1
TITLE PAGE
Title: Temporal gating of synaptic competition in the lateral amygdala by
cannabinoid receptor modulation of the thalamic input
Author Affiliation: Ana Drumonda,b, Natália Madeiraa,b, Rosalina Fonsecaa,b
a Instituto Gulbenkian de Ciência, Rua Quinta Grande, 6 2780-156 Oeiras
bChronic Diseases Research Center (CEDOC), NOVA Medical School,
Universidade Nova de Lisboa, Campo dos Mártires da Pátria, 130 1169-056
Lisboa
Corresponding Author: Rosalina Fonseca, Campo dos Mártires da Pátria, 130
1169-056 Lisboa, phone: +351 218803101 email: [email protected]
Keywords: Amygdala, Cannabinoid receptor, Long-term potentiation, Synaptic
competition, Synaptic cooperation,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
2
ABSTRACT
The acquisition of fear memories involves plasticity of the thalamic and cortical
pathways to the lateral amygdala (LA). The maintenance of synaptic plasticity
requires the interplay between input-specific synaptic tags and the allocation of
plasticity-related proteins (PRPs). Based on this interplay, weakly activated
synapses can express long-lasting synaptic plasticity by cooperation with strongly
activated ones. Increasing the number of activated synapses can shift
cooperation to competition. Synaptic cooperation and competition can determine
whether two events, separated in time, are linked or selected. The rules that
determine whether synapses cooperate or compete are unknown. We found that
synaptic cooperation and competition, in the LA, are determined by the temporal
sequence of cortical and thalamic stimulation and that the strength of the synaptic
tag is modulated by the endocannabinoid signalling. This modulation is
particularly effective in thalamic synapses, suggesting a critical role of
endocannabinoids in restricting thalamic plasticity. Also, we found that PRPs
availability is modulated by the action-potential firing of neurons, shifting
competition to cooperation. Our data present the first evidence that pre-synaptic
modulation of synaptic activation, by the cannabinoid signalling, function as a
temporal gating mechanism limiting synaptic cooperation and competition.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
3
INTRODUCTION
Learning is a key process allowing individuals to adapt to environmental
challenges (McKenzie and Eichenbaum, 2011). Cellular models of memory, such
as long-term potentiation (LTP), share common principles with memory
consolidation (Bliss and Collingridge, 1993). As for memory, not all events are
maintained in long-lasting forms of LTP. Depending on the strength of synaptic
activation, LTP induction can set a local, input-specific, synaptic tag or it can also
induce the synthesis of PRPs (Frey and Morris, 1997; Sajikumar and Frey, 2004).
If protein synthesis is triggered, LTP is maintained. Since the setting of the
“synaptic tag” and the capture of PRPs are two independent processes, then
induction of a maintained form of LTP in one set of synapses can stabilize a
transient form of LTP induced in a second independent set of synapses (Redondo
and Morris, 2011). This cooperative maintenance allows neurons to integrate
events that occur within large time windows, ranging from 30 to 60 minutes (Frey
and Morris, 1998; Sajikumar et al., 2007). If PRPs are limited, for example by
blocking protein-synthesis, or by increasing the pool of activated synapses, then
synaptic competition is observed (Fonseca et al., 2004; Govindarajan et al., 2011;
Sajikumar et al., 2014). In this working model, synaptic cooperation and
competition are determined by a balance between PRPs availability (source) and
the strength of the tag (sink). In turn, the tag strength is a combination of the
number of activated synapses and the ability that each synapse has in capturing
PRPs (Fonseca, 2012; Szabó et al., 2016). Previous studies looking at
competition in CA1 hippocampal synapses have shown that competition was
correlated with the strength of the tag (Fonseca et al., 2004). According to this,
synapses were stronger or weaker depending on their relative activation and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
4
therefore, winning synapses block the maintenance of LTP in loser synapses.
However, this was not observed in a second study, where competition displayed
a “winner-take-all” form. In this case, strong and weak activated synapses were
similar in their ability to capture PRPs and neither were maintained in a
competitive setting (Sajikumar et al., 2014).
Cooperation in the lateral-amygdala follows similar principles as shown for
hippocampal synapses. Lateral amygdala pyramidal cells receive input from the
auditory cortex (C) and the auditory thalamus (T), a neuronal circuitry that is
involved in the acquisition of fear memories (Amano et al., 2010). Lesion studies
have shown that the activation of either thalamic or cortical inputs is sufficient to
acquire a fear memory, whereas both inputs are necessary to discriminate
between fearful and neutral events (Antunes and Moita, 2010). We have
previously shown that cortical and thalamic synapses can cooperate by a
synaptic tagging and capture mechanism (Fonseca, 2013). Interestingly, thalamic
and cortical synapses were not identical in their ability to capture PRPs and
maintain LTP by cooperation. Thalamic synapses had a much shorter time
window of cooperation and this time-window was restricted by the activation of
presynaptic cannabinoid receptors (CB1R). This finding gains further relevance
if one considers the established link between endocannabinoid signalling and the
acquisition of fear memories (Drumond et al., 2017; Mechoulam and Parker,
2013). Endocannabinoids are synthesised on-demand, triggered by neuronal
activation and once synthesised, cross the neuronal membrane and bind to their
receptors (type 1 cannabinoid receptor – CB1R – and type 2 cannabinoid
receptor – CB2R), resulting in a suppression of neurotransmitter release at
excitatory and inhibitory synapses (Kano et al., 2009; Turu and Hunyady, 2010).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
5
Their degradation by the Fatty acid amide hydrolase (FAAH) and the DAGL
(DAG-hydrolizing enzyme) terminate their signalling (Drumond et al., 2017).
CB1Rs are highly expressed in the amygdala with regional differences between
the lateral, basal and central amygdala (Marsicano and Lutz, 1999). There is
compelling evidence that CB1R activation reduces amygdala excitability (Azad et
al., 2003; Lutz et al., 2015) and together with a recent report that
endocannabinoid synthesis is sensitized by previous negative events
(Sumislawski et al., 2011), support the hypothesis that endocannabinoid
signalling restricts fear-memory acquisition.
Here, we assessed the temporal rules that determine whether cortical and
thalamic synapses interact by cooperation or competition. Further, we tested the
role of the endocannabinoid signalling in competition and their contribution in the
modulation of cortical and thalamic synaptic plasticity. Our data support the
hypothesis that synaptic competition is a cellular mechanism to select the events
that are maintained and its temporal restriction may promote memory selectivity.
MATERIALS AND METHODS
A total of 260 slices prepared from 104 male Sprague-Dawley rats (3-5
week old) were used for electrophysiological recordings. All procedures were
approved by the Portuguese Veterinary Office (Direcção Geral de Veterinária e
Alimentação - DGAV). Coronal brain slices (350 µm) containing the lateral
amygdala were prepared as described previously (Fonseca, 2013). Whole-cell
current-clamp synaptic responses were recorded using glass electrodes (7-
10MΩ; Harvard apparatus, UK), filled with internal solution containing (in mM): K-
gluconate 120, KCl 10, Hepes 15, Mg-ATP 3, Tris-GTP 0.3 Na-phosphocreatine
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
6
15, Creatine-Kinase 20U/ml (adjusted to 7.25 pH with KOH, 290mOsm). Putative
pyramidal cells were selected by assessing their firing properties in response to
steps of current (Supplementary Figure 1A). Voltage-clamp synaptic currents
were recorded using 2-3MΩ glass electrodes filled with internal solution. Only
cells that had a resting potential of less than -60mV without holding current were
taken further into the recordings. Neurons were kept at -70mV to -75mV with a
holding current below -0.25nA. In current clamp recordings, the series resistance
was monitored throughout the experiment and ranged from 30MΩ-40MΩ; in
voltage clamp recordings, series resistance ranged from 10-20MΩ; changes
exceeding 25% of the series resistance determined the end of the recording.
Electrophysiological data were collected using an RK-400 amplifier (Bio-Logic,
France) filtered at 1 kHz and digitized at 10kHz using a Lab-PCI-6014 data
acquisition board (National Instruments, Austin, TX) and stored on a PC. Offline
data analysis was performed using a customized LabView-program (National
Instruments, Austin, TX). To evoke synaptic EPSP, tungsten stimulating
electrodes (Science Products, GmbH, Germany) were placed on afferent fibers
from the internal capsule (thalamic input) and from the external capsule (cortical
input–Supplementary Figure 1A’). Pathway independence was checked by
applying two pulses with a 30ms interval to either thalamic or cortical inputs and
confirming the absence of crossed pair-pulse facilitation. EPSPs were recorded
with a test pulse frequency for each individual pathway of 0.033 Hz. After 20 min
of baseline, long-lasting LTP (L-LTP) was induced with a strong tetanic
stimulation (25 pulses at a frequency of 100 Hz, repeated five times, with an
interval of 3 sec), whereas transient LTP (E-LTP) was induced with a weak tetanic
stimulation (25 pulses at a frequency of 100 Hz, repeated three times with an
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
7
interval of 3 sec). DSE was induced by depolarizing the cell (in voltage-clamp
mode) to 0mV during 10 sec.
Drugs were dissolved in DMSO (0.01%) and diluted to achieve the final
concentration: Rapamycin (Tocris) 1µM, AM281 (Sigma) 0.5µM or 1 µM, URB597
(Tocris) 1µM, Picrotoxin (Sigma) 25µM, Win55,212-12 (Tocris) 5µM or 10 µM and
SCH-50911 (Tocris) 10 µM. In control experiments, only DMSO was added to the
ACSF.
As a measure of synaptic strength, the initial slope of the evoked EPSPs
was calculated and expressed as percent changes from the baseline mean. Error
bars denote SEM values. For the statistical analysis, LTP values were averaged
over 5 min data bins immediately after LTP induction (T Initial - see timeline for
each particular pathway and experimental condition) and at the end of the
recording (T Final 100-105 minutes). LTP decay was calculated by [(T Initial –T
Final)/T Final*100]. Normality was assessed using the Kolmogorov-Smirnov and
the Shapiro-Wilk test. Since some of our groups were not normal, group
differences were assessed using a non-parametric test (Kruskal–Wallis Test,
SPSS software). For the correlation analysis a multiple regression analysis was
used (SPSS Software). In Depolarization-suppression of excitation (DSE)
experiments, the maximal amplitude of currents evoked by thalamic or cortical
stimulation (EPSC) was compared before and after depolarization (Kruskal–
Wallis test). For the analysis of the decrease in EPSP slope and changes in PPF,
induced by CB1R agonist, we used the Friedman ANOVA (SPSS software) and
compared changes before and after agonist application.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
8
RESULTS
Stimulation of a second thalamic input leads to competition and
destabilization of a previous induced LTP at thalamic synapses
Both thalamic and cortical input afferents to the lateral amygdala can
express transient and maintained forms of LTP, depending on the stimulation
strength (Fonseca, 2013). Weak tetanic stimulation of thalamic or cortical inputs
led to an increase in EPSP slope that decayed to baseline values at the end of
the recording (Supplementary Figure 1B/C). If cortical synapses were stimulated
with a strong tetanic stimulation, prior to weak thalamic stimulation, thalamic LTP
was maintained despite the weak synaptic stimulation (Supplementary Figure
1D). This heterosynaptic cortical-to-thalamic cooperation is dependent on the
synthesis of plasticity-related proteins. Bath application of rapamycin, an mTOR
pathway inhibitor (Sosanya et al., 2015), led to the destabilization of the cortical
LTP as well as to the blockade of the cooperative maintenance of the weak
thalamic LTP (Supplementary Figure 1E). Analysis of the percentage decay of
LTP showed that the maintenance of LTP in cortical and thalamic synapses
depend on the synthesis of PRPs or its capture through synaptic cooperation
(Supplementary Figure 1F).
Since the maintenance of LTP is dependent on this interplay between
synaptic tags and the availability of PRPs, to test whether thalamic and cortical
synapses compete, we increased the pool of activated synapses, by stimulating
a second thalamic input (thalamic W2). Our previous results showed that the time
window for thalamic synapses to effectively capture PRPs was around 7.5
minutes (Fonseca, 2013). Thus, we stimulated a second thalamic input (W2), with
a weak tetanic stimulation, 7.5 minutes after the first weak thalamic (W1)
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
9
stimulation. We observed that the LTP in the thalamic W2 was not maintained
and that the LTP in the thalamic W1 was destabilized, leading to its decay.
Interestingly, the cortical LTP (S3) was not significantly destabilized by this
thalamic competition (Figure 1A) and competition was not observed in the
absence of W2 stimulation (Figure 1B).
Figure 1. Competition was induced by increasing the pool of activated thalamic synapses. A. Weak stimulation of the Thalamic W2 led to the destabilization of the thalamic W1 LTP, but not of the cortical S3 LTP. B. If the Thalamic W2 was not weakly stimulated, LTP in the cortical S3 and thalamic W1 was maintained, by cooperation. C. Cooperation was also observed if weak thalamic W2 was stimulated 22.5 minutes after cortical S3 strong stimulation. D. The decay of LTP in the cortical S3 was similar in all conditions tested, whereas the Thalamic LTP, in W1 and W2, decayed significantly more if both thalamic inputs were stimulated (*p=0.007; # p=0.01). Inserts represent EPSPs traces before (T1), after LTP induction (T2) and at the end of the recording (T3). Bars: 20 ms and 15mV. Error bars represent SEM, n=number of slices.
To rule out that the decay observed in the thalamic W2, under competition,
was due to the time of weak W2 stimulation, we did a cooperation experiment in
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
10
which the weak stimulation of W1 is not present. Thalamic W2 LTP was still
maintained even if stimulated 22.5 minutes after cortical strong stimulation
(Figure 1C), showing that at this later time point W2 is still able to capture PRPs.
No significant change in LTP decay was observed for the cortical input in all
conditions (Figure 1D). These observations suggest that competition results from
an imbalance between the availability of PRPs and the pool of activated synapses
in a source-to-sink distribution.
Modulation of the endocannabinoid system alters synaptic competition
We have previously found that inhibiting CB1R extends the time-window
of thalamic cooperation (Fonseca, 2013). To test whether modulation of the
endocannabinoid system also plays a role in synaptic competition, we did a
competition experiment while decreasing or increasing endocannabinoid
signalling. We found that pharmacological inhibition of CB1R led to an increase
in competition, resulting in the destabilization of LTP in all inputs, thalamic and
cortical (Figure 2A; Supplementary Figure S2). Conversely, inhibition of
anandamine degradation, by URB597 a specific inhibitor of FAAH, reduced
competition and allowed the maintenance of LTP in all stimulated inputs (Figure
2B). Importantly, the inhibition of CB1R by AM281 did not block LTP maintenance
if competition was not induced (Figure 2C). Analysis of the percentage decay of
LTP showed that URB597 treatment significantly decreased the decay of LTP in
both thalamic W1 and W2 whereas AM281 application significantly increased
cortical S3 decay under competition (Figure 2D). These results show that PRPs
synthesised upon strong cortical stimulation are able to promote the maintenance
of all activated synapses if CB1R activation is increased. Since the availability of
PRPs is similar in all conditions tested, this observation suggests that CB1R
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
11
activation modulates the strength of the tag gating the balance towards
cooperation, if CB1R activation is increased, or towards competition, if CB1R
activation is decreased.
Figure 2. Competition was modulated by the activation of CB1R. A. Inhibition of CB1R, by AM281application (0.5µM – solid line) led to an increase in competition and to the destabilization of the cortical S3 LTP. B. Conversely, the inhibition of anandamine degradation, by URB597 application (1µM – solid line), led to the maintenance of LTP in all inputs. C. Inhibition of CB1R, by AM281 application (solid line) did not induce competition if the thalamic W2 input was not stimulated. D. The decay of LTP in the strong cortical was significantly higher in the presence of AM281, whereas the thalamic LTP decayed significantly less in the presence of URB597 or when the thalamic W2 was not stimulated. Inserts represent EPSPs traces before (T1), after LTP induction (T2) and at the end of the recording (T3). Bars: 20 ms and 15mV. Error bars represent SEM, n=number of slices.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
12
Competition depends on the relative time of synaptic stimulation
Since the tag in restricted in time, the sequence of synaptic activation
between the multiple inputs will determine whether PRPs are still available, but
also if synapses are still capturing PRPs. In our experimental conditions,
competition is induced by the stimulation of the second thalamic input that creates
an additional pool of activated synapses that capture PRPs. To test whether
synaptic competition depends on the time interval of the different input
stimulation, we delayed the weak thalamic W2 stimulation to 65 minutes. Based
on our previous findings, by increasing the interval between the first and second
thalamic stimulation to 30 minutes, we predicted that W1 is no longer destabilized
by W2 stimulation. Indeed, we found that synaptic competition is abolished and
no destabilization was observed in the thalamic W1 LTP maintenance (Figure
3A).
Figure 3. Delaying the stimulation of the thalamic W2 prevents competition A. If thalamic W2 weak stimulation was delayed to 65 minutes (30 minutes after
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
13
W1 stimulation) competition was not observed. B. Thalamic W2 was not maintained by cooperation if weak stimulation was delayed to 65 minutes (45 minutes after cortical stimulation). C. Application of AM281 (solid line) increases competition even if thalamic W2 weak stimulation is delayed to 65 minutes, leading to the destabilization of the cortical and thalamic W1 LTP. D. The decay of LTP in the strong cortical and the thalamic W1 was significantly higher in the presence of AM281, whereas the decay of the thalamic W2 did not differ in all conditions. Inserts represent EPSPs traces before (T1), after LTP induction (T2) and at the end of the recording (T3). Bars: 20 ms and 15mV. Error bars represent SEM, n=number of slices.
Interestingly, thalamic W2 LTP was not stabilized, suggesting that PRPs,
at this time point, are not sufficient to maintain the thalamic W2 LTP. This
interpretation is corroborated by the observation that cooperation also did not
occur if strong cortical S3 and weak W2 thalamic stimulation were separated by
45 minutes (Figure 3B). As in the previous experiment, inhibiting CB1R increased
competition. In the presence of AM281, stimulation of thalamic W2 promoted the
destabilization of thalamic W1 and cortical S3 LTP (Figure 3C). Weak stimulation
of thalamic W2, under CB1R inhibition, significantly increased the LTP decay of
the thalamic W1 and cortical S3 LTP (Figure 3D).
Competition depend on the availability of PRPs
Since competition relies on the balance between the source and the sink
of PRPs, then increasing the availability of PRPs, the source, should abolish
competition. In our experimental condition, PRPs synthesis was induced by the
strong cortical stimulation. To test this, we shortened the time interval between
weak thalamic and strong cortical synaptic stimulation. We found that, when
thalamic S1 and S2 were stimulated in the close temporal vicinity to cortical
strong stimulation, LTP in all activated inputs was maintained (Figure 4A). Once
more, CB1R inhibition led to an increase in competition and to the destabilization
of LTP in all stimulated inputs (Figure 4B). Reducing the time interval of weak
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
14
thalamic stimulation in relation to strong cortical LTP, decreased competition,
reducing LTP decay. Inhibiting CB1R activation restored competition, suggesting
that blocking CB1R created a new imbalance between source and sink (Figure
4C).
Figure 4. Competition is abolished if PRPs availability is increased. A. Shortening the time interval between the weak thalamic W1 and W2 and the cortical S3 stimulation abolished competition. B. Application of AM281 increased competition and led to the destabilization of LTP in all inputs. C. Percentage decay of LTP for all conditions tested. The decay of LTP is significantly higher for all inputs in the presence of AM281. Inserts represent EPSPs traces before (T1), after LTP induction (T2) and at the end of the recording (T3). Bars: 20 ms and 15mV. Error bars represent SEM, n=number of slices.
Since the inhibition of CB1R increased competition and considering that
competition is determined by a source-to-sink balance, then CB1R inhibition may
increase the sink (tag) or decrease the source (PRPs). Although CB1R inhibition
is associated with a decrease in protein synthesis, by modulation of the mTOR
pathway, the observation that inhibiting CB1R promotes cooperation (Fonseca,
2013), does not support the hypothesis that CB1R inhibition decreases the
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
15
source. An alternative hypothesis is that inhibition of CB1R increases the sink,
which goes in line with CB1R inhibition promoting cooperation. Given that CB1R
inhibition, in a competitive setting, led to the destabilization of the cortical LTP,
we hypothesised that CB1R modulation is higher in thalamic synapses than in
cortical synapses. If indeed CB1R preferentially modulates the tag strength of
thalamic synapses, then inhibiting CB1R during competition would create an
imbalance towards thalamic synapses forcing the cortical synapses to lose. To
test this hypothesis, we assessed the modulation of thalamic and cortical
synapses by CB1R. CB1R activation is necessary for the induction of the
depolarizing-induced suppression of excitation (DSE). This short-term plasticity
is associated with a decrease in presynaptic glutamate release induced by
transient depolarization (Kodirov et al., 2010). The induction of DSE resulted in a
significant decrease in EPSC amplitude both in the thalamic and cortical activated
synapse (Supplementary Figure 3A). Inhibition of CB1R with 0.5µM AM281
application blocked the induction of DSE in cortical but not in thalamic inputs
(Supplementary Figure 3B), whereas inhibition of CB1R with 1µM AM281
application blocked the induction of DSE in both cortical and thalamic inputs
(Supplementary Figure 3C). Since a higher concentration of AM281 was
necessary to fully block DSE in thalamic synapses, our results support the
hypothesis that the expression, or activity, of CB1R is higher in thalamic inputs.
To further confirm these findings, we analysed the effect of a CB1R agonist
in EPSPs evoked by thalamic and cortical input stimulation. Bath application of
Win55,212-2 reduced EPSP slope evoked by thalamic and cortical input
stimulation (Supplementary Figure 4A). As for the DSE experiment, a stronger
effect of the CB1R agonist was observed in thalamic synapses (Supplementary
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
16
Figure 4A/B). Analysis of pair pulse facilitation (PPF), evoked by stimulation of
thalamic or cortical inputs showed a significant decrease in PPF, suggesting a
presynaptic effect of CB1R activation (Supplementary Figure 4C). Taken
together, our results suggest that CB1R activation has a major role in modulating
thalamic synapses with a clear impact in restricting both cooperation and
competition between cortical and thalamic synapses.
Decreasing inhibition promotes synaptic cooperation
It is well established that CB1R activation can modulate excitation as well
as inhibition (Azad et al., 2003), resulting in an overall decrease in amygdala
excitability (Pistis et al., 2004). Thus, one possibility is that the inhibition of CB1R
increases competition by increasing excitability. To test this hypothesis, we used
Picrotoxin, an inhibitor of GABA A receptors, to increase excitability in the lateral
amygdala during competition. Since decreasing inhibition reduces the threshold
of LTP induction, we used the experimental design of late W2 thalamic
stimulation, described in Figure 3A, to avoid overlap of picrotoxin with LTP
induction. Picrotoxin application (25µM) was restricted to the time interval
between the W1 and W2 weak thalamic stimulation (from 35 to 60 minutes).
Additionally, we confirmed that weak thalamic stimulation, in this experimental
condition, did not led to a maintained LTP (Supplementary Figure 5A). We then
tested whether weak stimulation of W2 would promote competition, under
decreased inhibition. We did not observe competition, rather, we found that,
under reduced inhibition, the weak thalamic W2 was now maintained, despite its
late induction (Figure 5A). Co-application of rapamycin with picrotoxin abolished
the maintenance of the LTP in the thalamic W2, showing that weak thalamic LTP
maintenance occurs through cooperative sharing and capture of PRPs
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
17
(Supplementary Figure 5B). Taken together, decreasing inhibition did not
increase competition but rather allowed the maintenance of the weak thalamic
W2, in a protein-synthesis dependent manner. This suggests that picrotoxin
increases the availability of PRPs which were captured by the weak thalamic S2.
Previous studies have found an association between protein synthesis and
neuronal activity, in particular action-potential firing (Fonseca et al., 2006;
Karpova, 2006). Since our cells fire more action potentials, under picrotoxin
application, we repeated the experiment described above but prevented the cells
from firing during the time period of picrotoxin application. Cells were voltage-
clamped to -75mV, from 40 to 60 minutes. Preventing neuronal firing abolished
the maintenance of the weak thalamic S2 LTP (Figure 5B). In this condition the
cooperative maintenance of the thalamic W2 LTP, promoted by picrotoxin, was
abolished. Further, the number of spikes negatively correlated with the
percentage of LTP decay in the weak thalamic W2 pathway (Figure 5C). This
argues for an increase PRPs synthesis induced by spiking, favouring the
maintenance of the transient weak thalamic LTP through synaptic cooperation.
Inhibition of protein synthesis, at this later time point, had no effect in thalamic
W1 and cortical LTP, showing that LTP maintenance at this point was no longer
dependent on protein synthesis (Figure 5D).
If decreasing inhibition favours an increase in PRPs availability, it is
conceivable that it abolishes the increase in competition observed under
inhibition of CB1R. To test this, we did a late competition experiment with the co-
application of AM281 and picrotoxin. The inhibition of CB1R with AM281 covers
the induction of LTP in both W1 and W2 thalamic inputs whereas
AM281/picrotoxin co-application was restricted to the time window between S1
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
18
and S2 thalamic stimulation. In this experimental setting, we still observed
competition similar to AM281 application alone (Figure 5E/ Supplementary Figure
5C). A similar result was obtained by co-application of SCH-50911, an inhibitor
of GABA B receptors (Supplementary Figure 5D). Analysis showed no
differences, in the percentage decay of LTP, between AM281, AM281/picrotoxin
and AM281/SCH-50911 co-application (Figure 5F). This shows that the decay in
LTP induced by CB1R inhibition, in competition, is not due to an increase in
inhibition. Rather, CB1R inhibition creates a new unbalance by increasing
thalamic activation.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
19
Figure 5. Decreasing inhibition promotes cooperation. A. Decreasing inhibition, by Picrotoxin application (25µM – solid line), allowed the maintenance of LTP in all inputs. B. Maintaining the cells in voltage-clamp mode abolished the cooperative maintenance induced by picrotoxin application. C. Correlation plot of the total number of spikes throughout the recorded time against the percentage decay of the thalamic W2. The significant negative correlation showed that a higher spike number decreases thalamic W2 decay. D. Percentage decay of LTP for all conditions tested. The decay of LTP in the thalamic W2 is significantly lower if slices are treated with picrotoxin, before thalamic W2 stimulation. No significant change in LTP decay was observed in the cortical and thalamic W1. E. Inhibition of CB1R, by AM281 (dashed line), increased competition even if GABA A receptors are also inhibited (picrotoxin - solid line). F. No difference was observed in LTP decay for AM281 or co-application of AM281 with picrotoxin or SCH-50911. Inserts represent EPSPs traces before (T1), after LTP induction (T2) and at the end of the recording (T3). Bars: 20 ms and 15mV. Error bars represent SEM, n=number of slices.
DISCUSSION
In this study, we present the first evidence that thalamic and cortical inputs,
to the lateral amygdala, interact by synaptic competition. Synaptic cooperation
and competition are two forms of heterosynaptic plasticity that occur within large
time intervals and result from the synaptic integration of biochemical signals. As
described in hippocampal CA3 to CA1 synapses (Fonseca et al., 2004; Sajikumar
et al., 2014), competition in the amygdala was observed when the pool of
activated synapses was higher than the availability of PRPs. Our results support
the hypothesis that synaptic cooperation and competition are determined by a
balance between the pool of activated synapses (sink) and the availability of
PRPs (source). In our experimental setting, PRPs synthesis was induced by the
strong stimulation of the cortical input to the lateral amygdala (sink and source),
whereas two thalamic inputs were used as two additional sinks. Because cortical
fibers are packed in a bundle we were unable to stimulate two independent
cortical pathways, a pre-requisite to assess synaptic cooperation or
competition. When a second thalamic input was stimulated with a weak tetanus,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
20
sufficient to induce a synaptic tag but not PRPs synthesis, competition led to the
decay of a previously induced thalamic LTP (Figure 6A). Interestingly, cortical
LTP was not destabilized by thalamic competition. One possibility is that elapsed
time consolidated cortical LTP rendering it resistant to disruption by competition.
This hypothesis is supported by the observation that inhibiting protein synthesis
after cortical strong stimulation had no effect in cortical maintenance. An
alternative hypothesis is that the cortical LTP was maintained because the
strength of the cortical tag, or the ability of cortical synapses to capture PRPs,
was higher than in thalamic synapses. In this scenario, the strength of the tag
determines winners and losers. We have previously found, in hippocampal
synapses, that the strength of synaptic stimulation correlates with the strength of
the tag and with competitive load (Fonseca et al., 2004). Thus, if the thalamic
weak stimulation induces a weaker synaptic tag than the strong cortical
stimulation, cortical synapses will be able to capture more PRPs, win the
competition and maintain LTP.
Figure 6. Model of synaptic cooperation and competition. A. In competition, the stimulation of the thalamic W2 leads to the induction of an additional sink that creates an imbalance between the availability of PRPs and the number of tags destabilizing thalamic LTP. B. If the second thalamic stimulation is delayed, the
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
21
strong cortical and the weak thalamic W1 are stabilized and resistant to competition. C. Inhibition of CB1R increases the duration of the tag promoting competition. D. Decreasing inhibition promotes PRPs synthesis, decreasing the competitive load and promoting cooperation.
The hypothesis that the degree of synaptic activation correlates with the
strength of the tag is supported by a previous study showing that suspending
synaptic activation abolishes PRPs capture and plasticity maintenance (Szabó et
al., 2016). It also goes in line with our previous finding, that inhibiting CB1R
extends the time-window in which thalamic synapses can capture PRPs
(Fonseca, 2013). If CB1R activation acts as a negative feedback signal,
decreasing excitatory synaptic transmission, then inhibiting CB1R increases
synaptic activation and therefore the strength of the synaptic tag. In this scenario,
inhibiting CB1R, in a competitive setting, would lead to an increase of the tag
(sink) and to the competitive load. This is indeed what we observed under CB1R
inhibition, where competition is promoted (Figure 6C). In this functional model,
the strength of the tag is modulated by on-going synaptic activation, determining
whether a subsequent synaptic stimulation will be maintained by cooperation or
destabilized by competition. Our data show that in amygdala synapses, activation
of presynaptic CB1R decreases the tag and contribute to a restriction in the time
window of cooperation and competition.
It is also possible that inhibition of CB1R decreases the availability of PRPs
promoting competition. Indeed, two previous reports have shown that CB1R
inhibition downregulates the activity of the mTOR pathway (Busquets-Garcia et
al., 2013; Puighermanal et al., 2013). However, since CB1R inhibition promotes
cooperation (Fonseca, 2013), it is highly unlikely that it decreases PRPs
synthesis. Also, we did not observe any effect on LTP maintenance, under
AM281 application, when the thalamic S2 input was not stimulated.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
22
Competition is also modulated by the availability of PRPs, which in turn, is
time and activity-dependent. In the case where weak thalamic stimulation was
temporally closer to the strong cortical stimulation, competition was not observed,
suggesting that the initial availability of PRPs was sufficient to maintain LTP in all
activated synapses. Interestingly, the overall neuronal activity can also modulate
PRPs synthesis. When inhibition was blocked, PRPs availability increased,
allowing the cooperative maintenance of a weak thalamic input at much later time
points. The link between neuronal excitability and the synthesis of PRPs has been
widely reported (Barco et al., 2002; Ehlers, 2003; Han et al., 2008). Our results
suggest that inhibition in the amygdala also modulates the synthesis of PRPs,
restricting synaptic cooperation and thus, promoting competition. It is also
interesting that reducing inhibition, by either GABA A or B receptor inhibition, did
not prevent the increase in competition induced by the CB1R blockade. This
finding rules out the possibility that the increase in competition, observed under
CB1R inhibition, was a reflection of an increase in network inhibition.
It is now clear that cooperation and competition are mechanisms involved
in memory acquisition and maintenance (Cai et al., 2016; Kastellakis et al., 2015;
Rashid et al., 2016; Tonegawa et al., 2015). However, the rules by which these
two processes are orchestrated are still unclear. Our results show that the time
interval between the occurrence of two events will determine whether synaptic
cooperation or competition is favoured. Recent studies have also presented
compelling evidence that neuronal excitability determines which neurons are
recruited to participate in the encoding of a particular memory (Yiu et al., 2014).
Our data are consistent with this hypothesis and extends it in a significant
manner. By activating CB1R, pyramidal cells in the amygdala regulate their
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
23
intrinsic excitability reducing the probability that subsequent events are linked.
Taken together, our results set up a strong conceptual model, built from a detailed
analysis of synaptic plasticity in the amygdala, that provides clear-cut predictions
regarding the rules of memory acquisition and maintenance. Given that
endocannabinoid signalling is involved in anxiety, our results support the
hypothesis that by limiting cooperation and competition, the activation of
cannabinoid receptors may restrict fear generalization, one of the mechanisms
underlying the development of anxiety disorders.
FUNDING
This work is supported by a grant from the Portuguese Research Council
(Fundação para a Ciência e Tecnologia – 02/SAICT/2017/030772) and a Young
Investigator Grant from the Brain and Behaviour Research Foundation (NARSAD
grant number 25118). Rosalina Fonseca is supported by an FCT Investigator
grant (IF/01359/2014) and Natália Madeira is supported by a PhD fellowship
(SFRH/BD/130911/2017).
ACKNOWLEDGMENTS
We are grateful to Dr. Rita Teodoro for helpful comment on the manuscript. We
gratefully thank Dr Tobias Bonhoeffer and Dr Volker Staiger for providing
software for the electrophysiology experiments.
REFERENCES
Amano, T., Unal, C.T., Paré, D., 2010. Synaptic correlates of fear extinction in the amygdala. Nat. Neurosci. 13, 489–94. doi:10.1038/nn.2499 Antunes, R., Moita, M.A., 2010. Discriminative auditory fear learning requires both tuned and nontuned auditory pathways to the amygdala. J. Neurosci. 30,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
24
9782–7. doi:10.1523/JNEUROSCI.1037-10.2010 Azad, S.C., Eder, M., Marsicano, G., Lutz, B., Zieglgänsberger, W., Rammes, G., 2003. Activation of the Cannabinoid Receptor Type 1 Decreases Glutamatergic and GABAergic Synaptic Transmission in the Lateral Amygdala of the Mouse. Learn. Mem. 10, 116–128. doi:10.1101/lm.53303 Barco, A., Alarcon, J.M., Kandel, E.R., 2002. Expression of constitutively active CREB protein facilitates the late phase of long-term potentiation by enhancing synaptic capture. Cell 108, 689–703. Bliss, T. V, Collingridge, G.L., 1993. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31–9. doi:10.1038/361031a0 Busquets-Garcia, A., Gomis-González, M., Guegan, T., Agustín-Pavón, C., Pastor, A., Mato, S., Pérez-Samartín, A., Matute, C., De La Torre, R., Dierssen, M., Maldonado, R., Ozaita, A., 2013. Targeting the endocannabinoid system in the treatment of fragile X syndrome. Nat. Med. 19, 603–607. doi:10.1038/nm.3127 Cai, D.J., Aharoni, D., Shuman, T., Shobe, J., Biane, J., Song, W., Wei, B., Veshkini, M., La-Vu, M., Lou, J., Flores, S.E., Kim, I., Sano, Y., Zhou, M., Baumgaertel, K., Lavi, A., Kamata, M., Tuszynski, M., Mayford, M., Golshani, P., Silva, A.J., 2016. A shared neural ensemble links distinct contextual memories encoded close in time. Nature 534, 115–118. doi:10.1038/nature17955 Drumond, A., Madeira, N., Fonseca, R., 2017. Endocannabinoid signaling and memory dynamics: A synaptic perspective. Neurobiol. Learn. Mem. 138, 62–77. doi:10.1016/j.nlm.2016.07.031 Ehlers, M.D., 2003. Activity level controls postsynaptic composition and signaling via the ubiquitin-proteasome system. Nat. Neurosci. 6, 231–42. doi:10.1038/nn1013 Fonseca, R., 2013. Asymmetrical synaptic cooperation between cortical and thalamic inputs to the amygdale. Neuropsychopharmacology 38, 2675–87. doi:10.1038/npp.2013.178 Fonseca, R., 2012. Activity-dependent actin dynamics are required for the maintenance of long-term plasticity and for synaptic capture. Eur. J. Neurosci. 35, 195–206. doi:10.1111/j.1460-9568.2011.07955.x Fonseca, R., Nägerl, U.V., Bonhoeffer, T., 2006. Neuronal activity determines the protein synthesis dependence of long-term potentiation. Nat. Neurosci. 9, 478–80. doi:10.1038/nn1667 Fonseca, R., Nägerl, U.V., Morris, R.G.M., Bonhoeffer, T., 2004. Competing for memory: hippocampal LTP under regimes of reduced protein synthesis. Neuron 44, 1011–20. doi:10.1016/j.neuron.2004.10.033 Frey, U., Morris, R.G., 1998. Weak before strong: dissociating synaptic tagging and plasticity-factor accounts of late-LTP. Neuropharmacology 37, 545–52. Frey, U., Morris, R.G.M., 1997. Synaptic tagging and long-term potentiation. Nature 385, 533–536. doi:10.1038/385533a0 Govindarajan, A., Israely, I., Huang, S.-Y., Tonegawa, S., 2011. The dendritic branch is the preferred integrative unit for protein synthesis-dependent LTP. Neuron 69, 132–46. doi:10.1016/j.neuron.2010.12.008 Han, J.-H., Yiu, A.P., Cole, C.J., Hsiang, H.-L., Neve, R.L., Josselyn, S.A., 2008. Increasing CREB in the auditory thalamus enhances memory and generalization
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
25
of auditory conditioned fear. Learn. Mem. 15, 443–453. doi:10.1101/lm.993608 Kano, M., Ohno-Shosaku, T., Hashimotodani, Y., Uchigashima, M., Watanabe, M., 2009. Endocannabinoid-Mediated Control of Synaptic Transmission. Physiol. Rev. 89, 309–380. doi:10.1152/physrev.00019.2008 Karpova, A., 2006. Involvement of Protein Synthesis and Degradation in Long-Term Potentiation of Schaffer Collateral CA1 Synapses. J. Neurosci. 26, 4949–4955. doi:10.1523/JNEUROSCI.4573-05.2006 Kastellakis, G., Cai, D.J., Mednick, S.C., Silva, A.J., Poirazi, P., 2015. Synaptic clustering within dendrites: An emerging theory of memory formation. Prog. Neurobiol. doi:10.1016/j.pneurobio.2014.12.002 Kodirov, S.A., Jasiewicz, J., Amirmahani, P., Psyrakis, D., Bonni, K., Wehrmeister, M., Lutz, B., 2010. Endogenous cannabinoids trigger the depolarization-induced suppression of excitation in the lateral amygdala. Learn. Mem. 17, 43–9. doi:10.1101/lm.1663410 Lutz, B., Marsicano, G., Maldonado, R., Hillard, C.J., 2015. The endocannabinoid system in guarding against fear, anxiety and stress. Nat. Rev. Neurosci. 16, 705–718. doi:10.1038/nrn4036 Marsicano, G., Lutz, B., 1999. Expression of the cannabinoid receptor CB1 in distinct neuronal subpopulations in the adult mouse forebrain. Eur. J. Neurosci. 11, 4213–25. McKenzie, S., Eichenbaum, H., 2011. Consolidation and reconsolidation: two lives of memories? Neuron 71, 224–33. doi:10.1016/j.neuron.2011.06.037 Mechoulam, R., Parker, L.A., 2013. The endocannabinoid system and the brain. Annu. Rev. Psychol. 64, 21–47. doi:10.1146/annurev-psych-113011-143739 Pistis, M., Perra, S., Pillolla, G., Melis, M., Gessa, G.L., Muntoni, A.L., 2004. Cannabinoids modulate neuronal firing in the rat basolateral amygdala: evidence for CB1- and non-CB1-mediated actions. Neuropharmacology 46, 115–25. Puighermanal, E., Busquets-Garcia, A., Gomis-González, M., Marsicano, G., Maldonado, R., Ozaita, A., 2013. Dissociation of the pharmacological effects of THC by mTOR blockade. Neuropsychopharmacology 38, 1334–43. doi:10.1038/npp.2013.31 Rashid, A.J., Yan, C., Mercaldo, V., Hsiang, H.-L., Park, S., Cole, C.J., De Cristofaro, A., Yu, J., Ramakrishnan, C., Lee, S.Y., Deisseroth, K., Frankland, P.W., Josselyn, S.A., 2016. Competition between engrams influences fear memory formation and recall. Science (80-. ). 353, 383–387. doi:10.1126/science.aaf0594 Redondo, R.L., Morris, R.G.M., 2011. Making memories last: the synaptic tagging and capture hypothesis. Nat. Rev. Neurosci. 12, 17–30. doi:10.1038/nrn2963 Sajikumar, S., Frey, J.U., 2004. Late-associativity, synaptic tagging, and the role of dopamine during LTP and LTD. Neurobiol. Learn. Mem. 82, 12–25. doi:10.1016/j.nlm.2004.03.003 Sajikumar, S., Morris, R.G.M., Korte, M., 2014. Competition between recently potentiated synaptic inputs reveals a winner-take-all phase of synaptic tagging and capture. Proc. Natl. Acad. Sci. U. S. A. 111, 12217–21. doi:10.1073/pnas.1403643111 Sajikumar, S., Navakkode, S., Frey, J.U., 2007. Identification of compartment- and process-specific molecules required for “synaptic tagging” during long-term
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624
26
potentiation and long-term depression in hippocampal CA1. J. Neurosci. 27, 5068–80. doi:10.1523/JNEUROSCI.4940-06.2007 Sosanya, N.M., Cacheaux, L.P., Workman, E.R., Niere, F., Perrone-Bizzozero, N.I., Raab-Graham, K.F., 2015. Mammalian Target of Rapamycin (mTOR) Tagging Promotes Dendritic Branch Variability through the Capture of Ca2+/Calmodulin-dependent Protein Kinase II α (CaMKIIα) mRNAs by the RNA-binding Protein HuD. J. Biol. Chem. 290, 16357–71. doi:10.1074/jbc.M114.599399 Sumislawski, J.J., Ramikie, T.S., Patel, S., 2011. Reversible gating of endocannabinoid plasticity in the amygdala by chronic stress: a potential role for monoacylglycerol lipase inhibition in the prevention of stress-induced behavioral adaptation. Neuropsychopharmacology 36, 2750–61. doi:10.1038/npp.2011.166 Szabó, E.C., Manguinhas, R., Fonseca, R., 2016. The interplay between neuronal activity and actin dynamics mimic the setting of an LTD synaptic tag. Sci. Rep. 6, 33685. doi:10.1038/srep33685 Tonegawa, S., Pignatelli, M., Roy, D.S., Ryan, T.J., 2015. Memory engram storage and retrieval. Curr. Opin. Neurobiol. 35, 101–109. doi:10.1016/j.conb.2015.07.009 Turu, G., Hunyady, L., 2010. Signal transduction of the CB1 cannabinoid receptor. J. Mol. Endocrinol. 44, 75–85. doi:10.1677/JME-08-0190 Yiu, A.P., Mercaldo, V., Yan, C., Richards, B., Rashid, A.J., Hsiang, H.-L.L., Pressey, J., Mahadevan, V., Tran, M.M., Kushner, S.A., Woodin, M.A., Frankland, P.W., Josselyn, S.A., 2014. Neurons are recruited to a memory trace based on relative neuronal excitability immediately before training. Neuron 83, 722–35. doi:10.1016/j.neuron.2014.07.017
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted January 23, 2019. ; https://doi.org/10.1101/526624doi: bioRxiv preprint
https://doi.org/10.1101/526624