in rats showing that learning- related neural replay is evident in the first slow-wave sleep episode that follows learning (Peyrache et al., 2009). In a sleep-deprivation study that also points to an almost immediate role for systems-level con- solidation processes, Sterpenich et al. (2009), using human subjects, investigated memory for emo- tional and neutral pictures 6 months after encod- ing. Half the subjects were deprived of sleep on the first postencoding night, and half were allowed to sleep (and then all subjects slept normally each night thereafter). Six months later, subjects com- pleted a recognition test in the scanner in which each test item was given a judgment of “remem- ber” (previously seen and subjectively recollected), “know” (previously seen but not subjectively recol- lected), or “new” (not previously seen). A contrast between activity associated with remembered items and known items yielded a smaller difference in the sleep-deprived subjects across a variety of brain areas (ventral mPFC, precuneus, amygdala, and occipital cortex), even though the items had been memorized 6 months earlier, and these results were interpreted to mean that sleep during the first postencoding night influences the long-term systems-level con- solidation of emotional memory. Th e unmistakable implication from all of these studies is that the process thought to underlie sys- tems consolidation—namely, neural replay (or neu- ral reactivation)—begins to unfold in a measurable way along a time course ordinarily associated with cellular consolidation. That is, in the hours after a trace is formed, hippocampal LTP stabilizes, and neural replay in the hippocampus gets underway. Th ese findings would seem to raise the possibility that the molecular cascade that underlies cellular consolidation also plays a role in initiating neural replay (Mednick, Cai, Shuman, Anagnostaras, & Wixted, 2011). If interference occurs while the trace is still fragile, then LTP will not stabilize, and pre- sumably, neural replay will not be initiated. In that case, the memory will be lost. But if hippocampal
particular point in the environment (i.e., they were place cells), whereas the striatal cells generally fired in response to rewards. During slow-wave sleep (but not during REM sleep), they found that the hip- pocampal and striatal cells reactivated together. The coordinated firing was particularly evident for pairs in which the hippocampal place cell fired before the striatal reward-related neuron. Thus, the hippocam- pus leads reactivation in a projection area, and this mechanism may underlie the systems consolidation of place–reward associations. One concern about studies of neural replay is that the animals are generally overtrained, so little or no learning actually occurs. Thus, it is not clear whether learning-related neural replay takes place. However, Peyrache et al. (2009) recorded neurons in prefrontal cortex during the course of learning. Rats were trained on a Y-maze task in which they learned to select the rewarded arm using one rule (e.g., choose the left arm) that changed to a differ- ent rule as soon as a criterion level of performance was achieved (e.g., choose the right arm). They identified sets of neuronal assemblies with reliable coactivations in prefrontal cortex, and some of these coactivations became stronger when the rat started the first run of correct trials associated with the acquisition of the new rule. Following these sessions, replay during slow-wave sleep mainly involved the learning-related coactivations. Thus, learning-related replay—the mechanism that may underlie systems consolidation—can be identi- fied and appears to get underway very soon after learning. Other evidence suggests that something akin to neural replay occurs in humans as well. An intrigu- ing study by Rasch et al. (2007) showed that cuing recently formed odor-associated memories by odor re-exposure during slow-wave sleep—but not dur- ing REM sleep—prompted hippocampal activation (as measured by fMRI) and resulted in less forget- ting after sleep compared with a control group. Th is result is consistent with the notion that sys- tems consolidation results from the reactivation of newly encoded hippocampal representations during slow-wave sleep. In a conceptually related study, Peigneux et al. (2004) measured regional cerebral blood flow and showed that hippocampal areas that were activated during route learning in a virtual town (a hippocampus-dependent, spatial learning task) were activated again during subsequent slow- wave sleep. Moreover, the degree of activation dur- ing slow-wave sleep correlated with performance on the task the next day.
448 memory consolidation
within the hippocampal-entorhinal output network, and synchronized neural discharges tend to occur along this pathway during sharp-wave/ripple events (Buzsáki, 1986; Chrobak & Buzsáki, 1996). Thus, once again, rhythmic activity seems to coordinate communication between adjacent brain structures, and such communication has been found to occur between more distant brain structures as well. For example, ripples observed during hippocampal sharp waves have been correlated with the occurrence of spindles in prefrontal cortex (Siapas & Wilson, 1998). Moreover, the neural replay discussed earlier preferentially takes place during the high-frequency bursts of spindle waves (Wilson & McNaughton, 1994). All of this suggests that rhythmically based feedback activity from the hippocampus may serve to “train” the neocortex and thus facilitate the pro- cess of systems consolidation. When it occurs in the hours after learning, this kind of systems-level communication presumably involves hippocampal neurons that have encoded information and that are successfully undergoing the process of cellular consolidation. If so, then, again, cellular consolida- tion could be regarded as an early component of the systems consolidation process.
Sleep-Related Consolidation of
Nondeclarative Memory
A novel line of research concerned with the role of sleep in consolidation was initiated by a study suggesting that sleep also plays a role in the con- solidation of nondeclarative memories. Karni et al. (1994) presented subjects with computer-generated stimulus displays that sometimes contained a small target consisting of three adjacent diagonal bars (arranged either vertically or horizontally) embed- ded within a background of many horizontal bars. Th e displays were presented very briefly (10 ms) and then occluded by a visual mask, and the subject’s job on a given trial was to indicate whether the tar- get items were arranged vertically or horizontally in the just-presented display. Performance on this task improves with practice in that subjects can correctly identify the target with shorter and shorter delays between the stimulus and the mask. Th e detection of element orientation differences in these visual displays is a preattentive process that occurs rapidly and automatically (i.e., no deliber- ate search is required). In addition, the learning that takes place with practice presumably reflects plasticity in the early processing areas of the visual cortex, which would account for why the learning is extremely specific to the trained stimuli (e.g., if the
the encoding and consolidating states are also asso- ciated with characteristic rhythmic activity, and a basic assumption of this account is that communi- cation between the hippocampus and neocortex is mediated by coordinated oscillatory rhythms across different structures of the brain (Sirota, Csicsvari, Buhl, & Buzsáki, 2003). In the encoding state, the cortex is characterized by beta oscillations (i.e., 12 to 20 Hz), whereas the hippocampus is characterized by theta oscillations (i.e., 4 to 8 Hz). Hippocampal theta oscillations are thought to synchronize neural firing along an input pathway into the hippocampus. For example, in the presence of theta (but not in its absence), the hippocampus receives rhythmic input from neu- rons in the input layers of the adjacent entorhinal cortex (Chrobak & Buzsáki, 1996). In addition, Siapas, Lubenov, and Wilson (2005) showed that neural activity in the prefrontal cortex of rats was “phase-locked” to theta oscillations in the hip- pocampus in freely behaving (i.e., active-awake) rats. Findings like these are consistent with the idea that theta rhythms coordinate the flow of informa- tion into the hippocampus, and still other findings suggest that theta rhythms may facilitate the encod- ing of information flowing into the hippocampus. During the high-Ach encoding state—which is a time when hippocampal synaptic plasticity is high (Rasmusson, 2000)—electrical stimuli delivered at intervals equal to theta frequency are more likely to induce LTP than stimulation delivered at other frequencies (Larson & Lynch, 1986). Thus, theta appears to play a role both in organizing the flow of information into the hippocampus and in facilitat- ing the encoding of that information. Lower levels of Ach prevail during quite-awake and slow-wave sleep, and this is thought to shift the hippocampus into the consolidating state (see Rasch, Born, & Gais, 2006). In this state, activity along input pathways (ordinarily facilitated by theta rhythms) is suppressed, and hippocampal plasticity is low (i.e., hippocampal LTP is not readily induced). As such, and as indicated earlier, recently induced LTP is protected from interference and is given a chance to stabilize as the process of cellular consoli- dation unfolds. In addition, under these conditions, the cortex is characterized by low-frequency spindle oscillations (i.e., 7 to 14 Hz) and delta oscillations (i.e., 4 Hz or less), whereas the hippocampus is asso- ciated with a more broad-spectrum pattern punc- tuated by brief, high-frequency sharp waves (i.e., 30 Hz or more) and very-high-frequency “ripples” (about 200 Hz). Th ese sharp wave oscillations occur
wixted, cai 449
such as the sequential finger-tapping task (Walker et al., 2002, 2003a, 2003b). In this task, subjects learn a sequence of finger presses, and performance improves with training (i.e., the sequence is com- pleted with increasing speed) and improves still further following a night of sleep, with the degree of improvement often correlating with time spent in stage 2 sleep. Fischer, Hallschmid, Elsner, and Born (2002) reported similar results, except that performance gains correlated with amount of REM sleep. However, one aspect of this motor-sequence- learning phenomenon—namely, the fact that per- formance improves beyond what was observed at the end of training—has been called into question. Rickard et al. (2008) recently presented evidence suggesting that the apparent absolute enhance- ment of performance on this task following sleep may have resulted from a combination of averag- ing artifacts, time-of-day confounds (cf. Keisler, Ashe, & Willingham, 2007; Song et al., 2007), and the buildup of fatigue (creating the impression of less presleep learning than actually occurred). This result does not necessarily question the special role of sleep in the consolidation of motor-sequence learning, but it does call into question the absolute increase in performance that has been observed fol- lowing a period of sleep. Somewhat more puzzling for the idea that REM plays a special role in the consolidation of nondeclarative memory is that Rasch, Pommer, Diekelmann, and Born (2008) found that the use of antidepressant drugs, which virtually eliminate REM sleep, did not eliminate the apparent sleep- related enhancement of performance on two non- declarative memory tasks (mirror tracing and motor sequence learning). This result would appear to sug- gest that REM sleep, per se, is not critical for the consolidation of learning on either task. Instead, conditions that happen to prevail during REM sleep (rather than REM sleep per se) may be criti- cal. Consistent with this possibility, Rasch, Gais, and Born (2009) showed that cholinergic receptor blockade during REM significantly impaired motor skill consolidation. This fi nding suggests that the consolidation of motor skill depends on the high cholinergic activity that typically occurs during REM (and that presumably occurs even when REM is eliminated by antidepressant drugs). What consolidation mechanism is responsible for sleep-related enhancement of performance on perceptual learning tasks? Hippocampal replay discussed earlier seems like an unlikely candidate because this is not a hippocampus-dependent
targets always appear in one quadrant of the screen during training, no transfer of learning is appar- ent when the targets are presented in a different quadrant). Thus, the visual segregation task is not a hippocampus-dependent task involving conscious memory (i.e., it is not a declarative memory task); instead, it is a nondeclarative memory task. A remarkable finding reported by Karni et al. (1994; Karni & Sagi, 1993) was that, following a night of normal sleep, performance improved on this task to levels that were higher than the level that had been achieved at the end of training—as if further learning took place offline during sleep. Th is is unlike what is typically observed on declara- tive memory tasks, which only rarely show an actual performance enhancement. Various control condi- tions showed that the enhanced learning effect was not simply due to a reduction in general fatigue. Instead, some kind of performance-enhancing con- solidation apparently occurred while the subjects slept. Karni et al. (1994) found that depriving subjects of slow-wave sleep after learning did not prevent the improvement of postsleep performance from occurring, but depriving them of REM sleep did. Th us, REM sleep seems critical for the sleep-related enhancement of procedural learning to occur, and similar results have been reported in a number of other studies (Atienza et al., 2004; Gais et al., 2000; Mednick et al., 2002, 2003; Stickgold, James, & Hobson, 2000; Walker et al., 2005). These fi nd- ings have been taken to mean that nondeclarative memories require a period of consolidation and that REM sleep in particular is critical for such consoli- dation to occur. Although most work has pointed to REM, some work has suggested a role for slow-wave sleep as well. For example, using the same texture- discrimination task, Stickgold et al. (2000) found that the sleep-dependent gains were correlated with the amount of slow-wave sleep early in the night and with the amount of REM sleep late in the night (cf. Gais et al., 2000). In the case of nondeclarative memories, the evi- dence for consolidation does not consist of decreas- ing dependence on one brain system (as in systems consolidation) or of increasing resistance to interfer- ence (as in cellular consolidation). Instead, the evi- dence consists of an enhancement of learning beyond the level that was achieved at the end of training. At the time Karni et al. (1994) published their findings, this was an altogether new phenomenon, and it was followed by similar demonstrations of sleep-related enhancement using other procedural memory tasks,
wixted, cai 451
more persistent in the consolidation group com- pared with the reconsolidation group. Still, this study adds to a large and growing literature showing that reactivated memories are in some way vulner- able in a way that was not fully appreciated until Nader et al. (2000) drove the point home with their compelling study.
Conclusion
Th e idea that memories require time to consoli- date was proposed more than a century ago, but empirical inquiry into the mechanisms of consoli- dation is now more intense than ever. With that inquiry has come the realization that the issue is complex, so much so that, used in isolation, the word “consolidation” no longer has a clear meaning. One can speak of consolidation in terms of memory becoming less dependent on the hippocampus (sys- tems consolidation) or in terms of a trace becoming stabilized (cellular consolidation). Alternatively, one can speak of consolidation in terms of enhanced performance (over and above the level of perfor- mance achieved at the end of training), in terms of increased resistance to interference (i.e., less for- getting), or in terms of a presumed mechanism, such as neural replay or neural reactivation. A clear implication is that any use of the word consolidation should be accompanied by a statement of what it means. Similarly, any suggestion that consolidation “strengthens” the memory trace should be accom- panied by a clear statement of the way (or ways) in which the trace is thought be stronger than it was before. A more precise use of the terminology commonly used in this domain of investigation will help to make sense of the rapidly burgeoning lit- erature on the always fascinating topic of memory consolidation.
References
Anagnostaras S. G., Maren S., & Fanselow M. S. ( 1999 ). Temporally-graded retrograde amnesia of contextual fear after hippocampal damage in rats: Within-subjects examina- tion. Journal of Neuroscience , 19 , 1106 –1114. Anderson , J. R., & Schooler, L. J. ( 1991 ). Reflections of the envi- ronment in memory. Psychological Science , 2 , 396 –408. Atienza , M., Cantero , J. L., & Stickgold , R. ( 2004 ) Posttraining sleep enhances automaticity in perceptual discrimination. Journal of Cognitive Neuroscience , 16 , 53 –64. Barrett , T. R., & Ekstrand , B. R. ( 1972 ). Effect of sleep on memory: III. Controlling for time-of-day effects. Journal of Experimental Psychology , 96 , 321 –327. Bayley, P. J. Gold , J. J., Hopkins , R. O., & Squire , L. R. ( 2005 ). The neuroanatomy of remote memory. Neuron , 46 , 799 –810. Bernard , F. A., Bullmore , E. T., Graham , K. S., Thompson , S. A., Hodges , J. R., & Fletcher, P. C. ( 2004 ). The hippocampal
All these results parallel the effects of anisomycin on tone–shock memory when it is infused after a con- ditioning trial (Schafe & LeDoux, 2000). What was remarkable about the Nader et al. (2000) results was that similar consolidation effects were also observed well after conditioning and in response to the reacti- vation of memory caused by the presentation of the tone. Similar findings have now been reported for other tasks and other species (see Nader & Hardt, 2009, for a review). Th e notion that a consolidated memory becomes fragile again merely because it is reactivated might seem implausible because personal experience does not suggest that we place our memories at risk by retrieving them. In fact, the well-known testing eff ect —the finding that successful retrieval enhances memory more than additional study—seems to sug- gest that the opposite may be true (e.g., Roediger & Karpicke, 2006). However, a fragile trace is also a malleable trace, and it has been suggested that the updating of memory—not its erasure—may be a benefi t of what otherwise seems like a problem- atic state of affairs. As noted by Dudai (2004), the susceptibility to corruption of a retrieved memory “might be the price paid for modifiability” (p. 75). If the reactivated trace is susceptible only to agents such as anisomycin, which is not a drug that is encountered on a regular basis, then the price for modifiability might be low indeed. On the other hand, if the trace is vulnerable to corruption by new learning, as a newly learned memory trace appears to be, then the price could be considerably higher. In an intriguing new study, Monfils, Cowansage, Klann, and LeDoux (2009) showed that contextual fear memories in rats can be more readily eliminated by extinction trials if the fear memory is first reac- tivated by a reminder trial. For the first time, this raises the possibility that reactivated memories are vulnerable to disruption and modification by new learning (not just by protein synthesis inhibitors). Much remains unknown about reconsolidation, and there is some debate as to whether the disrup- tion of a recently retrieved trace is a permanent or a transient phenomenon. For example, Stafford and Lattal (2009) recently compared the eff ects of anisomycin administered shortly after fear con- ditioning (which would disrupt the consolidation of a new memory) or shortly after a reminder trial (which would disrupt the consolidation of a newly retrieved memory). With both groups equated on important variables such as prior learning experi- ence, they found that the anisomycin-induced defi- cit on a test of long-term memory was larger and
452 memory consolidation
Ekstrand , B. R. ( 1972 ). To sleep, perchance to dream (about why we forget). In C. P. Duncan, L. Sechrest, & A. W. Melton (Eds.), Human memory: Festschrift for Benton J. Underwood (pp. 59 –82). New York : Appelton-Century-Crofts. Eslinger, P. J. ( 1998 ). Autobiographical memory after temporal lobe lesions. Neurocase , 4 , 481 –495. Evans , M. S., & Viola-McCabe , K. E. ( 1996 ). Midazolam inhib- its long-term potentiation through modulation of GABAA receptors. Neuropharmacology , 35 , 347 –357. Fillmore , M. T., Kelly, T. H., Rush , C. R., & Hays , L. ( 2001 ). Retrograde facilitation of memory by triazolam: Effects on automatic processes. Psychopharmacology , 158 , 314 –321. Fischer, S., Hallschmid , M., Elsner, A. L., & Born , J. ( 2002 ). Sleep forms memory for finger skills. Proceedings of the National Academy of Sciences U S A , 99 , 11987 –11991. Frankland , P. W., & Bontempi , B. ( 2005 ). The organization of recent and remote memory. Nature Reviews Neuroscience , 6 , 119 –130. Frankland , P. W., & Bontempi , B. ( 2006 ). Fast track to the medial prefrontal cortex. Proceedings of the National Academy of Sciences U S A , 103 , 509 –510. Frankland , P. W., Bontempi , B., Talton , L. E., Kaczmarek , L., & Silva , A. J. ( 2004 ). The involvement of the anterior cingu- late cortex in remote contextual fear memory. Science , 304 , 881 –883. Gais , S., Lucas , B., & Born , J. ( 2006 ). Sleep after learning aids memory recall. Learning and Memory , 13 , 259 –262. Gais , S., Plihal , W., Wagner, U., Born , J. ( 2000 ). Early sleep triggers memory for early visual discrimination skills. Nature Neuroscience , 3 , 1335 –1339. Ghoneim , M. M., Hinrichs , J. V., & Mewaldt , S. P. ( 1984 ). Dose-response analysis of the behavioral effects of diazepam: I. Learning and memory. Psychopharmacology , 82 , 291 –295. Givens , B., & McMahon , K. ( 1995 ). Ethanol suppresses the induction of long-term potentiation in vivo. Brain Research , 688 , 27 –33. Haist , F., Bowden Gore , J., & Mao , H. ( 2001 ). Consolidation of human memory over decades revealed by functional magnetic resonance imaging. Nature Neuroscience , 4 , 1139 –1145. Hasselmo , M. E. ( 1999 ) Neuromodulation: Acetylcholine and memory consolidation. Trends in Cognitive Sciences , 3 , 351 –359. Hinrichs , J. V., Ghoneim , M. M., & Mewaldt , S. P. ( 1984 ). Diazepam and memory: Retrograde facilitation produced by interference reduction. Psychopharmacology , 84 , 158 –162. Hirano , M., & Noguchi , K. ( 1998 ). Dissociation between spe- cific personal episodes and other aspects of remote memory in a patient with hippocampal amnesia. Perceptual and Motor Skills , 87 , 99 –107. Hoffman , K. L., & McNaughton , B. L. ( 2002 ). Coordinated reactivation of distributed memory traces in primate neocor- tex. Science , 297 , 2070 –2073. Izquierdo , I., Schröder, N., Netto , C. A., & Medina , J. H. ( 1999 ). Novelty causes time-dependent retrograde amnesia for one-trial avoidance in rats through NMDA receptor- and CaMKII-dependent mechanisms in the hippocampus. European Journal of Neuroscience , 11 , 3323 –3328. Jenkins , J. B., & Dallenbach , K. M. ( 1924 ). Oblivescence dur- ing sleep and waking. American Journal of Psychology , 35 , 605 –612. Ji , D., & Wilson , M. A. ( 2007 ). Coordinated memory replay in the visual cortex and hippocampus during sleep. Nature Neuroscience , 10 , 100 –107.
region is involved in successful recognition of both remote and recent famous faces. NeuroImage , 22 , 1704 –1714. Bliss , T. V. P., & Collingridge , G. L. ( 1993 ). A synaptic model of memory: Long-term potentiation in the hippocampus. Nature , 361 , 31 –39. Bliss , T. V. P., Collingridge , G. L., & Morris , R. G. ( 2003 ). Long- term potentiation: enhancing neuroscience for 30 years – Introduction. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences , 358 , 607 –611. Bliss , T. V. P., & Lomo , T. ( 1973 ). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. Journal of Physiology , 232 , 331 –356. Bramham , C. R., & Srebo , B. ( 1989 ). Synaptic plasticity in the hippocampus is modulated by behavioral state. Brain Research , 493 , 74 –86. Bruce , K. R., & Pihl , R. O. ( 1997 ). Forget “drinking to forget”: Enhanced consolidation of emotionally charged memory by alcohol. Experimental and Clinical Psychopharmacology , 5 , 242 –250. Buzsáki, G. (1986). Hippocampal sharp waves: Their origin and significance. Brain Research , 398 , 242 –252. Buzsáki , G. ( 1989 ). A two-stage model of memory trace for- mation: A role for “noisy” brain states. Neuroscience , 31 , 551 –570. Cai ., D. J., Mednick ., S. A., Harrison , E. M., Kanady, J. C., & Mednick , S. C. ( 2009 ). REM, not incubation, improves creativity by priming associative networks. Proceedings of the National Academy of Sciences U S A , 106 , 10130 –10134. Cai , D. J., Shuman , T., Gorman , M. R., Sage , J. R., & Anagnostaras , S. G. ( 2009 ). Sleep selectively enhances hippocampus-dependent memory in mice. Behavioral Neuroscience , 123 , 713 –719. Chrobak , J. J., & Buzsáki , G. ( 1996 ) High-frequency oscilla- tions in the output networks of the hippocampal-entorhinal axis of the freely-behaving rat. Journal of Neuroscience , 16 , 3056 –3066. Cipolotti , L., Shallice , T., Chan , D., Fox , N., Scahill , R., Harrison , G., Stevens , J., & Rudge , P. ( 2001 ). Long-term retrograde amnesia: the crucial role of the hippocampus. Neuropsychologia , 39 , 151 –172. Coenen , A. M. L., & Van Luijtelaar, E. L. J. M. ( 1997 ). Effects of benzodiazepines, sleep and sleep deprivation on vigilance and memory. Acta Neurologica Belgica , 97 , 123 –129. Damasio , A.R. ( 1989 ). Time-locked multiregional retroactiva- tion: A systems-level proposal for the neural substrates of recall and recognition. Cognition , 33 , 25 –62. Del Cerro , S., Jung , M., & Lynch , L. ( 1992 ). Benzodiazepines block long-term potentiation in slices of hippocampus and piriform cortex. Neuroscience , 49 , 1 –6. Douville , K., Woodard , J. L., Seidenberg , M., Miller, S. K., Leveroni , C. L., Nielson , K. A., Franczak , M., Antuono , P., & Rao , S. M. ( 2005 ). Medial temporal lobe activity for recognition of recent and remote famous names: An event related fMRI study. Neuropsychologia , 43 , 693 –703. Dudai , Y. ( 2004 ). The neurobiology of consolidations, or, how stable is the engram? Annual Review of Psychology , 55 , 51 – 86. Ebbinghaus , H. ( 1885 ). Über das Gedchtnis. Untersuchungen zur experimentellen Psychologie. Leipzig : Duncker & Humblot. English edition: Ebbinghaus , H. ( 1913 ). Memory: A contribu- tion to experimental psychology. New York : Teachers College, Columbia University.
454 memory consolidation
synthesis and protein kinase A in the amygdala. Journal of Neuroscience , 20 , RC96 , 1–5. Scoville , W. B., & Milner, B. ( 1957 ). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery and Psychiatry , 20 , 11 –21. Siapas , A. G., Lubenov, E. V., & Wilson , M. A., ( 2005 ). Prefrontal phase locking to hippocampal theta oscillations. Neuron , 46 , 141 –151. Siapas , A. G., & Wilson , M. A. ( 1998 ). Coordinated interactions between hippocampal ripples and cortical spindles during slow-wave sleep. Neuron , 21 , 1123 –1128. Sinclair, J. G., & Lo , G. F. ( 1986 ). Ethanol blocks tetanic and calcium-induced long-term potentiation in the hippocampal slice. General Pharmacology , 17 , 231 –233. Sirota , A., Csicsvari , J. Buhl , D., & Buzsáki , G. ( 2003 ). Communication between neocortex and hippocampus dur- ing sleep in rats and mice. Proceedings of the National Academy of Sciences U S A , 100 , 2065 –2069. Smith , C. N., & Squire , L. R. ( 2009 ). Medial temporal lobe activity during retrieval of semantic memory is related to the age of the memory. Journal of Neuroscience , 29 , 930 –938. Song , S. S., Howard , J. H., Jr., & Howard , D. V. ( 2007 ). Sleep does not benefit probabilistic motor sequence learning. Journal of Neuroscience , 27 , 12475 –12483. Squire , L. R. ( 1992 ) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychological Review , 99 , 195 –231. Squire , L. R. ( 2009 ). Th e legacy of patient H.M. for neurosci- ence. Neuron , 61 , 6 –9. Squire, L. R., & Alvarez, P. (1995). Retrograde amnesia and memory consolidation: A neurobiological perspective. Current Opinion in Neurobiology , 5 , 169 –177. Squire , L. R., Clark , R. E., & Knowlton , B. J. ( 2001 ). Retrograde amnesia. Hippocampus , 11 , 50 –55. Squire , L. R., & Wixted , J. T. ( 2011 ). The cognitive neuro- science of human memory since H.M. Annual Review of Neuroscience , 34 , 259 –288. Squire , L. R., & Zola , S. M. ( 1996 ). Structure and function of declarative and nondeclarative memory systems. Proceedings of the National Academy of Sciences U S A , 93 , 13515 –13522. Staff ord , J. M., & Lattal , K. M. ( 2009 ). Direct comparisons of the size and persistence of anisomycin-induced consolida- tion and reconsolidation deficits. Learning and Memory , 16 , 494 –503. Steinvorth , S., Levine , B., & Corkin , S. ( 2005 ). Medial tem- poral lobe structures are needed to re-experience remote autobiographical memories: evidence from H.M. and W.R. Neuropsychologia , 43 , 479 –496. Sterpenich , V., Albouy, G., Darsaud , A., Schmidt , C., Vandewalle , G., Dang Vu , T. T., Desseilles , M., Phillips , C., Degueldre , C., Balteau , E., Collette , F., Luxen , A., & Maquet , P. ( 2009 ). Sleep promotes the neural reorganization of remote emo- tional memory. Journal of Neuroscience , 16 , 5143 –5152. Stickgold , R., James , L., & Hobson , J. A. ( 2000 ). Visual dis- crimination learning requires sleep after training. Nature Neuroscience , 3 , 1237 –1238. Takashima , A., Nieuwenhuis , I. L. C., Jensen , O., Talamini , L. M., Rijpkema, M., & Fernández. G. (2009). Shift from hip- pocampal to neocortical centered retrieval network with con- solidation. Journal of Neuroscience , 29 , 10087 –10093. Takashima , A., Petersson , K. M., Rutters , F., Tendolkar, I., Jensen, O, Zwarts, M. J., McNaughton, B. L., & Fernández, G. ( 2006 ). Declarative memory consolidation in humans: A
O’Keefe , J., & Dostrovsky, J. ( 1971 ). The hippocampus as a spatial map: Preliminary evidence from unit activity in the freely-moving rat. Brain Research , 34 , 171 –175. Parker, E. S., Birnbaum , I. M., Weingartner, H., Hartley, J. T., Stillman , R. C., & Wyatt , R. J. ( 1980 ). Retrograde enhance- ment of human memory with alcohol. Psychopharmacology , 69 , 219 –222. Parker, E. S., Morihisa , J. M., Wyatt , R. J., Schwartz , B. L., Weingartner, H., & Stillman , R. C. ( 1981 ). The alco- hol facilitation eff ect on memory: A dose-response study. Psychopharmacology , 74 , 88 –92. Peigneux , P., Laureys , S., Fuchs , S., Collette , F., Perrin , F., et al. ( 2004 ). Are spatial memories strengthened in the human hip- pocampus during slow wave sleep? Neuron , 44 , 535 –545. Peyrache , A., Khamassi , M., Benchenane , K., Wiener, S. I., & Battaglia , F. P. ( 2009 ). Replay of rule-learning related neu- ral patterns in the prefrontal cortex during slee p. Nature Neuroscience , 12 , 919 –926. Plihal , W., & Born , J. ( 1997 ). Effects of early and late noctur- nal sleep on declarative and procedural memory. Journal of Cognitive Neuroscience , 9 , 534 –547. Plihal, W., & Born, J. (1999). Effects of early and late nocturnal sleep on priming and spatial memory. Psychophysiology , 36 , 571 –582. Rasch , B. H., Born , J., & Gais , S. ( 2006 ). Combined blockade of cholinergic receptors shifts the brain from stimulus encoding to memory consolidation. Journal of Cognitive Neuroscience , 18 , 793 –802. Rasch , B., Buchel , C., Gais , S., & Born , J. ( 2007 ). Odor cues during slow-wave sleep prompt declarative memory consoli- dation. Science , 315 , 1426 –1429. Rasch , B., Gais , S., & Born , J. ( 2009 ) Impaired off-line con- solidation of motor memories after combined blockade of cholinergic receptors during REM sleep-rich sleep. Neuropsychopharmacology , 34 , 1843 –1853. Rasch , B., Pommer, J., Diekelmann , S., & Born , J. ( 2008 ). Pharmacological REM sleep suppression paradoxi- cally improves rather than impairs skill memory. Nature Neuroscience , 12 , 396 –397. Rasmusson, D. D. (2000). The role of acetylcholine in cortical synaptic plasticity. Behavioural Brain Research , 115 , 205–218. Ribot , T. ( 1881 ). Les maladies de la memoire [ Diseases of memory ]. New York : Appleton-Century-Crofts. Ribot , T. ( 1882 ). Diseases of memory: An essay in positive psychol- ogy. London : Kegan Paul, Trench & Co. Rickard , T. C., Cai , D. J., Rieth , C. A., Jones , J., & Ard , M. C. (2008). Sleep does not enhance motor sequence learning. Journal of Experimental Psychology: Learning, Memory, and Cognition , 34 , 834 –842. Roberto , M., Nelson , T. E., Ur, C. L., & Gruol , D. L. ( 2002 ). Long-term potentiation in the rat hippocampus is reversibly depressed by chronic intermittent ethanol exposure. Journal of Neurophysiology , 87 , 2385 –2397. Roediger, H. L., & Karpicke , J.D. ( 2006 ). Test-enhanced learn- ing: Taking memory tests improves long-term retention. Psychological Science , 17 , 249 –255. Rosenbaum , R.S., McKinnon , M. C., Levine , B., & Moscovitch , M. ( 2004 ). Visual imagery deficits, impaired strategic retrieval, or memory loss: Disentangling the nature of an amnesic person’s autobiographical memory deficit. Neuropsychologia , 42 , 1619 –1635. Schafe , G. E., & LeDoux , J. E. ( 2000 ). Memory consolidation of auditory Pavlovian fear conditioning requires protein
wixted, cai 455
Weingartner, H. J., Sirocco , K., Curran , V., & Wolkowitz , O. ( 1995 ). Memory facilitation following the administration of the benzodiazepine triazolam. Experimental and Clinical Psychopharmacology , 3 , 298 –303. Whitlock J. R., Heynen A. J., Schuler M. G., & Bear M. F. ( 2006 ). Learning induces long-term potentiation in the hip- pocampus. Science , 313 , 1058 –1059. Wickelgren , W. A. ( 1974 ). Single-trace fragility theory of mem- ory dynamics. Memory and Cognition , 2 , 775 –780. Wilson , M. A., & McNaughton , B. L. ( 1994 ). Reactivation of hippocampal ensemble memories during sleep. Science , 265 , 676 –679. Wixted , J. T. ( 2004a ). The psychology and neuroscience of for- getting. Annual Review of Psychology , 55 , 235 –269. Wixted , J. T. ( 2004b ). On common ground: Jost’s (1897) law of forgetting and Ribot’s (1881) law of retrograde amnesia. Psychological Review , 111 , 864 –879. Wixted , J. T., & Carpenter, S. K. ( 2007 ). The Wickelgren power law and the Ebbinghaus savings function. Psychological Science , 18 , 133 –134. Wixted , J. T., & Ebbesen , E. ( 1991). On the form of forgetting. Psychological Science , 2 , 409 –415. Xu , L., Anwyl , R., & Rowan , M. J. ( 1998 ). Spatial exploration induces a persistent reversal of long-term potentiation in rat hippocampus. Nature , 394 , 891 –894. Yamashita , K., Hirose , S., Kunimatsu , A., Aoki , S., Chikazoe , J., Jimura , K., Masutani , Y., Abe , O., Ohtomo , K., Miyashita , Y., & Konishi , S. ( 2009 ). Formation of long-term memory representation in human temporal cortex related to pictorial paired associates. Journal of Neuroscience , 29 , 10335 –10340. Yaroush , R., Sullivan , M. J., & Ekstrand , B. R. ( 1971 ). The eff ect of sleep on memory: II. Differential effect of the first and second half of the night. Journal of Experimental Psychology , 88 , 361 –366. Yuste , R., & Bonhoeffer, T. ( 2001 ). Morphological changes in dendritic spines associated with long-term synaptic plasticity. Annual Review of Neuroscience , 24 , 1071 –1089.
prospective functional magnetic resonance imaging study. Proceedings of the National Academy of Sciences U S A , 103 , 756 –761. Takehara , K., Kawahara , S., & Kirino , Y. ( 2003 ). Time- dependent reorganization of the brain components underly- ing memory retention in trace eyeblink conditioning. Journal of Neuroscience , 23 , 9897 –9905. Takehara-Nishiuchi , K., & McNaughton , B. L. ( 2008 ). Spontaneous changes of neocortical code for associative memory during consolidation. Science , 322 , 960 –963. Talamini , L. M., Nieuwenhuis , I. L., Takashima , A., & Jensen , O. ( 2008 ). Sleep directly following learning benefits consoli- dation of spatial associative memory. Learning and Memory , 15 , 233 –237. Tse, D., Langston, R. F., Kakeyama, M., Bethus, I., Spooner, P. A., Wood, E. R., Witter, M. P., & Morris, R. G. (2007). Schemas and memory consolidation. Science , 316 , 76–82. Underwood , B. J. ( 1957 ). Interference and forgetting. Psychological Review , 64 , 49 –60. Wagner U., Gais , S., Haider, H., Verleger, R., & Born , J. ( 2004 ) Sleep inspires insight. Nature , 427 , 352 –355. Walker, M. P., Brakefi eld , T., Hobson , J. A., & Stickgold , R. ( 2003a ). Dissociable stages of human memory consolidation and reconsolidation. Nature , 425 , 616 –620. Walker, M. P., Brakefi eld , T., Morgan , A., Hobson , J. A., & Stickgold , R. ( 2002 ). Practice with sleep makes perfect: Sleep- dependent motor skill learning. Neuron , 35 , 205 –211. Walker, M. P., Brakefield , T., Seidman , J., Morgon , A., Hobson , J. A., & Stickgold , R. ( 2003b ). Sleep and the time course of motor skill learning. Learning and Memory , 10 , 275 –284. Walker, M. P., Stickgold , R., Jolesz , F. A., & Yoo , S. S. ( 2005 ). The functional anatomy of sleep-dependent visual skill learn- ing. Cerebral Cortex , 15 , 1666 –1675. Watkins , C., & Watkins , M. J. ( 1975 ). Buildup of proactive inhibition as a cue-overload effect. Journal of Experimental Psychology: Human Learning and Memory , 1 , 442 –452.
