-
ORIGINAL RESEARCH ARTICLEpublished: 06 February 2013
doi: 10.3389/fncir.2013.00010
Chronic stress disrupts neural coherence betweencortico-limbic
structuresJoão Filipe Oliveira 1,2†, Nuno Sérgio Dias1,2,3†,
Mariana Correia 1,2, Filipa Gama-Pereira 1,2,Vanessa Morais
Sardinha1,2, Ana Lima1,2, Ana Filipa Oliveira 1,2, Luís Ricardo
Jacinto1,2,4,Daniela Silva Ferreira 1,2, Ana Maria Silva1,2, Joana
Santos Reis1,2, João José Cerqueira 1,2 andNuno Sousa1,2*1 School
of Health Sciences, Life and Health Sciences Research Institute
(ICVS), University of Minho, Braga, Portugal2 ICVS/3B’s - PT
Government Associate Laboratory, Braga/Guimarães, Portugal3 DIGARC,
Polytechnic Institute of Cávado and Ave, Barcelos, Portugal4
Department of Industrial Electronics, University of Minho - Campus
de Azurém, Guimarães, Portugal
Edited by:Michael Brecht, HumboldtUniversity Berlin, Germany
Reviewed by:Christiaan P. De Kock, VU UniversityAmsterdam,
NetherlandsDaoyun Ji, Baylor Collegeof Medicine, USA
*Correspondence:Nuno Sousa, Instituto de Ciênciasda Saúde e da
Vida (ICVS),Escola de Ciências daSaúde - Universidade do
Minho,Campus de Gualtar,4710-103 Braga, Portugal.e-mail:
[email protected]†These authors equally contributedto this
work.
Chronic stress impairs cognitive function, namely on tasks that
rely on the integrityof cortico-limbic networks. To unravel the
functional impact of progressive stress incortico-limbic networks
we measured neural activity and spectral coherences between
theventral hippocampus (vHIP) and the medial prefrontal cortex
(mPFC) in rats subjected toshort term stress (STS) and chronic
unpredictable stress (CUS). CUS exposure consistentlydisrupted the
spectral coherence between both areas for a wide range of
frequencies,whereas STS exposure failed to trigger such effect. The
chronic stress-induced coherencedecrease correlated inversely with
the vHIP power spectrum, but not with the mPFCpower spectrum, which
supports the view that hippocampal dysfunction is the primaryevent
after stress exposure. Importantly, we additionally show that the
variations invHIP-to-mPFC coherence and power spectrum in the vHIP
correlated with stress-inducedbehavioral deficits in a spatial
reference memory task. Altogether, these findings resultin an
innovative readout to measure, and follow, the functional events
that underlie thestress-induced reference memory impairments.
Keywords: chronic stress, coherence, power spectrum,
hippocampus, prefrontal cortex
INTRODUCTIONStress is a constant in the daily life of the modern
societies. Eachsubject is constantly challenged and threatened by a
large vari-ety of unpredicted events. Under stressful situations, a
primaryresponse is set up, in order to restore homeostasis and
pro-mote behavioral adaptation; however, prolonged stress
exposuremay trigger maladaptive responses that lead to severe
manifes-tations such as learning and memory deficits or anxious
anddepressive-like behavior (Popoli et al., 2011; Sousa and
Almeida,2012). Whereas hippocampal structural damage and
impairedplasticity were initially recognized to underlie these
manifesta-tions (Sousa et al., 2000), subsequent studies
demonstrated thatthe medial prefrontal cortex (mPFC), an area
intimately relatedwith working memory processes and cognitive
stimuli integra-tion, is also a key target of chronic stress
(Cerqueira et al., 2007a,b;Liston et al., 2009). At a functional
level, such neuronal compro-mise was correlated with a strong
impairment of plasticity in thehippocampus-PFC pathway (Cerqueira
et al., 2007a).
Although mechanisms of plasticity are generally accepted
asreadouts of interregional connections, much information
onneuronal dynamics is lost due to the supra-physiological
pro-tocols typically used. Since previous data suggests that
chronicunpredictable stress (CUS) triggers disconnections in
specificbrain circuits, additional measures such as power spectra
andphase coherence of local field potentials (LFPs) under
stressfulconditions urge to be assessed. LFPs reproduce
summated
individual conductance and synaptic inputs of networks com-posed
by ensembles of firing neurons and surrounding glia, andtherefore
are an excellent readout of network dynamics (Buzsáki,2006, 2010;
Perea et al., 2009; Jia et al., 2011). LFPs reflect thetemporal
pattern of activity that acts on local networks that aredirectly
connected (Varela et al., 2001); importantly, the
ventralhippocampus (vHIP) and the prelimbic subfield of the
medialPFC are linked by monosynaptic connection (Jay and
Witter,1991; Thierry et al., 2000; Tierney et al., 2004). Moreover,
phasecoherence between the hippocampus and cortex was proposedto
correlate with acquisition of information and memory for-mation
(Buzsáki, 1996; Popa et al., 2010; Fell and Axmacher,2011). In
fact, previous studies using maze-based paradigmsto test different
types of memory, reveal that rats displayedincreased theta phase
coherence between the PFC and the HIPby the time they take a
decision based in a previous expe-rience (Jones and Wilson, 2005;
Benchenane et al., 2010); inaddition, hippocampal theta rhythms
were shown to synchro-nize with PFC single cell (Siapas et al.,
2005) and field activities(Hyman et al., 2011). Beta coherence
between hippocampusand mPFC was also shown to be necessary for
effective com-munication during visual object processing (Sehatpour
et al.,2008). Interestingly, impaired coherence in the
hippocampus-mPFC was observed after loss of PFC function
(Brockmannet al., 2011) and in models of schizophrenia (Sigurdsson
et al.,2010).
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 1
NEURAL CIRCUITS
http://www.frontiersin.org/Neural_Circuits/editorialboardhttp://www.frontiersin.org/Neural_Circuits/editorialboardhttp://www.frontiersin.org/Neural_Circuits/editorialboardhttp://www.frontiersin.org/Neural_Circuits/abouthttp://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.org/Neural_Circuits/10.3389/fncir.2013.00010/abstracthttp://www.frontiersin.org/Community/WhosWhoActivity.aspx?sname=Jo�oOliveira_1&UID=72372http://www.frontiersin.org/Community/WhosWhoActivity.aspx?sname=NunoDias&UID=75957http://www.frontiersin.org/Community/WhosWhoActivity.aspx?sname=LuisJacinto&UID=75940http://www.frontiersin.org/Community/WhosWhoActivity.aspx?sname=Jo�oCerqueira&UID=2488http://www.frontiersin.org/Community/WhosWhoActivity.aspx?sname=NunoSousa&UID=2465mailto:[email protected]:[email protected]://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
Based on these assumptions we hypothesize that stress-induced
deleterious effects on the function of the PFC andhippocampus may
be in a large extent due to alterations inphase coherence between
these areas. The present study testedthis hypothesis by recording
neural activity simultaneously fromthe mPFC and vHIP in rats
exposed to short-term and chronicstress; the use of different
periods of exposure to stress providedan insight of the temporal
dynamics of the onset of stress-inducedchanges. Power spectrum
densities and coherence were also ana-lyzed, and correlations
between them were studied and comparedwith classical long-term
potentiation measurements. In addi-tion, in a separate set of rats,
variations in coherence and powerspectrum were correlated with
behavioral performance.
RESULTSSTRESSED RATS DISPLAY INCREASED POWER SPECTRUM
DENSITIESIN THE vHIP AND mPFCPower spectra translate the amplitude
of the signals recorded ina brain region on the frequency domain.
power spectrum density
(PSD) analysis of the recorded LFPs from the mPFC and vHIP ofall
studied rats allowed the thorough characterization of poweractivity
in a wide range of frequencies (Delta, 1–4 Hz; Theta,4–12 Hz; Beta,
12–20 Hz; Low Gamma, 20–40 Hz; High Gamma,40–90 Hz) for those
regions at a basal state.
Regarding the function of the vHIP, short term stress
(STS)triggered an increase of PSD in the theta, beta, low gamma
andhigh gamma frequencies, while in the delta band no signifi-cant
variation to controls (CON) was recorded (Figures 1A–C;see Table 1
for statistical values). Importantly, CUS exposure-induced a
persistent increase in PSD in all frequency bandsanalyzed which was
always higher than controls and STS(Figures 1A–C; see Table 1).
In contrast, PSD in the mPFC was only affected by the expo-sure
to STS in beta and low gamma frequencies when comparedto control
rats (Figures 1D–F; Table 1); however, exposure toCUS triggered an
increase in the PSD in the theta, beta andlow gamma and high gamma
frequency bands (Figures 1D–F;Table 1).
FIGURE 1 | Stressed rats show increased PSD in multiple
frequencybands both in the ventral hippocampus and medial PFC.
Representativetraces of raw data (black line) recorded
simultaneously from the vHIP (A) andmPFC (D) of a rat of each
group; the red line represents theta filteredcomponent, as example.
Power Spectral Density (PSD) values of the vHIP(B) recordings and
the mPFC (E) recordings, for controls (CON), short-term
stress (STS), and chronic unpredictable stress (CUS); each
horizontal line inthe Y-axis represents the spectrogram of an
individual rat. Group comparisonof the PSD values from vHIP (C) and
mPFC (F) in the delta (1–4 Hz), theta(4–12 Hz), beta (12–20 Hz),
low gamma (20–40 Hz), and high gamma(40–90 Hz) frequency bands.
∗Statistically different from CON, p < 0.05;#Statistically
different from CUS, p < 0.05; error bars represent SD.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 2
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
Table 1 | Statistical values of the group comparisons of PSD and
coherence for each frequency band in the first set of animals.
CON vs. STS
Band vHIP PSD mPFC PSD
delta theta beta low gamma high gamma delta theta beta low gamma
high gamma
Z -value −1.47 −2.06 −2.29 −2.18 −2.41 0.00 −1.59 −2.06 −2.06
−0.53p-value 0.14 0.03 0.01 0.02 0.01 1.00 0.11 0.03 0.03 0.60
Band Coherence
delta theta beta low gamma high gamma
Z -value 2.76 2.18 1.94 1.23 −0.05p-value 0.00 0.06 0.08 0.22
0.95
CON vs. CUS
Band vHIP PSD mPFC PSD
delta theta beta low gamma high gamma delta theta beta low gamma
high gamma
Z -value −2.95 −3.32 −3.32 −3.32 −3.32 −0.98 −2.38 −2.38 −2.76
−1.92p-value 0.00 0.00 0.00 0.00 0.00 0.33 0.01 0.01 0.00 0.05
Band Coherence
delta theta beta low gamma high gamma
Z -value 2.76 2.76 3.04 2.29 1.26
p-value 0.00 0.00 0.00 0.01 0.21
STS vs. CUS
Band vHIP PSD mPFC PSD
delta theta beta low gamma high gamma delta theta beta low gamma
high gamma
Z-value −1.88 −2.52 −2.66 −2.59 −2.02 −1.81 −1.31 −0.03 −0.81
−1.59p-value 0.05 0.00 0.00 0.00 0.04 0.06 0.19 0.97 0.42 0.11
Band Coherence
delta theta beta low gamma high gamma
Z -value −1.03 1.10 1.81 1.88 1.17p-value 0.31 0.27 0.06 0.05
0.24
Each table represents the Z- and p-values of pairwise
comparisons of vHIP PSD, mPFC PSD, and coherence values in the
control (CON), short-term stress (STS),
and chronic unpredictable stress (CUS) groups. These values
correspond to the graphs presented in Figures 1 and 2. Comparisons
made by Mann–Whitney test;
statistical significance is highlighted in yellow (p ≤
0.05).
STRESS DECREASES SPECTRAL COHERENCE BETWEEN THE vHIPAND
mPFCCoherence between brain regions measures the matching
oftemporal structure in signals recorded from those regions.
Thetemporal structure, or phase, is a powerful measure since it
doesnot depend on signal amplitude and two signals are said to
besynchronous if their rhythms’ phase match (Varela et al.,
2001).Coherence analysis of the signals obtained simultaneously
fromthe vHIP and mPFC was used to study phase coherence
betweenthese regions of the brain (Figure 2). The exposure of the
ratsto STS caused a significant decrease of phase coherence in
the
delta band when compared to the observed in controls
(CON)(Figure 2A for individual analysis, Figure 2B for group
compari-son by frequency band; statistic results on Table 1).
Interestingly,exposure to CUS-induced a pronounced decrease in
phase coher-ence when comparing to CON rats in the delta, theta,
alpha,beta and low gamma bands, and in the low gamma band
whencompared to STS; in the high gamma band, coherence levelsof CUS
subjects were statistically undistinguishable from thoseobtained in
the CON rats (Figures 2A,B; Table 1) and thereforefurther analysis
was focused on the data observed for frequenciesbelow 40 Hz.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 3
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
FIGURE 2 | Chronic stress decreases the coherence between
theventral hippocampus (vHIP) and the medial prefrontal
cortex(mPFC). (A) Spectral coherence of controls (CON), short term
stress(STS), and chronic unpredictable stress (CUS); each
horizontal line in theY-axis represents the spectrogram of an
individual rat (B) Group
comparison of the coherence values between mPFC and vHIP for
delta(1–4 Hz), theta (4–12 Hz), beta (12–20 Hz), low gamma (20–40
Hz), andhigh gamma (40–90 Hz) frequency bands. ∗Statistically
different fromCON, p < 0.05; #Statistically different from CUS,
p < 0.05; error barsrepresent SD.
STRESS-INDUCED LOSS OF COHERENCE CORRELATES WITHINCREASED PSD IN
THE vHIPIn order to disclose whether stress affects the
relationship betweenpower activity and coherence observed in the
vHIP and mPFCof CON rats, a correlation between both measures was
explored(Figure 3).
PSD recorded from the vHIP of CON rats did not correlate
sig-nificantly with vHIP-mPFC coherence in the range of
frequencybands analyzed (0–40Hz; CON in Figures 3A,B). Exposure
tostress triggers a negative correlation between theta vHIP-PSD
anddelta-theta and low gamma coherence which links the
observa-tions in Figures 1 and 2, where coherence decreases as the
powerincreases in vHIP as a result of exposure to stress. Such
negativecorrelations are absent between vHIP-mPFC coherence and
PSDrecorded from the mPFC of stressed animals suggesting that
thesetwo measures are not linked. Interestingly, CON rats show a
clearpositive correlation between theta coherence and all
frequen-cies in mPFC PSD which seems to be disrupted by the
chronicexposure to stress (Figures 3C,D).
STRESS EXPOSURE IMPAIRED LTP INDUCTION IN THEHIPPOCAMPAL-mPFC
LINKTo compare the above described results with classical
electrophys-iological measures of synaptic plasticity, we have
analyzed theability to induce LTP in this neuronal connection. When
sub-jected to high frequency stimulation (HFS) in the vHIP, the
slopeof evoked PSPs recorded in the mPFC of CON rats increasedabout
52.1 ± 9.0% and remained elevated for at least 90 min, inthe form
of long term potentiation (LTP; as a measure of neural
plasticity between the two regions; Figure A1A). The CUS
ratsdisplayed the previously described impairment of LTP betweenthe
vHIP and mPFC (CUS, 19.6 ± 4.7%; p < 0.05) (Cerqueiraet al.,
2007a), while the STS rats maintained the LTP valueswhen compared
to the CON rat (STS, 43.5 ± 3.5%; p > 0.05;Figure A1B).
BEHAVIORAL PERFORMANCE CORRELATES WITH VARIATIONS INCOHERENCE
AND POWER SPECTRUM DENSITYAn independent set of control and
chronically stressed rats wastested for cognitive function in order
to assess an eventual linkbetween the chronic stress-induced
coherence impairment andpower increase observed in the first set of
rats and the previ-ously described cognitive impairments triggered
by the chronicstress protocols described in the literature
(Cerqueira et al., 2007a;Dias-Ferreira et al., 2009). The analysis
of corticosterone lev-els as a measure of efficacy of the stress
protocol confirmedthat the CUS protocol induced a chronic stress
state in thestressed rats (CON, 48.2 ± 5.9 ng/ml; CUS, 414.8 ± 41.5
ng/ml;p = 0.002).
Chronic stress triggered deficits in the spatial reference
mem-ory task, since CUS rats swim longer that controls to find
theplatform and the learning curves of CON and CUS are
signifi-cantly different (Figure 4A; p = 0.02). Post-hoc analysis
compar-isons revealed significant impairments in reference memory
inCUS rats on day 2 (Figure 4A; day 2, t-value = 3.126, p <
0.05;day 1, t-value = 2.178; day 3, t-value = 1.285; day 4, t-value
=1.480; p > 0.05). Importantly, the analysis of swimming speed
ofeach group reported that both CON and CUS rats swim at a same
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 4
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
FIGURE 3 | Correlations between spectral coherence and power
spectraldensities in ventral hippocampus (vHIP) and medial
prefrontal cortex(mPFC). The graphs present Pearson values for
correlations betweenvHIP-mPFC spectral coherence and vHIP (A,B) or
mPFC (C,D) power spectraldensities and respective p-values against
null hypothesis (corr = 0), for CON
and CUS rats (A,C and B,D, respectively). For each brain region,
statisticallysignificant correlations are represented in the right
panel by color above darkblue (p < 0.05); the direction of the
correlation, positive or negative, is givenin the left panel by the
color code (green to red, positive correlation; green toblue,
negative correlation).
speed in the water maze (Figure A2; p > 0.05) and post-hoc
anal-ysis sustained this observation for each day (day 1, t-value =
0.81;day 2, t-value = 0.93; day 3, t-value = 2.38; day 4, t-value =
1.83;p > 0.05), excluding any locomotor deficit due to the
chronicstress treatment. Additionally, chronic stress triggered a
behav-ioral flexibility impairment, since CUS rats spent less time
in thenew quadrant that control rats (Figure 4B; p = 0.04) in the
rever-sal learning task of the Morris Water Maze, in agreement with
datapreviously reported (Cerqueira et al., 2007a).
Again, we confirmed in this additional set of rats that
CUStriggered a general increase in power recorded in the vHIPwhen
compared to the CON rats (Figures 4C, A3A; Table 2).Similarly, the
CUS rats displayed an increased PSD in the mPFCwhen compared to the
CON rats (Figures 4D, Figure A3B;Table 2). Regarding the coherence
between the vHIP and mPFC(Figure 4E), the stressed rats displayed a
general decrease ofcoherence in a wide range of frequency bands
when comparedto controls, except for the low gamma band, where a
similar ten-dency was observed although not significant (Figure 4F;
Table 2).These results were in accordance with the data recorded
for thefirst set of rats, sustaining a profound affection of the
networkafter chronic exposure to unpredictable stressors.
Subsequently, we searched for pairwise correlations
amongcognitive performance in each day of the spatial reference
mem-ory task, cognitive performance in the reversal learning
task,coherence between the mPFC and vHIP and power spectra inthese
brain regions (Figures 4G,H). Data shows that the dif-ferent
performance of stressed and control rats on the second
day of the reference memory task inversely correlates with
thevalue of coherence measured between the vHIP and mPFC foreach of
those rats (Figure 4H) in the delta (r = −0.79), theta(r = −0.88)
and high gamma (r = −0.77) bands, meaning thatthe coherence between
the two regions is crucial for the good per-formance in this task.
Additionally, we observed that the increasein theta power in the
vHIP directly correlates with higher escapelatencies in the second
day of test (Figure 4H; r = 0.71), disclos-ing a link between
stress-triggered increase in theta vHIP powerand worse performance
in this task by CUS rats.
DISCUSSIONExposure to CUS was previously shown to induce
deleteriouseffects at a morphological level in the hippocampus
(Sousa et al.,2000) and mPFC (Cerqueira et al., 2007b), which were
sug-gested to underlie the deficits observed in behavior
paradigmsthat rely on those areas in rodents (Cerqueira et al.,
2007a; Dias-Ferreira et al., 2009) and in humans (Soares et al.,
2012). In thiswork, using anesthetized rats, we show an unequivocal
decreasein phase coherence in stressed subjects when compared
withtheir non-stressed rats. Additionally, PSD extracted from
LFPsrecorded in the mPFC of the CUS subjects increased
signifi-cantly between 4 and 40 Hz, while a global increase of PSD
inthe vHIP was observed for both STS and CUS rats. We showfinally
that the decrease in delta, theta and high gamma coherenceand the
increase in theta power in the vHIP correlate with theworse
performance in the reference memory task of the MorrisWater
Maze.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 5
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
FIGURE 4 | CUS-induced changes in PSD and coherence correlate
withimpairments in cognitive performance. (A,B) Cognitive
performance of thestudied rats in water maze based tests for
spatial reference memory andbehavioral flexibility; (A) Learning
curves of the reference memory task ofcontrol (CON) and chronically
stressed (CUS) rats. (B) Average trial time inthe new quadrant
given as a percentage of the total escape latency for thereversal
learning task. (C–F) Analysis of the LFP signals recorded in the
vHIPand mPFC of the CON and CUS rats; Power Spectral Density (PSD)
values ofthe vHIP (C) recordings and the mPFC (D) recordings, for
controls (CON)and chronically stressed (CUS); (E) Spectral
coherence of controls (CON) and
chronic unpredictable stress (CUS); in (C–E) each horizontal
line in the Y-axisrepresents the spectrogram of an individual rat;
(F) Group comparison of thecoherence values between mPFC and vHIP
for delta (1–4 Hz), theta(4–12 Hz), beta (12–20 Hz), low gamma
(20–40 Hz), and high gamma(40–90 Hz) frequency bands. (G,H)
Correlation between behavior andelectrophysiological performances
of the recorded rats; (G) p-values forPearson correlations between
behavior and electrophysiologicalperformances for each rat;
significant correlations highlighted in yellow;(H) Correlation plot
for each significant correlation observed in G.*Statistically
different from CON, p < 0.05; error bars represent SD.
The impact of stress in PSD and coherence between
intercon-nected brain regions is currently under scrutiny. A recent
studyreported a decrease in coherence around 4 Hz between both
areasbut no significant changes in PSD (Lee et al., 2011). The
presentdata is quite distinct to the findings of that study.
Although bothsets of data report to the connection between the HIP
and PFC,differences in experimental conditions may explain these
appar-ently discrepant findings. First, in the present study we
haverecorded from a more ventral hippocampal region than the
onerecorded in the Lee et al. study; this is very relevant, since
ourplacement was chosen based on evidence that the monosynap-tic
hippocampus-PFC connection originates in the vHIP (Jay andWitter,
1991). Moreover, it has been shown that the electrophys-iological
response of the vHIP to stress is remarkably different
from the one observed in more dorsal regions (Maggio and
Segal,2009). The second methodological aspect that deserves merit
tobe highlighted is that the stress paradigms used in both
studieswere also different. While Lee et al. (2011) used a repeated
stressparadigm, the unpredictability of stressor presentation of
the CUSprotocol used in our work was shown to be crucial for the
stress-consequence severity in the neural regions affected (Maier
andWatkins, 2005; Amat et al., 2006); this might also justify
theseverity of effects observed in our data.
CUS rats presented significant increases of PSD, when com-pared
to non-stressed rats, in both vHIP and mPFC. Interestingly,the
present results show that STS also increase PSD, but
moreextensively in the vHIP. Since power activity translates the
ampli-tude of the signals recorded in a brain region for a
given
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 6
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
Table 2 | Statistical values of the group comparisons of PSD and
coherence for each frequency band in the second set of animals.
CON vs. CUS
Band vHIP PSD mPFC PSD
delta theta beta low gamma high gamma delta theta beta low gamma
high gamma
Z -value −1.87 −1.87 −1.87 −1.87 −1.87 −1.87 −2.16 −2.16 −2.16
−1.87p-value 0.05 0.05 0.05 0.05 0.05 0.05 0.02 0.02 0.02 0.05
Band Coherence
delta theta beta low gamma high gamma
Z -value 1.87 2.16 1.87 1.01 2.16
p-value 0.05 0.02 0.05 0.34 0.02
Each table represents the Z- and p-values of pairwise
comparisons of vHIP PSD, mPFC PSD, and coherence values in the
control (CON) and chronic unpredictable
stress (CUS) groups. These values correspond to the graphs
presented in Figure 4. Comparisons made by Mann–Whitney test;
statistical significance is highlighted
in yellow (p ≤ 0.05).
frequency, which in turn represents the extent of neuron
recruit-ment to fire on a particular rhythm, one may conclude
thatexposure to stress caused an increase in neuronal activity,
morepronounced in the vHIP, that progresses with time also to
otherbrain regions, such as the mPFC. This temporal sequence
hasbeen previously proposed by our behavioral data (Cerqueira et
al.,2007a) and is of great relevance to understand the
progressionof cognitive deficits in stress-related pathologies. The
under-lying mechanisms for the augmented PSD are still
unknown.However, it seems reasonable to admit that an imbalance
inexcitatory/inhibitory inputs is present. Several lines of
evidencepoint on this direction. On one hand, stress exposure was
shownto enhance the release of excitatory neurotransmitters (such
asglutamate, dopamine, norepinephrine, or acetylcholine) in
thehippocampus and PFC (Finlay et al., 1995; Mark et al.,
1996;Feenstra, 2000; Moghaddam, 2002; Dazzi et al., 2004); on
theother hand, it is known that GABAergic activity is reduced
afterstress exposure in these brain regions (Otero Losada, 1988;
Hasleret al., 2010). Interestingly, once the increase in PSD is
establishedin the vHIP, the changes in the mPFC might also be due
to an aug-mented afferent excitatory input; in this case an early
increase inthe activation of hippocampal neurons could lead
consequentlyto overactivation of monosynaptically connected (Jay
and Witter,1991; Thierry et al., 2000; Tierney et al., 2004)
neurons in themPFC. Such uncontrolled and desynchronized neuronal
activityincrease is likely to underlie the striking decrease in
coherencebetween the vHIP and the mPFC observed after CUS, which
maytrigger a compensatory mechanism, translated by an
increasedpower activity, in order to attempt to restore the
decreased coher-ence in this neuronal network. This detrimental
effect fits with theloss of important connections during neuronal
atrophy verified instressed subjects (Cerqueira et al.,
2007a,b).
Coherence, as a measure of synchronization between the
oscil-lations of cortical and limbic regions, was shown to
underliegood behavior performance, namely supporting memory
pro-cessing (Buzsáki, 1996, 2010; Jones and Wilson, 2005; Siapas et
al.,2005; Benchenane et al., 2010; Popa et al., 2010; Hyman et
al.,2011). In this scope, it is predictable that a decrease of
coherencemay cause impairments in the functions attributed to the
regions
that are connected by such circuits. Indeed, decreased
coher-ence in the hippocampus-mPFC was shown to affect directly
thePFC function (Brockmann et al., 2011). Based in the
presentobservations, we suggest that the pronounced loss of
vHIP-mPFC coherence between 1–40 Hz after chronic stress
exposureunderlie the previously described stress-related behavior
impair-ments, which are dependent on the normal hippocampal
andprefrontal function in those frequency bands (Cerqueira et
al.,2007a). Importantly, the CUS-induced decrease in
vHIP-mPFCcoherence is associated with an enhancement of neuronal
activ-ity, measured by PSD, in coincident frequencies in both
regions.The correlation of these two electrophysiological
phenomenarevealed an inverse correlation between vHIP-mPFC
coherenceand the hippocampal function, but not between mPFC
activityand vHIP-mPFC coherence. This indicates a putative leading
roleof hippocampal firing on the disruption of this direct
pathwayby chronic stress exposure; the hippocampal disruption
would,in turn, compromise the behavior outputs dependent on
thispathway. In fact, these observations confirm previously
reportedbehavioral data, in which reference memory
(hippocampal-dependent) was impaired after 3 days of stress, while
behaviorflexibility (PFC-dependent) was only after a longer period
ofexposure to stress (Cerqueira et al., 2007a). Accordingly,
thissequential behavior impairment is supported by a reduction
ofspine density and apical dendrite arborization of pyramidal
neu-rons of layers II/III of the PFC, precisely where the
hippocampalafferents establish synapses (Radley et al., 2004;
Cerqueira et al.,2007b). Finally, it is noteworthy that after STS
rats show a gen-eral increase in vHIP PSD, which is less extent in
the mPFC of thesame rats, and may indicate an early affection of
the hippocampalneural activity by stress. Interestingly, in
contrast to CUS, STS ratsdo not present yet signs of LTP
impairment.
In addition, we confirmed that the decrease of
vHIP-mPFCcoherence and increase in PSD is linked to the poorer
cognitiveperformance of stressed rats, as we found significant
correlationsbetween reference memory performance and
electrophysiologi-cal data. This evidence is of major relevance
since it establishesfor the first time the link between the
affection of the neuralnetwork after chronic stress exposure and
the cognitive decline
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 7
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
observed in the stressed subjects. Importantly, it confirms
thatcoherence and neural activity, namely in the theta band, is
severelyaffected in the stressed rats and is disruptive for the
entrainmentof the hippocampal-cortical connection and,
consequently, forgood behavior performance (Buzsáki, 1996; Jones
and Wilson,2005; Siapas et al., 2005; Benchenane et al., 2010; Popa
et al., 2010;Hyman et al., 2011).
In summary, the present data reveals that chronic stress
pri-marily affects hippocampal activity, which in turn will
inducedamage in prefrontal dendritic trees, disrupting the
HIP-PFCconnection and triggering a decrease in coherence between
thetwo areas. Since neural coherence is a measure of regional
inter-play, behavior outcomes dependent on this interplay are
severelyaffected, as showed by direct correlations between
decreasein vHIP-mPFC coherence and reference memory
performance.These findings provide a novel readout of the dynamics
of stressdeleterious effects in corticolimbic networks and are of
relevanceto better understand the mechanisms underlying
stress-inducedimpairment in cognitive performance.
MATERIALS AND METHODSANIMALS AND TREATMENTSAll experiments were
conducted in accordance with localregulations (European Union
Directive 86/609/EEC) andNational Institutes of Health guidelines
on animal care andexperimentation.
Two months-old male Wistar–Han rats (Charles RiverLaboratories,
Spain) were housed in groups of two under stan-dard laboratory
conditions (lights on from 8:00 A.M. to 8:00 P.M.;room temperature
22◦C; ad libitum access to food and water). Agroup of rats was
submitted to 28 days of CUS. Briefly, rats wereexposed to a
stressor (1 h/day) of one of several aversive stimuli[cold water
(18◦C), restraint, overcrowding, exposure to a hot airstream, noise
or shaking]; the stressors were presented in a ran-dom order. This
stress paradigm was shown previously to inducepersistently a
pattern of consequences characterized by elevatedplasma levels of
corticosterone, the primary glucocorticoid of therat, and reduced
thymus weights (Cerqueira et al., 2007a). STSwas induced to a
second group of rats with one stressor daily (sim-ilar to CUS) for
6 days. A third group of rats was handled daily andserved as
controls (CON). Electrophysiological recordings wereperformed for
each rat the day after the end of treatment.
An additional set of rats was used to search for
correlationsbetween behavioral performance and variations in power
spec-tra in vHIP and mPFC and in coherence between these
regions.Corticosterone levels were measured in blood serum
sampledbetween 9:00 and 10:00 A.M. using a commercially
availableELISA kit (Cayman Chemical, USA).
SURGERYRats were anesthetized with sodium pentobarbital (60
mg/kg,i.p., supplemented each 60 min throughout the experiment)
andplaced in a stereotaxic frame. Rectal temperature was
main-tained at 37◦C by a homoeothermic blanket (Stoelting,
Ireland).Experimental procedures for implantation and recording
extra-cellular field potentials in the prelimbic area (PL) of
themedial PFC have been described previously (Rocher et al.,
2004;Cerqueira et al., 2007a). Briefly, Platinum/Iridium
recording
electrodes (Science Products, Germany) were placed in the
PL(coordinates, 3.3 mm anterior to bregma, 0.8 mm lateral tothe
midline, 4.0 mm below bregma) and a concentric bipo-lar
tungsten/stainless-steel electrode (WPI, USA) was positionedinto
the ipsilateral CA1/subicular region of the vHIP (coordi-nates, 6.5
mm posterior to bregma, 5.5 mm lateral to the midline,5.3 mm below
bregma), according to the atlas of Paxinos andWatson (2005).
In vivo RECORDING OF LOCAL FIELD POTENTIALS IN THE mPFC
ANDVENTRAL HIPPOCAMPUSLFP signals obtained from both electrodes
were amplified, filtered(0.1–3000 Hz, LP511 Grass Amplifier,
Astro-Med, Germany),acquired (Micro 1401 mkII, CED, UK) and
recorded on apersonal computer running Signal Software (CED, UK).
Afterreaching fully anesthetized state, 100 s of local field
activity wasrecorded at the sampling rate of 250 Hz (for
representative tracessee Figures 1A–D). After the
electrophysiological protocols, abiphasic 1 mA stimulus was
delivered to both electrodes. Ratswere sacrificed and perfused a
solution of paraformaldehyde 4%.Brains were carefully removed and
sectioned in 50 µm slices.Brain slices containing the PFC and
hippocampus were processedwith a light Giemsa staining for
determination of electrode posi-tion; the lesion caused by the
biphasic stimulus identified theelectrode position and the
recordings were discarded wheneverone of the electrodes failed the
targeted position (about 12%of the recordings). In total we
analyzed 6 CON, 9 STS, and 12CUS rats; an additional set of 4 CON
and 4 CUS were usedto establish correlations between behavioral
performance andelectrophysiological data (Figure A4).
In vivo STUDY OF SYNAPTIC PLASTICITY BETWEEN THE mPFC ANDVENTRAL
HIPPOCAMPUSSynaptic plasticity was tested in anesthetized rats by
inducing LTPas described previously (Cerqueira et al., 2007a).
Briefly, the elec-trode inserted in the CA1/subicular region was
used to deliverstimuli to induce a characteristic monosynaptic
field excitatorypostsynaptic potential (fEPSP) in the PFC (Figure
1A; inset). Testpulses (100 ms) were delivered every 30 s at an
intensity enoughto evoke a potential about 70% of its maximum
(250–500 µA;S88X Grass Stimulator, Astro-Med, Germany). The evoked
poten-tial to such stimulation is likely to reflect summated PSPs.
Basalresponses were recorded during 30 min and followed by
LTPinduction, which was obtained performing HSF, that consistedof
two series of 10 trains (250 Hz, 200 ms) at 0.1 Hz, 6 min
apart,delivered at test intensity. The size of LTP induction was
measuredby changes in the slope of responses to additional 90 min
stimu-lating each 30 s. PSP slopes were analyzed using Signal
software(CED, UK; sampling rate, 10 KHz) and expressed as a
percent-age change of the mean responses to basal stimulation
before andafter HFS.
BEHAVIORAL ANALYSISBehavioral tests were conducted as described
previously(Cerqueira et al., 2007a), using water maze-based tests
to assessspatial reference memory and behavior flexibility. The
spatialreference memory task, highly dependent of the integrity
ofthe function of the hippocampus (Morris, 1984), assesses the
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 8
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
ability of rats to learn the position of the hidden platform,
basedon external cues placed outside the pool. Rats were placed,
facingthe wall of the maze, in each of the four imaginary pool
quadrantsat the beginning of each of the four daily trials. A trial
is con-sidered complete when the rat escapes onto the platform;
whenthis escape fails to occur within 120 s, the rat is gently
guided tothe platform and an escape latency of 120 s is recorded
for thattrial. Rats are allowed to spend 30 s on the escape
platform beforebeing positioned at a new starting point. Time spent
to reach theplatform (escape latency) was recorded in the
consecutive trials.During the reference memory task, the platform
was set in thesame imaginary pool quadrant and rats were tested
during 4 days.Data on average escape latencies to the platform on
days 1–4 wereanalyzed as a reference memory test.
To perform the reversal learning task, a PFC-dependent func-tion
(De Bruin et al., 1994), on day 5, the escape platform
waspositioned in the opposite (new) quadrant and rats were testedin
a 4-trial paradigm, similar to that described above. For
thisreverse-learning task, time spent swimming in the new
quadrantwere recorded and analyzed as a measure of behavioral
flexibility.
DATA ANALYSIS AND STATISTICSThe PSD of PFC and hippocampus
regions, as well as the coher-ence between both regions, were
performed on the 2-channel100 s long LFP signals acquired at the
start of the experimentfor each rat. Each measure was applied on 1
s long segments andthe average of all segments was considered for
statistical groupanalysis. Both measures were assessed in a wide
range of fre-quencies: delta (1–4 Hz); theta (4–12 Hz); beta (12–20
Hz); lowgamma (20–40 Hz); high gamma (40–90 Hz). All LFP
recordswere thoroughly inspected and those that presented
significantnoise corruption were excluded from further analyses.
PSD andcoherence were calculated with custom-written
MATLAB-basedprograms (MathWorks, Natick, MA).
PSD analysisThe PSD of each channel was calculated through the
10 × log10of the multiplication between the complex Fourier
transform ofeach 1 s long data segment and its complex conjugate.
The meanPSD of each channel was evaluated for all frequencies from1
to 90 Hz.
Spectral coherence analysisCoherence analysis was based on
multi-taper Fourier analysis.Coherence was calculated by
custom-written MATLAB scripts,using the MATLAB toolbox Chronux
(http://www.chronux.org)
(Mitra and Pesaran, 1999). Coherence has been calculated foreach
1 s long segments and their mean was evaluated for allfrequencies
from 1 to 90 Hz.
Coherence-PSD correlationsThe correlation values were calculated
between the follow-ing paired frequency-domain vectors:
coherence—mPFC PSD;coherence—vHIP PSD. The correlation measures
were calculatedthrough a custom-written MATLAB program that
computes pair-wise Pearson’s coefficient between frequency-domain
(1 Hz bins)coherence and PSD vector pairs. The p-values were
calculatedthrough a two-tailed t-test.
Statistical group analysisThe group mean comparisons were
calculated through non-parametric Mann–Whitney tests.
Multiple comparisons, for instance to compare LTP inductionin
the three groups (Figure A1), were made using
non-parametricKruskal–Wallis test followed by Dunn’s
corrections.
Performance of each group in the reference memory task
andrespective swim speeds were compared using
repeated-measuresANOVA, being the difference between groups in each
day calcu-lated by Bonferroni post-hoc tests.
Pearson correlation coefficients were calculated
betweenbehavioral parameters and PSD, for both PFC and
hippocampusregions, and vHIP-mPFC coherence.
ACKNOWLEDGMENTSThe authors work was supported by FEDER funds
throughOperational program for competivity factors—COMPETEand by
national funds through FCT—Foundation forScience and Technology
(FCT) fellowships (João FilipeOliveira by SFRH/BPD/66151/2009; Luís
Ricardo Jacinto bySFRH/BD/40459/2007), a Marie Curie Fellowship
(João FilipeOliveira by PIEF-GA-2010-273936) and Grants from
BIALFoundation (138/2008 to João José Cerqueira and 61/2010 toJoão
Filipe Oliveira and Vanessa Morais Sardinha) and FCT (JoãoFilipe
Oliveira, Ana Lima, and Ana Filipa Oliveira by
PTDC/SAU-NSC/118194/2010; Nuno Sérgio Dias, Luís Ricardo
Jacinto,João José Cerqueira, and Nuno Sousa by
FCT/PTDC/SAU-ENB/118383/2010; Nuno Sérgio Dias, Daniela Silva
Ferreira,Joana Santos Reis, João José Cerqueira, and Nuno Sousa
byFCOMP-01-0124-FEDER-022674). The authors would like tothank Rui
Gomes for the help in the application of electrophysio-logical
techniques and Luís Martins and Miguel Carneiro for thehistological
preparations.
REFERENCESAmat, J., Paul, E., Zarza, C., Watkins, L.
R., and Maier, S. F. (2006). Previousexperience with behavioral
controlover stress blocks the behavioraland dorsal raphe nucleus
activatingeffects of later uncontrollable stress:role of the
ventral medial prefrontalcortex. J. Neurosci. 26, 13264–13272.
Benchenane, K., Peyrache, A.,Khamassi, M., Tierney, P.
L.,Gioanni, Y., Battaglia, F. P., et al.
(2010). Coherent theta oscillationsand reorganization of spike
timingin the hippocampal-prefrontalnetwork upon learning. Neuron
66,921–936.
Brockmann, M. D., Pöschel, B.,Cichon, N., and Hanganu-Opatz,I.
L. (2011). Coupled oscillationsmediate directed interactionsbetween
prefrontal cortex andhippocampus of the neonatal rat.Neuron 71,
332–347.
Buzsáki, G. (1996). The Hippocampo-neocortical dialogue. Cereb.
Cortex6, 81–92.
Buzsáki, G. (2006). Rhythms of theBrain. 1st Edn. USA:
OxfordUniversity Press.
Buzsáki, G. (2010). Neural syntax:cell assemblies, synapsembles,
andreaders. Neuron 68, 362–385.
Cerqueira, J. J., Mailliet, F., Almeida,O. F. X., Jay, T. M.,
and Sousa,N. (2007a). The prefrontal cortex
as a key target of the maladaptiveresponse to stress. J.
Neurosci. 27,2781–2787.
Cerqueira, J. J., Taipa, R., Uylings,H. B. M., Almeida, O. F.
X.,and Sousa, N. (2007b). Specificconfiguration of dendritic
degen-eration in pyramidal neuronsof the medial prefrontal
cortexinduced by differing corticosteroidregimens. Cereb. Cortex
17,1998–2006.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 9
http://www.
chronux.orghttp://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
Dazzi, L., Seu, E., Cherchi, G., andBiggio, G. (2004).
Inhibitionof stress-induced dopamineoutput in the rat
prefrontalcortex by chronic treatment witholanzapine. Biol.
Psychiatry 55,477–483.
De Bruin, J. P., Sànchez-Santed, F.,Heinsbroek, R. P., Donker,
A., andPostmes, P. (1994). A behaviouralanalysis of rats with
damage to themedial prefrontal cortex using theMorris water maze:
evidence forbehavioural flexibility, but not forimpaired spatial
navigation. BrainRes. 652, 323–333.
Dias-Ferreira, E., Sousa, J. C., Melo,I., Morgado, P., Mesquita,
A.R., Cerqueira, J. J., et al. (2009).Chronic stress causes
frontostri-atal reorganization and affectsdecision-making. Science
325,621–625.
Feenstra, M. G. (2000). Dopamine andnoradrenaline release in the
pre-frontal cortex in relation to uncon-ditioned and conditioned
stressand reward. Prog. Brain Res. 126,133–163.
Fell, J., and Axmacher, N. (2011).The role of phase
synchronizationin memory processes. Nat. Rev.Neurosci. 12,
105–118.
Finlay, J. M., Zigmond, M. J., andAbercrombie, E. D.
(1995).Increased dopamine and nore-pinephrine release in
medialprefrontal cortex induced byacute and chronic stress:
effectsof diazepam. Neuroscience 64,619–628.
Hasler, G., van der Veen, J. W.,Grillon, C., Drevets, W. C.,
andShen, J. (2010). Effect of acutepsychological stress on
prefrontalGABA concentration determined byproton magnetic resonance
spec-troscopy. Am. J. Psychiatry 167,1226–1231.
Hyman, J. M., Hasselmo, M. E., andSeamans, J. K. (2011). What is
thefunctional relevance of prefrontalcortex entrainment to
hippocampaltheta rhythms? Front. Neurosci. 5:24.doi:
10.3389/fnins.2011.00024
Jay, T. M., and Witter, M. P. (1991).Distribution of hippocampal
CA1and subicular efferents in the pre-frontal cortex of the rat
studied bymeans of anterograde transport ofPhaseolus
vulgaris-leucoagglutinin.J. Comp. Neurol. 313, 574–586.
Jia, X., Smith, M. A., and Kohn,A. (2011). Stimulus selectiv-ity
and spatial coherence ofgamma components of the localfield
potential. J. Neurosci. 31,9390–9403.
Jones, M. W., and Wilson, M. A.(2005). Theta rhythms coordi-nate
hippocampal–prefrontalinteractions in a spatial mem-ory task. PLoS
Biol. 3:e402. doi:10.1371/journal.pbio.0030402
Lee, Y. A., Poirier, P., Otani, S., andGoto, Y. (2011).
Dorsal-ventral dis-tinction of chronic
stress-inducedelectrophysiological alterations inthe rat medial
prefrontal cortex.Neuroscience 183, 108–120.
Liston, C., McEwen, B. S., and Casey,B. J. (2009). Psychosocial
stressreversibly disrupts prefrontal pro-cessing and attentional
control.Proc. Natl. Acad. Sci. U.S.A. 106,912–917.
Maggio, N., and Segal, M. (2009).Differential modulation of
long-term depression by acute stressin the rat dorsal and
ventralhippocampus. J. Neurosci. 29,8633–8638.
Maier, S. F., and Watkins, L. R.(2005). Stressor controllability
andlearned helplessness: the roles ofthe dorsal raphe nucleus,
sero-tonin, and corticotropin-releasingfactor. Neurosci. Biobehav.
Rev. 29,829–841.
Mark, G. P., Rada, P. V., and Shors, T. J.(1996). Inescapable
stress enhancesextracellular acetylcholine in the rathippocampus
and prefrontal cor-tex but not the nucleus accum-bens or amygdala.
Neuroscience 74,767–774.
Mitra, P. P., and Pesaran, B. (1999).Analysis of dynamic
brainimaging data. Biophys. J. 76,691–708.
Moghaddam, B. (2002). Stressactivation of glutamate
neuro-transmission in the prefrontalcortex: implications for
dopamine-associated psychiatric disorders.Biol. Psychiatry 51,
775–787.
Morris, R. (1984). Developments ofa water-maze procedure for
study-ing spatial learning in the rat.J. Neurosci. Methods 11,
47–60.
Otero Losada, M. E. (1988). Changes incentral GABAergic function
follow-ing acute and repeated stress. Br. J.Pharmacol. 93,
483–490.
Paxinos, V. G., and Watson, C. (2005).The Rat Brain in
StereotaxicCoordinates. 5th Edn. New York,NY: Academic Press.
Perea, G., Navarrete, M., and Araque,A. (2009). Tripartite
synapses: astro-cytes process and control synapticinformation.
Trends Neurosci. 32,421–431.
Popa, D., Duvarci, S., Popescu, A.T., Léna, C., and Paré, D.
(2010).Coherent amygdalocortical thetapromotes fear memory
consoli-dation during paradoxical sleep.Proc. Natl. Acad. Sci.
U.S.A. 107,6516–6519.
Popoli, M., Yan, Z., McEwen, B. S., andSanacora, G. (2011). The
stressedsynapse: the impact of stress andglucocorticoids on
glutamate trans-mission. Nat. Rev. Neurosci. 13,22–37.
Radley, J. J., Sisti, H. M., Hao, J.,Rocher, A. B., McCall, T.,
Hof, P.R., et al. (2004). Chronic behav-ioral stress induces apical
dendriticreorganization in pyramidal neu-rons of the medial
prefrontal cortex.Neuroscience 125, 1–6.
Rocher, C., Spedding, M., Munoz,C., and Jay, T. M. (2004).
Acutestress-induced changes in hip-pocampal/prefrontal circuits in
rats:effects of antidepressants. Cereb.Cortex 14, 224–229.
Sehatpour, P., Molholm, S., Schwartz,T. H., Mahoney, J. R.,
Mehta, A.D., Javitt, D. C., et al. (2008). Ahuman intracranial
study of long-range oscillatory coherence acrossa
frontal–occipital–hippocampalbrain network during visual
objectprocessing. Proc. Natl. Acad. Sci.U.S.A. 105, 4399–4404.
Siapas, A. G., Lubenov, E. V., andWilson, M. A. (2005).
Prefrontalphase locking to hippocampal thetaoscillations. Neuron
46, 141–151.
Sigurdsson, T., Stark, K. L.,Karayiorgou, M., Gogos, J. A.,and
Gordon, J. A. (2010). Impairedhippocampal-prefrontal syn-chrony in
a genetic mouse modelof schizophrenia. Nature 464,763–767.
Soares, J. M., Sampaio, A., Ferreira,L. M., Santos, N. C.,
Marques,F., Palha, J. A., et al. (2012).Stress-induced changes in
humandecision-making are reversible.Transl. Psychiatry 2:e131.
doi:10.1038/tp.2012.59
Sousa, N., and Almeida, O. F. X.(2012). Disconnection and
recon-nection: the morphological basis of(mal)adaptation to stress.
TrendsNeurosci. 35, 742–751.
Sousa, N., Lukoyanov, N. V., Madeira,M. D., Almeida, O. F.,
andPaula-Barbosa, M. M. (2000).Reorganization of the morphol-ogy of
hippocampal neurites andsynapses after stress-induced dam-age
correlates with behavioralimprovement. Neuroscience 97,253–266.
Thierry, A. M., Gioanni, Y., Dégénétais,E., and Glowinski, J.
(2000).Hippocampo-prefrontal cortexpathway: anatomical and
elec-trophysiological characteristics.Hippocampus 10, 411–419.
Tierney, P. L., Dégenètais, E., Thierry,A.-M., Glowinski, J.,
and Gioanni, Y.(2004). Influence of the hippocam-pus on
interneurons of the rat pre-frontal cortex. Eur. J. Neurosci.
20,514–524.
Varela, F., Lachaux, J.-P., Rodriguez,E., and Martinerie, J.
(2001). Thebrainweb: phase synchronizationand large-scale
integration. Nat.Rev. Neurosci. 2, 229–239.
Conflict of Interest Statement: Theauthors declare that the
researchwas conducted in the absence of anycommercial or financial
relationshipsthat could be construed as a potentialconflict of
interest.
Received: 07 November 2012; accepted:16 January 2013; published
online: 06February 2013.Citation: Oliveira JF, Dias NS, CorreiaM,
Gama-Pereira F, Sardinha VM, LimaA, Oliveira AF, Jacinto LR,
Ferreira DS,Silva AM, Reis JS, Cerqueira JJ andSousa N (2013)
Chronic stress disruptsneural coherence between
cortico-limbicstructures. Front. Neural Circuits 7:10.doi:
10.3389/fncir.2013.00010Copyright © 2013 Oliveira, Dias,Correia,
Gama-Pereira, Sardinha, Lima,Oliveira, Jacinto, Ferreira, Silva,
Reis,Cerqueira and Sousa. This is an open-access article
distributed under the termsof the Creative Commons
AttributionLicense, which permits use, distributionand reproduction
in other forums, pro-vided the original authors and sourceare
credited and subject to any copy-right notices concerning any
third-partygraphics etc.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 10
http://dx.doi.org/10.3389/fncir.2013.00010http://dx.doi.org/10.3389/fncir.2013.00010http://dx.doi.org/10.3389/fncir.2013.00010http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
APPENDIX
FIGURE A1 | Chronic unpredictable stress decreases
synapticplasticity between the ventral hippocampus (vHIP) and the
medialprefrontal cortex (mPFC). (A) Time course of LTP induction
after highfrequency stimulation (HFS); each circle represents the
average of 10normalized fEPSP slopes with SEM; three groups are
depicted: controls
(CON, blue), short term stress (STS, green) and chronic
unpredictablestress (CUS, red); (inset) representative recordings
of fEPSPs pre- andpost-HFS (scale bar: 1 mV; 5 ms) of a CON rat;
(B) Normalized values ofLTP averaged for the three groups of rats:
CON, STS, and CUS;∗statistically different from controls, p <
0.05; error bars represent SD.
FIGURE A2 | Chronic stress exposure does not induce
locomotordeficits. Average swimming speeds of CON (blue) and CUS
(red) ratsduring the 4 days spatial reference memory task. No
differences areobserved between groups in each day (p > 0.05);
error barsrepresent SD.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 11
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
-
Oliveira et al. Chronic stress disrupts cortico-limbic
coherence
FIGURE A3 | Stressed rats show increased PSD in
multiplefrequency bands both in the ventral hippocampus andmedial
PFC in the second set of rats. Group comparison ofthe power
spectral density (PSD) values from vHIP (A) and
mPFC (B) in the delta (1–4 Hz), theta (4–12 Hz), beta (12–20
Hz),low gamma (20–40 Hz) and high gamma (40–90 Hz) frequencybands.
∗Statistically different from CON, p < 0.05; error barsrepresent
SD.
FIGURE A4 | Identification of electrode recording sites. Left
panel,recording site location in control (CON, blue), short term
stress (STS, green)and chronic unpredictable stress (CUS, red)
rats, within the medial prefrontal
cortex (A, mPFC) and ventral hippocampus (B, vHIP); Images
adapted fromPaxinos and Watson, 2005. Right panel, Cresyl Violet
stained 50 µm sectionswith an electrolytic lesion at the recording
site.
Frontiers in Neural Circuits www.frontiersin.org February 2013 |
Volume 7 | Article 10 | 12
http://www.frontiersin.org/Neural_Circuitshttp://www.frontiersin.orghttp://www.frontiersin.org/Neural_Circuits/archive
Chronic stress disrupts neural coherence between cortico-limbic
structuresIntroductionResultsStressed Rats Display Increased Power
Spectrum Densities in the vHIP and mPFCStress Decreases Spectral
Coherence Between the vHIP and mPFCStress-Induced Loss of Coherence
Correlates with Increased PSD in the vHIPStress Exposure Impaired
LTP Induction in the Hippocampal-mPFC LinkBehavioral Performance
Correlates with Variations in Coherence and Power Spectrum
Density
DiscussionMaterials and MethodsAnimals and TreatmentsSurgeryIn
vivo Recording of Local Field Potentials in the mPFC and Ventral
HippocampusIn vivo Study of Synaptic Plasticity between the mPFC
and Ventral HippocampusBehavioral AnalysisData Analysis and
StatisticsPSD analysisSpectral coherence analysisCoherence-PSD
correlationsStatistical group analysis
AcknowledgmentsReferencesAppendix