Molecular Neurobiology

Copyright © 2005 Humana Press Inc.

All rights of any nature whatsoever reserved.
ISSN0893-7648/05/32(2): 123–132/$30.00

Removing Pathogenic Memories

A Neurobiology of Psychotherapy

Diego Centonze,1,2,* Alberto Siracusano,3

Paolo Calabresi,2,4 and Giorgio Bernardi1,2
1Clinica Neurologica, Dipartimento di Neuroscienze, Università Tor Vergata,
2Fondazione Santa Lucia, Centro Europeo per la Ricerca sul Cervello, and
3Clinica Psichiatrica, Dipartimento di Neuroscienze, Università Tor Vergata,

Rome, Italy; 4Clinica Neurologica, Università di Perugia, Perugia, Italy

Abstract
Experimental research examining the neural bases of nondeclarative memory has offered
intriguing insight into how functional and dysfunctional implicit learning affects the brain. Long-
term modifications of synaptic transmission, in particular, are currently considered the most plau-
sible mechanism underlying memory trace encoding and compulsions, addiction, anxiety, and
phobias. Therefore, an effective psychotherapy must be directed to erase maladaptive implicit
memories and aberrant synaptic plasticity.
This article describes the neurobiological bases of pathogenic memory disruption to provide
some insight into how psychotherapy works. At least two mechanisms of unwanted memory
erasing appear to be implicated in the effects of psychotherapy: inhibition of memory consolida-
tion/reconsolidation and extinction. Behavioral evidence demonstrated that these two ways to
forget are profoundly distinct in nature, and it is increasingly clear that their cellular, synaptic,
and molecular underpinnings are different. Accordingly, the blockade of consolidation/reconsol-
idation erases memories by reversing the plasticity associated with memory maintenance,
whereas extinction is a totally new form of plasticity that, similar to the plasticity underlying the
old memory, requires protein synthesis-dependent synaptic remodeling.

Index Entries: Extinction; forgetting; long-term depression; long-term potentiation; reconsol-

idation.

Received December 21, 2004; Accepted March 2, 2005.

*Author to whom all correspondence and reprint requests should be addressed. E-mail: [email protected]

Molecular Neurobiology 123 Volume 32, 2005

124 Centonze et al.

“I only know that such ‘oblivion’ was not ganglia loop (8–11). Conversely, addictive drugs
reached by the patients I analysed, but that instead abuse the molecular mechanisms of reward-
it led to various pathological reactions” (Freud, based associative learning by inducing long-
1894, The Neuro-Psychoses of Defence). term changes in synaptic effectiveness in those
brain areas serving basic biological needs, such
“The hysteric suffers from reminiscences” as feeding and sexual interaction (midbrain
(Freud, 1895, Studies on Hysteria. Freud, 1909, dopamine neurons and several brain structures
Five Lectures on Psycho-Analysis). receiving dopamine-releasing axon terminals)
(12–15). Finally, anxiety, panic disorder, and
Introduction phobias are viewed as uncontrolled and repeti-
tive defensive reactions secondary to abnormal
In Freud’s theory on the origin of neuroses, fear conditioning (a form of implicit associa-
real or imaginary traumatic experiences per- tive learning, encoded as long-term potentiation
manently alter the functioning of the psychic [LTP] in the lateral amygdala in which emotion-
apparatus. Therefore, unconscious memories ally neutral stimuli acquire the capacity to elicit
can be pathogenic: “All hysterics and neurotics defensive responses after association with an
remember the painful experiences of their dis- aversive event) (16–18).
tant past…; they are unable to free themselves Therefore, a main corollary of these consid-
from the past and thus neglect reality and the erations is that an effective psychotherapy
present. This fixation of psychic life for patho- must be directed to erase maladaptive implicit
genic traumas is one of the most important memories and aberrant synaptic plasticity.
characteristics and practically the most signifi- This article discusses the processes of mem-
cant of neurosis” (1). ory disruption following behavioral treatment.
It is increasingly accepted that many psychi- The cellular and molecular mechanisms of the
atric symptoms originate from the formation process of “forgetting” are also described to
and consolidation of implicit dysfunctional provide biological insight into how psychother-
memories. Implicit, or nondeclarative, learn- apy works.
ing refers to the ability of the brain to establish
unconscious associations between distinct
events. Indeed, learning by association is an
Inhibition of Consolidation/
essential process for retaining information Reconsolidation and Extinction
from past experiences to reinforce homeostatic Are Two Distinct Ways to Forget
behaviors and escape from aversive condi-
tions (2–4). Over the past decade, remarkable progress
Experimental research examining the neural has been made toward our understanding of
bases of nondeclarative memory (such as habit the mechanisms of memory formation and
formation, classical conditioning, and fear con- consolidation. More recently, attention has also
ditioning) has offered intriguing insight into focused on the process of forgetfulness. The
how functional and dysfunctional implicit learn- emerging scenario is particularly interesting,
ing affects the brain. Long-term modifications of because it is increasingly clear that forgetting
synaptic transmission in particular, are cur- does not simply result from the passive degra-
rently considered the most plausible mecha- dation of the plasticity associated with mem-
nisms underlying memory trace encoding ory acquisition; rather, it represents a very
compulsions, addiction, anxiety, and phobias complex process that actively counters the
(5–7). In this line, compulsions and other stereo- learning course.
typies are viewed as pathological habits (nearly At least two mechanisms of unwanted mem-
automated implicit motor abilities) encoded as ory disruption might be implicated in the
aberrant synaptic plasticity in the cortico-basal effects of psychotherapy: inhibition of memory

Molecular Neurobiology Volume 32, 2005

Removing Pathogenic Memories 125

consolidation/reconsolidation and extinction. is now increasingly accepted that it represents a

Both processes must be clearly distinguished phenomenon shared by different types of mem-
from repression, a process by which unac- ories (28). Thus, instead of occurring only once,
cepted memories are kept out of consciousness memory storage is a process that is reiterated
and, therefore, are not properly forgotten. Pro- with each use of the memory.
posed by Freud more than a century ago, the These discoveries substantially contributed to
psychological process of memory repression revision of the concept that long-term memories
has come to be clarified only in recent years, are irreversibly stored in the wiring of the brain
based on the demonstration of a series of psy- and, rather, indicated that memories can be con-
chological inhibitory mechanisms actively tinually updated, modified, and even erased.
recruited to prevent unwanted memories from Incidentally, the susceptibility of the reconsoli-
entering awareness (19). dation process to interference may account for
the occurrence of false (including pathogenic)
memories. Specifically, during re-activation of a
Inhibition of Memory memory trace, its content can be changed
Consolidation/Reconsolidation through suggestion or other means (32).
If the original memory traces become sus-
It has long been recognized that long-term ceptible to changes each time the memory is
information storage involves a process by retrieved and this is true also for pathogenic
which a labile short-term memory is converted memories, the psychotherapy treatment
into a stable, fixed trace that is particularly (which enables patients to recall and re-elabo-
impervious to amnesic treatments (20–22). This rate past traumatic experiences) can help to
view was originally proposed based on evi- extinguish their pathogenic potential by inter-
dence that treatments such as electroconvulsive fering with the reconsolidation process.
shock that produce amnesia shortly after learn- Notably, since the very beginning of his theo-
ing have no effect when delivered several rization, Freud discovered that memory re-
hours later (23). The transition from short- to activation and re-elaboration can be beneficial.
long-term memory is called consolidation and In Recollection, Repetition and Working Through
is believed to follow the transformation of tem- (1914), he wrote: “We know therefore that the
porary alterations in synaptic transmission into patient in analysis repeats instead of remem-
persistent modifications of synaptic architec- bering…From repetitive reactions, the known
ture in specific brain areas (7,24–26). This paths (psychoanalytical therapy) lead to the
model, which is more than 100 yr old (27,28), reawakening of memories” (33). In other
was initially challenged in a 1968 study show- terms, the pathogenic effects of dysfunctional
ing that although electroconvulsive shock was unconscious memories, which are responsible
ineffective in disrupting memory hours after for repetition typical of neurotic symptoms,
the learning process, it could induce amnesia if can be overcome by a process or memory trace
a fully consolidated and stable long-term mem- retrieval and re-elaboration.
ory was re-activated by a reminder process
prior to amnesic treatments (29). Therefore, this
finding indicates that old memory re-activation Memory Extinction
returns it to a labile state, which undergoes
another round of consolidation similar to that Although the inhibition of reconsolidation is
occurring after new learning. This process has a process of interference with the mechanism
been referred to as reconsolidation (30,31). In of memory restorage that follows the reminder,
recent years, the process of reconsolidation was extinction is a completely different phenome-
clearly demonstrated by using a large number non. Accordingly, inhibition of consolidation
of behavioral paradigms and amnesic agents; it as well as reconsolidation tends to render a

Molecular Neurobiology Volume 32, 2005

126 Centonze et al.

certain memory “unlearnt,” whereas extinction is generally more labile than the excitatory one,
is a new learning process that does not disrupt indicating why the extinguished response tends
previously acquired plasticity but, rather, trig- to re-appear with time or changing the environ-
gers additional plasticity. Experimentally, mental context (34,36–38). Therefore, the high
extinction consists in the progressive decrease rate of relapse observed in patients treated with
and eventual disappearance of a conditioned behavioral therapy might rely on this peculiar
response (CR; e.g., fear) when a neutral condi- feature of the learning process that is activated
tioned stimulus (CS; e.g., a light) predictive of by extinction.
a certain significant event (unconditioned Although distinct in nature, it is likely that
stimulus [US]; e.g., a footshock) is repeatedly the inhibition of consolidation/reconsolida-
presented alone. tion and extinction occur simultaneously dur-
Since the pioneering studies of Pavlov, ing disruption of maladaptive memories. In
extinction has been studied in several experi- fact, based on several behavioral findings, it
mental paradigms, including eyeblink condi- has been proposed that reconsolidation (which
tioning, several appetitive paradigms, and fear favors memory maintenance) and extinction
conditioning (34,35). Incidentally, the experi- (which facilitates memory inhibition) are two
mental protocol used to induce fear extinction processes competing for their expression fol-
in animals—that is, the iterative presentation of lowing memory re-activation. For example, it
the CS in the absence of the US—is very similar has been reported that far from inducing a
to the psychotherapeutic approach employed reconsolidation of the learned behavior, reit-
to treat fear disorders in humans, which com- eration of the CS alone solely causes extinc-
monly involves exposure to the feared object in tion of learned fear (39). Therefore, when
the absence of any overt danger. extinction is favored, reconsolidation is invari-
Several observations indicate that extinction ably inhibited (28,34,39,40). Although the
does not result from the progressive weaken- study of the factors favoring the emergence of
ing and ultimate degradation of the original a process over the other is still at its infancy, a
CS–US association. For example, extinction recent report proposed that reminder dura-
does not occur following simple passage of tion is a major determinant dictating whether
time but, conversely, is the extinction process reconsolidation or extinction is preferentially
that tends to dissipate, as indicated by the pro- activated (41).
gressive re-appearance of the extinguished
CRs over time (36). Similarly, an extinguished
CR re-appears when an animal (37) or a human Molecular and Synaptic Bases
(38) is tested in a context different from that in of Forgetfulness
which the extinction training occurred or when
unsignaled US are presented following the When studied at the molecular level, blockade
completion of the extinction training. of consolidation/reconsolidation and extinction
Together, these data indicate that extinction appear as two completely different phenomena.
does not permanently erase the ability of a CS to Intriguingly, although pharmacological inhibi-
drive a CR; rather, it triggers a second learning tion of protein synthesis favors the blockade of
process that inhibits the expression of the CR both consolidation and reconsolidation, this
without destroying the old memory, although it treatment prevents extinction. These findings
suppresses its expression (34). In other words, are generally interpreted as confirmation that
extinction is a process that generates a further the blockade of reconsolidation erases memories
inhibitory association between the CS and US, by reversing the plasticity associated with mem-
which actively counters to the tendency of the ory maintenance, whereas, extinction is a totally
CS to activate US representation. Notably, the new form of plasticity requiring protein synthe-
inhibitory association at the basis of extinction sis-dependent synaptic re-arrangements (similar

Molecular Neurobiology Volume 32, 2005

Removing Pathogenic Memories 127

to the plasticity underlying the old memory) dation and depotentiation is lacking, but in our
(28,34,39–41). opinion, some considerations exist for this asso-
ciation. First, consolidation and reconsolidation
are inhibited (48), and depotentiation is favored
Long-Term Potentiation of Excitatory (26,49) by blocking the transcription factor
Transmission As a Synaptic cyclic adenine monophosphate-response-ele-
Correlate of Memory ment binding protein. Second, memory trace
and LTP are susceptible to erasing processes
In 1973, Bliss and Lømo (42) reported the
(inhibition of reconsolidation and synaptic depo-
first finding that repetitive synaptic activation
tentiation, respectively) for only a short time fol-
induced a long-lasting increase in the efficacy
lowing memory acquisition (or retrieval) (28)
of excitatory transmission in the hippocampus,
and LTP induction (45,46). Finally, the process
a brain area known to be involved in learning
of extinction-independent forgetting is facili-
and memory processes. This phenomenon,
tated (50) in the same way as depotentiation
known as LTP, essentially consists of the
(47,51) by the activity of protein phosphatase 1,
enduring facilitation of the communication
an enzyme critical for the removal of phosphate
between two neurons in response to the sus-
groups from target proteins (52). Notably, the
tained activation of the synapses by which
stimulation of several protein kinases, which
they are interconnected. Several subsequent
conversely add phosphate groups on the same
scientific discoveries have made it possible to
target proteins, enables LTP formation (7).
correlate this synaptic phenomenon with
Depotentiation can apparently reset synap-
learning and memory (6,7,43). Notably, LTP
tic transmission to the naïve state, implying
has now been described at several other excita-
that memory trace disappears entirely follow-
tory synapses of the central nervous system,
ing inhibition of consolidation/reconsolidation.
including those of brain areas important for
However, recent studies have challenged this
implicit memory encoding.
view, indicating that the plasticity history of a
synapse is as important as its current physiolog-
Molecular and Synaptic ical state in determining its response to a given
Correlates of the Inhibition stimulation. An interesting study demonstrated
that excitatory synapses can occupy as many as
of Consolidation/Reconsolidation seven distinct states (silent, active, depressed,
If LTP subserves memory formation and recently potentiated, recently depotentiated,
consolidation, then it is likely that depotentia- recently unsilenced, and remotely unsilenced)
tion represents the synaptic underpinning of and that the lasting response to the same stim-
the process of inhibition of memory reconsoli- ulation protocol differs among them (53). With
dation. Depotentiation is a recently described respect to the plastic potential of a naïve and of
synaptic phenomenon consisting of the possi- a depotentiated synapse, low-frequency stimu-
bility to reverse a previously induced LTP by lation was found to induce LTD in both cases,
an appropriate protocol of synaptic stimula- but the depression observed in depotentiated
tion (44–47). Although depotentiation exhibits synapses was significantly smaller in amplitude
some similarities to long-term depression (53). Therefore, depotentiated synapses tend to
(LTD) of synaptic transmission, it represents a prefer the potentiated, rather than the depressed,
completely different phenomenon because the state (54). This evidence indicates that once
commonly employed depotentiation protocol employed for information storage, synapses
is unable to depress nonpotentiated synapses indefinitely retain a trace of this experience,
(43,47). even after LTP and memory erasing. This might
A definitive demonstration of the correlation provide an explanation for the evidence that
between inhibition of consolidation/reconsoli- memory disruption produced by blockade of

Molecular Neurobiology Volume 32, 2005

128 Centonze et al.

reconsolidation can recover (55) and for the clin- tex, cerebellum, deep cerebellar nucleus, lateral
ical experience that treated symptoms tend to superior olive, and brain stem (74); however, a
re-appear. Freud evocatively described the direct link between GABAergic plasticity and
impossible oblivion of many neurotics as: “the behavioral modifications was still missing. Sev-
eternal return of the same” (56,57). eral findings support the concept that iLTD lies
at the basis of the extinction process. For exam-
Molecular and Synaptic Correlates ple, NMDA glutamate receptor activation is
required for both extinction (34,60,64,65) and
of Extinction iLTD (74,75). Additionally, the recent evidence
As reported earlier, general agreement exists that extinction of cocaine-seeking behavior is
regarding the fact that extinction is a new learn- associated with upregulation of α-amino-3-
ing process. Not surprisingly, therefore, the hydroxy-5-methyl-4-isoxazole propionic acid
molecular determinants of memory extinction glutamate receptors (76) is also consistent with
parallel those of synaptic plasticity and memory the finding that these receptors mediate the ini-
formation. For example, the stimulation of N- tial depolarization during the conditioning pro-
methyl-D-aspartate (NMDA) glutamate recep- tocol of iLTD (77). Finally, pharmacological
tors is a critical requirement for amygdalal LTP agents that reduce the effectiveness of GABA
(58,59) and fear memory acquisition (60–63); transmission (such iLTD) have been found to
however, it has also been found to be crucial for facilitate the extinction process (78).
fear extinction (60,64,65). Additionally, the The idea that long-term inhibition of GABA
postreceptor events initiated by NMDA recep- transmission is a synaptic correlate of extinc-
tor stimulation and mediating LTP and memory tion also helps to explain the critical role
formation are substantially identical to those played by endocannabinoids in this process.
involved in extinction. Accordingly, protein Pharmacological or genetic inactivation of
kinase A, calcium/calmodulin-dependent pro- cannabinoid CB1 receptors prevents fear
tein kinase II, and mitogen-activated protein extinction, whereas the levels of two major
kinase are all required for LTP (43,66–69) as well endocannabinoids, anandamide and 2-arachi-
as memory formation (68,70,71) and extinction donylglycerol, increase significantly in the
(65,72). Finally, extinction, similarly to memory basolateral amygdala during fear extinction
consolidation (28) and the late phase of LTP (7), training (73). Therefore, evidence exists that
requires new protein synthesis (34). behavioral treatment of aversive memories
Although these data are all suggestive of the results in the activation of the endocannabi-
fact that extinction recapitulates the molecular noid system, which, in turn, facilitates mem-
and synaptic events at the basis of memory ory extinction by stimulating cannabinoid
and plasticity of glutamatergic synapses, an CB1 receptors in the amygdala. Notably,
attractive alternative hypothesis is that extinc- among the different effects of endocannabi-
tion-associated plasticity does not involve noids in central neurons, the short- and long-
glutamatergic transmission but, rather, γ-amino- term inhibition of GABAergic transmission is
butyric acid (GABA)ergic inhibitory synapses. particularly relevant. Cannabinoid CB1 recep-
A recent study showed that impairment of tors are abundantly expressed on GABAergic
fear extinction is associated with the disrup- nerve terminals in the brain, where they
tion of LTD of GABAergic transmission (iLTD) reduce transmitter release (79,80). Endo-
in the basolateral amygdala (73), implying this cannabinoids participate in the inhibitory
form of synaptic plasticity in aversive memory control of GABA transmission by acting as
elimination. retrograde messengers—that is, they are
Synaptic plasticity of GABA synapses has released from the somatodendritic region of
been reported to occur in several other regions the postsynaptic neuron and back propagate
of the brain, including the hippocampus, cor- to inhibit presynaptic GABA-releasing nerve

Molecular Neurobiology Volume 32, 2005

Removing Pathogenic Memories 129

terminals (81). This mechanism of action has patients perceive danger in situations that are
been proposed to explain two distinct synap- not dangerous; paranoid patients may miscon-
tic phenomena: depolarization-induced sup- strue situations in terms of being deceived or
pression of inhibition (DSI) and iLTD. Whereas attacked; depressed patients see evidence of
DSI consists in a transient inhibition of GABA personal defect in situations that offer no
release triggered by sustained depolarization objective reasons for self-deprecation. In other
of the postsynaptic neuron (82,83), iLTD is a words, once connected, two or more distinct
long-lasting phenomenon (74). The different representations proceed together and behave
time-courses of these two processes likely as one.
reflect the duration of endocannabinoid release. Therefore, in the face of the diverse technical
Accordingly, whereas DSI is presumably trig- approaches, the progressive dissolution of
gered by a brief (a few seconds) release of such pathogenic associations should be the
endocannabinoids (83,84), cannabinoid CB1 purpose of all psychotherapeutic treatments.
receptors must be activated for minutes to Notably, in 1922, Freud, in defining psychoan-
trigger iLTD (85). This is consistent with the alytic treatment, wrote: “The patient’s symp-
finding that fear-extinction-triggered iLTD in toms and pathological manifestations—like all
the amygdala is associated with an increase in his psychic activities—are of a highly compos-
endocannabinoids detectable even after the ite character […]. But the patient is either
killing of the experimental animals and brain unaware of these factors or has only an insuffi-
removal (73). cient knowledge of them. Therefore, we teach
him to understand the composition of these
very complicated psychic formations […], that
Conclusion is, we act like a chemist who isolates a simple
substance or chemical ‘element’ from the salt
Many psychiatric symptoms follow abnor- in which it has combined with other elements
malities of brain’s representation of the outside and thus become unrecognisable” (1).
world as a consequence of dysfunctional
unconscious associations between external
stimuli. At the end of the 19th century, Ramon
and Cajal (86) and Freud (87) proposed that References
information could be stored by modifying 1. Freud S. (1909) Five Lectures on Psycho-Analy-
interneuronal connections—a principle formal- sis. In: The Standard Edition of the Complete Psy-
ized by Hebb 50 yr later: “When an axon of cell chological Works of Sigmund Freud, vol. 11,
A is near enough to excite a cell B and repeat- Strachey J., ed. London: Hogarth Press, 1953,
edly and persistently takes part in firing it, pp. 3–55.
some growth process or metabolic change takes 2. Kandel E. R. (1999) Biology and the future of
place in one or both cells such that A’s efficacy, psychoanalysis: a new intellectual framework
as one of the cells firing B, is increased” (20). for psychiatry revisited. Am. J. Psychiatry 156,
Indeed, Hebb’s rule is believed to account 505–524.
for the associations of neural events. In fact, 3. Schall J. D. (2001) Neural basis of deciding,
with respect to the classical Pavlovian condi- choosing and acting. Nat. Rev. Neurosci. 2, 33–42.
4. Glimcher P.W. (2002) Decisions, decisions, deci-
tioning, if the activity of cell A is part of the
sions: choosing a biological science of choice.
representation of a neutral stimulus (e.g., a Neuron 36, 323–332.
sound) and the activity of cell B signals food, 5. Bliss T. V. and Collingridge G. L. (1993) A
once the connection between A and B has been synaptic model of memory: long-term potentia-
strengthened, each time A is activated by the tion in the hippocampus. Nature 362, 31–39.
sound, the representation of food will appear. 6. Martin S. J., Grimwood P. D., and Morris R. G.
Because of this kind of association, anxious M. (2000) Synaptic plasticity and memory: an

Molecular Neurobiology Volume 32, 2005

130 Centonze et al.

evaluation hypothesis. Annu. Rev. Neurosci. 23, 24. Davis H. P. and Squire L. R. (1984) Protein syn-
649–711. thesis and memory. A review. Psychol. Bull. 96,
7. Kandel E. R. (2001) The molecular biology of 518–559.
memory storage: a dialogue between genes and 25. Goelet P., Castellucci V. F., Schacher S., and
synapses. Science 294, 1030–1038. Kandel E. R. (1986) The long and short of long-
8. Calabresi P., De Murtas M., and Bernardi G. term memory—a molecular framework. Nature
(1997) The neostriatum beyond the motor func- 322, 419–422.
tion: experimental and clinical evidence. Neuro- 26. Bourtchuladze R., Frenguelli B., Cioffi D.,
science 78, 39–60. Blendy J., Schutz G., and Silva A. (1994) Defi-
9. Graybiel A. M. and Rauch S. L. (2000) Toward a cient long-term potentiation in mice with a tar-
neurobiology of obsessive-compulsive disorder. geted mutation of cAMP-responsive element
Neuron 28, 343–347. binding protein. Cell 79, 59–68.
10. Leckman J. F. and Riddle M. A. (2000) Tourette’s 27. Müller G. E. and Pilzecker A. (1900) Experi-
syndrome: when habit-forming systems form mentelle beitrage zur lehre vom gedachtnis. Z.
habits of their own? Neuron 28, 349–354. Psychol. 1, 117–131.
11. Holden C. (2001) “Behavioral” addictions: do 28. Nader K. (2003) Memory traces unbound.
they exist? Science 294, 980–982. Trends Neurosci. 26, 65–72.
12. Berke J. D. and Hyman S. E. (2000) Addiction, 29. Misanin J. R., Miller R. R., and Lewis D. J. (1968)
dopamine, and the molecular mechanisms of Retrograde amnesia produced by electrocon-
memory. Neuron 25, 515–532. vulsive shock after reactivation of a consoli-
13. Nestler E. J. (2001) Molecular basis of long-term dated memory trace. Science 160, 203–204.
plasticity underlying addiction. Nat. Rev. Neu- 30. Spear N. (1973) Retrieval of memory in animals.
rosci. 2, 119–128. Psychol. Rev. 80, 163–194.
14. Ungless M. A., Whistler J. L., Malenka R. C., 31. Przybyslawski J. and Sara S. J. (1997) Reconsoli-
and Bonci A. (2001) Single cocaine exposure in dation of memory after its reactivation. Behav.
vivo induces long-term potentiation in dopa- Brain Res. 84, 241–246.
mine neurons. Nature 411, 583–587. 32. Loftus E. F. and Yuille J. C. (1984) Departures
15. Gerdeman G. L., Partridge J. G., Lupica C. R., from reality in human perception and memory.
and Lovinger D. M. (2003) It could be habit In: Memory Consolidation: Psychobiology of Cogni-
forming: drugs of abuse and striatal synaptic tion, Weingartner H., and Parker E.S., eds. Hills-
plasticity. Trends Neurosci. 26, 184–192. dale, NJ: Lawrence Erbaum Associates, pp.
16. Rogan M. T., Stäubli U. V., and LeDoux J. E. 163–184.
(1997) Fear conditioning induces associative 33. Freud S. (1914) Recollection, Repetition and
long-term potentiation in the amygdala. Nature Working Through.In: The Standard Edition of the
390, 604–607. Complete Psychological Works of Sigmund Freud,
17. Bauer E. P., LeDoux J. E., Nader K. (2001) Fear vol. 12, Strachey J., ed. London: Hogarth Press,
conditioning and LTP in the lateral amygdala 1953, pp. 147–158.
are sensitive to the same stimulus contingen- 34. Myers K. M. and Davis M. (2002) Behavioral
cies. Nat. Neurosci. 4, 687,688. and neural analysis of extinction. Neuron 36,
18. Maren S. (2001) Neurobiology of Pavlovian fear 567–584.
conditioning. Annu. Rev. Neurosci. 24, 897–931. 35. Beggs J. M., Brown T. H., Byrne J. H., et al.
19. Anderson M. C. and Green C. (2001) Suppress- (1999) Learning and memory: basic mecha-
ing unwanted memories by executive control. nisms. In: Fundamental Neuroscience, Zigmond
Nature 410, 366–369. M. J., Bloom F. E., Landis S. C., Roberts J. L.,
20. Hebb D. O. (1949) Organization of Behavior: a and Squire L. R., eds. London: Academic Press,
Neuropsychological Theory. New York: John pp. 1411–1454.
Wiley and Sons Press. 36. Robbins S. J. (1990) Mechanisms underlying
21. McGaugh J. L. (1966) Time-dependent processes spontaneous recovery in autoshaping. J. Exp.
in memory storage. Science 153, 1351–1358. Psychol. Anim. Behav. Process 16, 235–249.
22. McGaugh J. L. (2000) Memory—a century of 37. Bouton M. E. and Bolles R. C. (1979) Contextual
consolidation. Science 287, 248–251. control of the extinction of conditioned fear.
23. Duncan C. P. (1949) The retroactive effect of Learn. Motiv. 10, 455–466.
electroconvulsive shock. J. Comp. Physiol. Psy- 38. Rodriguez B. I., Craske M. G., Mineka S., and
chol. 42, 32–44. Hladek D. (1999) Context-specificity of relapse:

Molecular Neurobiology Volume 32, 2005

Removing Pathogenic Memories 131

effects of therapist and environmental context 52. Silva A. J. and Josselyn S. A. (2002) The mole-
on return to fear. Behav. Res. Ther. 37, 845–862. cules of forgetfulness. Nature 418, 929,930.
39. Izquierdo I., Cammarota M., Vianna M. M., and 53. Montgomery J. M. and Madison D. V. (2002)
Bevilaqua L. R. (2004) The inhibition of State-dependent heterogeneity in synaptic
acquired fear. Neurotox. Res. 6, 175–188. depression between pyramidal cell pairs. Neu-
40. Milekic M. H., and Alberini C. M. (2002) Tem- ron 33, 765–777.
porally graded requirement for protein synthe- 54. Philpot B.D., and Bear M.F. (2002) Synaptic
sis following memory reactivation. Neuron 36, plasticity in an altered state. Neuron 33,
521–525. 665–671.
41. Pedreira M. E. and Maldonado H. (2003) Pro- 55. Anokhin K. V., Tiunova A. A., and Rose S. P.
tein synthesis subserves reconsolidation or (2002) Reminder effects—reconsolidation or
extinction depending on reminder duration. retrieval deficit? Pharmacological dissection
Neuron 38, 863–869. with protein synthesis inhibitors following
42. Bliss T. V. and Lømo T. (1973) Long-lasting reminder for a passive-avoidance task in young
potentiation of synaptic transmission in the chicks. Eur. J. Neurosci. 15, 1759–1765.
dentate area of the anaesthetized rabbit follow- 56. Freud S. (1919) The Uncanny. In: The Standard
ing stimulation of the perforant path. J. Physiol. Edition of the Complete Psychological Works of Sig-
232, 331–356. mund Freud, vol. 17, Strachey J., ed. London:
43. Centonze D., Picconi B., Gubellini P., Bernardi Hogarth Press, 1953, pp. 219–256.
G., and Calabresi P. (2001) Dopaminergic con- 57. Freud S. (1920) Beyond the Pleasure Principle.
trol of synaptic plasticity in the dorsal striatum. In: The Standard Edition of the Complete Psycho-
Eur. J. Neurosci. 13, 1071–1077. logical Works of Sigmund Freud, vol. 18, Strachey
44. Bashir Z. I. and Collingridge G. L. (1994) An J., ed. London: Hogarth Press, 1953, pp. 7–64.
investigation of depotentiation of long-term 58. Blair H. T., Schafe G. E., Bauer E. P., Rodrigues
potentiation in the CA1 region of the hip- S. M., and LeDoux J. E. (2001) Synaptic plastic-
pocampus. Exp. Brain Res. 100, 437–443. ity in the lateral amygdala: a cellular hypo-
45. Huang C. -C., and Hsu K. -S. (2001) Progress in thesis of fear conditioning. Learn. Mem. 8,
understanding the factors regulating reversibil- 229–242.
ity of long-term potentiation. Rev. Neurosci. 12, 59. Bauer E. P., Schafe G. E., and LeDoux J. E. (2002)
51–68. NMDA receptors and L-type voltage-gated cal-
46. Huang C. -C., Liang Y. -C., and Hsu K. -S. (2001) cium channels contribute to long-term potentia-
Characterization of the mechanism underlying tion and different components of fear memory
the reversal of long term potentiation by low formation in the lateral amygdala. J. Neurosci.
frequency stimulation at hippocampal CA1 22, 5239–5249.
synapses. J. Biol. Chem. 51, 48,108–48,117. 60. Lee H. and Kim J. (1998) Amygdalar NMDA
47. Picconi B., Centonze D., Hakansson K., et al. receptors are critical for new fear learning in
(2003) Loss of bidirectional striatal synaptic previously fear-conditioned rats. J. Neurosci. 18,
plasticity in L-DOPA-induced dyskinesia. Nat. 8444–8454.
Neurosci. 6, 501–506. 61. Miserendino M. J. D., Sananes C. B., Melia K. R.,
48. Kida S., Josselyn S. A., de Ortiz S. P., et al. (2002) and Davis M. (1990) Blocking of acquisition but
CREB required for the stability of new and reac- not of expression of conditioned fear-potenti-
tivated fear memories. Nat. Neurosci. 5, 348–355. ated startle by NMDA antagonists in the amyg-
49. Yin J .C., Wallach J. S., Del Vecchio M., et al T. dala. Nature 345, 716–718.
(1994) Induction of a dominant negative CREB 62. Lee H. J., Choi J. S., Brown T. H., and Kim J. J.
trangene specifically blocks long-term memory (2001) Amygdalar NMDA receptors are critical
in Drosophila. Cell 79, 49–58. for the expression of multiple conditioned fear
50. Genoux D., Haditsch U., Knobloch M., responses. J. Neurosci. 21, 4116–4124.
Michalon A., Storm D., and Mansuy I. M. (2002) 63. Rodrigues S. M., Schafe G. E., and LeDoux J. E.
Protein phosphatase 1 is a molecular constraint (2001) Intra-amygdala blockade of the NR2B
on learning and memory. Nature 418, 970–975. subunit of the NMDA receptor disrupts the
51. Otmakhova N. A. and Lisman J. E. (1998) acquisition but not the expression of fear condi-
D1/D5 dopamine receptors inhibit depotentia- tioning. J. Neurosci. 21, 6889–6896.
tion at CA1 synapses via cAMP-dependent 64. Falls W. A., Miserendino M. J. D., and Davis M.
mechanism. J. Neurosci. 18, 1270–1279. (1992) Extinction of fear-potentiated startle:

Molecular Neurobiology Volume 32, 2005

132 Centonze et al.

blockade by infusion of an NMDA antagonist in neonatal rat hippocampus. J. Neurosci. 19,

into the amygdala. J. Neurosci. 12, 854–863. 7568–7577.
65. Szapiro G., Vianna M. R. M., McGaugh J. L., 76. Sutton M. A., Schmidt E. F., Choi K. -H., et al.
Medina J. H., Izquierdo I. (2003) The role of (2003) Extinction-induced upregulation in AMPA
NMDA glutamate receptors, PKA, MAPK and receptors reduces cocaine-seeking behaviour.
CAMKII in the hippocampus in extinction of Nature 421, 70–75.
conditioned fear. Hippocampus 13, 53–58. 77. Wang J. H. and Stelzer A. (1996) Shared calcium
66. Hardingham N., Glazewski S., Pakhotin P., et signalling pathways in the induction of long-
al. (2003) Neocortical long-term potentiation term potentiation and synaptic disinhibition in
and experience-dependent synaptic plasticity CA1 pyramidal cell dendrites. J. Neurophysiol.
require alpha-calcium/calmodulin-dependent 75, 1687–1702.
protein kinase II autophosphorylation. J. Neu- 78. McGaugh J. L., Castellano C., and Brioni J.
rosci. 23, 4428–4436. (1990) Picrotoxin enhances latent extinction of
67. Liu S., Rao Y., and Daw N. (2003) Roles of pro- conditioned fear. Behav. Neurosci. 104, 264–267.
tein kinase A and protein kinase G in synaptic 79. Katona I., Rancz E. A., Acsady L., et al. (2001) Dis-
plasticity in the visual cortex. Cereb. Cortex 13, tribution of CB1 cannabinoid receptors in the
864–869. amygdala and their role in the control of GABAer-
68. Morozov A., Muzzio I. A., Bourtchouladze R., gic transmission. J. Neurosci. 21, 9506–9518.
et al. (2003) Rap1 couples cAMP signaling to a 80. Freund T. F., Katona I., and Piomelli D. (2003)
distinct pool of p42/44MAPK regulating Role of endogenous cannabinoids in synaptic
excitability, synaptic plasticity, learning, and signaling. Physiol. Rev. 83, 1017–1066.
memory. Neuron 39, 309–325. 81. Iversen L. (2003) Cannabis and the brain. Brain
69. Waltereit R. and Weller M. (2003) Signaling 126, 1252–1270.
from cAMP/PKA to MAPK and synaptic plas- 82. Alger B. E. and Pitler T. A. (1995) Retrograde
ticity. Mol. Neurobiol. 27, 99–106. signaling at GABAA-receptor synapses in the
70. Schafe G. E., Nader K., Blair H. T., and LeDoux mammalian CNS. Trends Neurosci. 18, 333–340.
J. E. (2001) Memory consolidation of Pavlovian 83. Wilson R. I. and Nicoll R. A. (2001) Endogenous
fear conditioning: a cellular and molecular per- cannabinoids mediate retrograde signalling at
spective. Trends Neurosci. 24, 540–546. hippocampal synapses. Nature 410, 588–592.
71. Dudai Y. (2002) Molecular bases of long-term 84. Ohno-Shosaku T., Maejima T., and Kano M.
memories: a question of persistence. Curr. Opin. (2001) Endogenous cannabinoids mediate retro-
Neurobiol. 12, 211–216. grade signals from depolarized postsynaptic
72. Lu K. -T., Walker D. L., and Davis M. (2001) neurons to presynaptic terminals. Neuron 29,
Mitogen-activated protein kinase cascade in the 729–738.
basolateral nucleus of the amygdala is involved 85. Chevaleyre V. and Castillo P. E. (2003) Het-
in extinction of fear-potentiated startle. J. Neu- erosynaptic LTD of hippocampal GABAergic
rosci. 21, RC162. synapses: a novel role of endocannabinoids in
73. Marsicano G., Wotjak C. T., Azad S. C., et al. regulating excitability. Neuron 38, 461–472.
(2002) The endogenous cannabinoid system 86. Ramon y Cajal S. (1892) A new concept of the
controls extinction of aversive memories. histology of the central nervous system. In:
Nature 418, 530–534. Neurological Classics in Modern Translation, Rot-
74. Gaiarsa J. -L., Caillard O., and Ben-Ari Y. (2002) tenberg D. A., and Hochberg F. H., eds. New
Long-term plasticity at GABAergic and glycin- York: Hafner, 1977, pp. 7–29.
ergic synapses: mechanisms and functional sig- 87. Freud S. (1895) Project for a Scientific Psychol-
nificance. Trends Neurosci. 25, 564–570. ogy In: The Standard Edition of the Complete Psy-
75. Caillard O., Ben-Ari Y., and Gaiarsa J. L. (1999) chological Works of Sigmund Freud, vol. 1,
Mechanisms of induction and expression of Strachey J., ed. London: Hogarth Press, 1953,
long-term depression at GABAergic synapses pp. 283–397.

Molecular Neurobiology Volume 32, 2005

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.