Neurobiology of Disease xxx (2013) xxx–xxx

YNBDI-02897; No. of pages: 10; 4C: 3, 4, 6, 7

Contents lists available at SciVerse ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r .com/ locate /ynbd i

Partial loss in septo-hippocampal cholinergic neurons alters memory-dependentmeasures of brain connectivity without overt memory deficits

Laurent Brayda-Bruno a,b, Nicole Mons c, Benjamin K. Yee d, Jacques Micheau c, Djoher Nora Abrous a,b,Xavier Nogues a,b, Aline Marighetto a,b,⁎a INSERM, Neurocentre Magendie, Physiopathologie de la Plasticité Neuronale, U862, F-33000 Bordeaux, Franceb Univ. Bordeaux, Neurocentre Magendie, Physiopathologie de la Plasticité Neuronale, U862, F-33000 Bordeaux, Francec CNRS UMR 5287, Avenue des Facultés 33405 Talence, Université Bordeaux, Franced Robert S. Dow Neurobiology Laboratories, Legacy Research Institute, 1225 NE 2nd Ave, Portland, OR 97232, USA

⁎ Corresponding author at: INSERM, Neurocentre MaPlasticité Neuronale, U862, F-33000 Bordeaux, France.

E-mail address: aline.marighetto@inserm.fr (A. MariAvailable online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2013 Elsevier Inc. Allhttp://dx.doi.org/10.1016/j.nbd.2013.01.010

Please cite this article as: Brayda-Bruno, L., ebrain connectivity without overt memory ...,

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 August 2012Revised 7 January 2013Accepted 10 January 2013Available online xxxx

Keywords:Declarative memoryAgingP75 saporin immunotoxinRadial arm mazeFos immunohistochemistryFunctional connectivitySpatial memoryMouse

The functional relevance of septo-hippocampal cholinergic (SHC) degeneration to the degradation of hippocampus-dependent declarative memory (DM) in aging and Alzheimer's disease (AD) remains ill-defined. Specifically, selec-tive SHC lesions often fail to induce overt memory impairments in animal models. In spite of apparent normal per-formance, however, neuronal activity within relevant brain structures might be altered by SHC disruption. Wehypothesized that partial SHC degenerationmay contribute to functional alterations withinmemory circuits occur-ring in aging before DM decline. In young adult mice, we studied the effects of behaviorally ineffective (saporin-induced) SHC lesions – similar in extent to that seen in aged animals – on activity patterns and functional connec-tivity between three main neural memory systems: the septo-hippocampal system, the striatum and the amygdalathat sustain declarative, procedural and emotionalmemory, respectively. Animalswere trained in a radialmaze pro-cedure dissociating the human equivalents of relational/DM and non-R/DM expressions in animals. Test-inducedFos activation pattern revealed that the partial SHC lesion significantly altered the brain's functional activities andconnectivity (co-activation pattern) despite the absence of overt behavioral deficit. Specifically, hippocampal CA3hyperactivity and abnormal septo-hippocampo-amygdalar inter-connectivity resemble those observed in agingand prodromal AD. Hence, SHC neurons critically coordinate hippocampal function in concert with extra-hippocampal structures in accordance with specific mnemonic demand. Although partial SHC degeneration is notsufficient to impact DM performance by itself, the connectivity change might predispose the emergence of subse-quent DM loss when, due to additional age-related insults, the brain can no longer compensate the holistic imbal-ance caused by cholinergic loss.

© 2013 Elsevier Inc. All rights reserved.

Introduction

Cholinesterase inhibitors are commonly administered to Alzheimer'sdisease (AD) patients to compensate for the loss of cholinergic neuronsin an attempt to rescue the loss of memory function (Bartus et al.,1985; Pepeu andGiovannini, 2009). However, the contribution of cholin-ergic degeneration to the characteristic deterioration of hippocampus-dependent declarativememory (DM) in aging and early AD is still poorlyunderstood. Indeed, the reduction in cholinergic markers seen in agedpatients and animals often correlates with the severity of their memoryimpairment (Gallagher and Rapp, 1997) but selective deafferentationof cholinergic input to the hippocampus by intraseptal injection of

gendie, Physiopathologie de la

ghetto).irect.com).

rights reserved.

t al., Partial loss in septo-hippNeurobiol. Dis. (2013), http:/

immunotoxin saporin in young animals does not reliably produce overtmemory loss (Parent and Baxter, 2004).

To better understand the role of septo-hippocampal cholinergic(SHC) neurons in senescent memory decline, we examined herewhether partial SHC degeneration similar in extent to that naturally oc-curring in aging – despite being inefficient to induce over memory im-pairment – could alter brain activity patterns. Indeed, senescence isassociated with alterations in functional activity patterns and connec-tivity within memory-related brain circuits and some of these alter-ations occur before the appearance of overt DM degradation (Grady,2008; Hedden and Gabrieli, 2005; Minati et al., 2007; Pihlajamaki etal., 2009; Sperling et al., 2010).We hypothesize that partial SHC degen-eration may contribute to the functional alterations that precede irre-versible DM decline. In particular, while hippocampal activationengaged by memory execution is typically reduced in cognitively im-paired subjects, a paradoxical increase in hippocampal activity at pro-dromal stage of AD has been observed, which may be predictive offuture pathological memory decline (Bassett et al., 2006; Dickerson

ocampal cholinergic neurons alters memory-dependent measures of/dx.doi.org/10.1016/j.nbd.2013.01.010

2 L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

and Sperling, 2008;Mondadori et al., 2006; Quiroz et al., 2010; Sperling,2007). Moreover, indiscriminate hippocampal hyperactivation underDM as well as non-DM learning and striatal hyperactivation under DMlearning have been previously observed in aged subjects, suggesting asenescence-related “de-differentiation” between DM and non-DMcircuits (Dennis and Cabeza, 2011). Finally, aging-related reduction inhippocampo-amygdalar functional connectivity linked to memory re-trieval has been reported in two studies (Murty et al., 2009; St Jacqueset al., 2010).

Although the precisemechanisms leading to functional connectivitydysregulation remain to bedelineated, the hypothesized contribution ofSHC degeneration is based on current knowledge that SHC manipula-tions can alter hippocampal functional activity and the coordinationamong neural systems underlying different memory processes. First,computational models (Buzsaki, 1989; Hasselmo, 2006; Hasselmo andBower, 1993) predict that the modulation of hippocampal function byacetylcholine modulates the balance between memory encoding andretrieval processes subserved by the hippocampus. Thus, SHC disrup-tion may induce a shift in hippocampal computation, detrimental tomemory encoding/promoting retrieval processes similar to that seenin aging (Gallagher et al., 2010; Ikonen et al., 2002; Toner et al., 2009;Yassa and Stark, 2011). Second, hippocampal cholinergic transmissionseems to coordinate activity between neural systems sustaining DMand non-DM (Micheau andMarighetto, 2011) by balancing the relativecontributions of the hippocampus and extra-hippocampal structures,such as the amygdala (Calandreau et al., 2006; McIntyre et al., 2002)and the striatum (Chang and Gold, 2003; McIntyre et al., 2003; Pychet al., 2005) involved in emotional and procedural memory, respective-ly (White and McDonald, 2002).

Here, we tested the hypothesis that senescence-related SHC degener-ation could alter functional activity in the hippocampus and functional

A B

Fig. 1. Impact of aging and intra-septal saporin infusion on the density of SHC neurons. In thereduction in ChAT immunopositive cells' density is observed in 23–24 month old mice compasame age (4–5 months, C). B: photomicrographs of the medial septum representative of the S

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:

connectivity between the septo-hippocampal, striatal and amygdalarneural systems. To this end, memory-induced activation patterns inthese brain areas were visualized by Fos imaging in young adult mice,and the impact of partial SHC neuronal loss (induced by intra-septalimmunotoxin saporin) comparable inmagnitude to that occurring natu-rally in aged animals examined. Thepartial lesionswere expected to alterinter-region Fos activity patterns without affecting memory perfor-mance. A radial maze discrimination learning procedure was used toseparately engage two forms ofmemory expression, thereby dissociatingrelational/DM and non-R/DM, through a change in the arrangement ofdiscriminanda with any difference in informativeness (Mingaud et al.,2007). Fos activation patterns were examined in a two-way (lesions×memory engagement) factorial design contrasting lesioned and controlmice having been subjected to either R/DM or non-R/DM retrieval incomparison with a treadmill non-learning control condition.

Methods and materials

Subjects

Male C57 BL/6J mice (Charles River, Lyon, France) were used.First we evaluated the effect of aging on the density of SHC neu-

rons by comparing the number of ChAT immunopositive neuronsper mm2 in the medial septum – the main source of the cholinergicinnervations to the hippocampus (Mesulam et al., 1983) – in young(4–5 months, n=9) and aged (23–24 months, n=11) mice.

Having ascertained that agedmice suffered approximately a 20% re-duction in the density of SHC cells, we went on to induce SHC loss ofsimilar magnitude in young mice by intra-septal saporin infusion(saporin-injected mice, n=27; vehicle-injected Sham controls, n=29). Saporin and Sham mice were submitted to behavioral training

C

medial septum (M-Sept) but not in the vertical limb of the diagonal band (VDB), a ≈20%red to 4-month old ones (A), and in saporin-treated mice compared to Shammice of theham and Saporin groups (scale bar: 50 μm).

ocampal cholinergic neurons alters memory-dependent measures of//dx.doi.org/10.1016/j.nbd.2013.01.010

3L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

either in radial maze discrimination learning tasks (assessing R/DM ornon-R/DM) or (“non-learning/active” control) treadmill training condi-tion. All mice were maintained at around 90% of their free feedingweight throughout training by appropriate food restriction in thehome cage. The final number of mice by group was: Sham-R/DM n=12, Sham-non R/DM n=8, Sham-Treadmill n=6, Saporin-R/DM n=7, Saporin-non R/DM n=9 and Saporin-Treadmill n=7.

All manipulations were carried out in accordance with theEuropean Communities Council Directive of the 24 November 1986(86/609/EEC).

Surgery

Ten days before behavioral training, partial SHC lesionsweremade tomimic the subtotal loss of cholinergic neurons in the medial septumfound in aged mice of the same strain (−17.6% vs young adults, seeFig. 1). Mice were deeply anesthetized with a ketamine–xylazine mix-ture and maintained under standard surgical conditions. Mu p75-Sap(Advanced Targeting Systems, San Diego, CA, USA) diluted in PBS(0.045 μg/0.3 μl) was injected into the medial septal area by stereotaxicguidance (midline, 1 mm posterior to the bregma, 4.5 mm below dura).The concentration and volume of toxin injected here were determinedby prior pilots (consisting of counting ChAT immunopositive cells inthe medial septum after local injection of various doses of mu p75-Sap)since all previous studies with the same toxin had chosen an intra-cerebroventricular route of administration (e.g., Ho et al., 2009;Moreau et al., 2008). In our experimental conditions, our chosen dosewas sufficiently low to limit damage to a subset of cholinergic cells andwas therefore unlikely to significantly affect non-cholinergic cells.

Fig. 2. Schematic representation of the radial-maze procedure. R/DM and non-R/DM groups wepositive, three negative) arms were repeatedly presented one by one successively, acco15 s-inter-trial-interval). After reaching the acquisition criterion (when the mean latency tomice were submitted on the following day to a retrieval test of either R/DM or non-R/DM. Tthe arms was modified to assess either R/DM (the arms were presented by pairs, with one pwere opened all six at a time, and the mouse was free to visit them in whichever order, until a

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:/

Behavior

ApparatusAn open 8-arm radial maze automatized by IMETRONIC (Pessac —

France) and described in Mingaud et al.(2007, 2008) was used for thelearning conditions. For the non-learning control condition, the animalswere placed on an electrical treadmill as previously described (Mingaudet al., 2007, 2008; Touzani et al., 2003). The treadmill (28 cm long×5 cmwide×20 cm high) was located in the same room that housed theradial-maze. The treadmill mice were maintained in the same conditions(including partial food deprivation) and submitted to the same amountand duration of training-sessions as the radial-maze learning groups.The running speed was increased progressively across sessions to matchthe general increase in running speed observed in mice subjected toradial-maze learning task. This was achieved by yoking each treadmillmouse randomly to a radial-maze mouse.

Radial maze test designThe procedures were chosen for its capability to dissociate two

forms of memory expression simply by manipulating the number ofarms presented at a given choice point, which can determine if per-formance is supported by the hippocampus or not (Etchamendy etal., 2003; Mingaud et al., 2007).

Each mouse was separately assigned six adjacent arms with threedesignated as positive (baited) arms and the remaining three negative(unbaited) arms. The reward valence of the arms remained unchangedthroughout the entire experiment, but the manner of presenting thearms was modified between the acquisition (Stage 1) and the retrievaltest (Stage 2) in order to solicit distinct expressions (hippocampal vsnon-hippocampal) of acquired memory about arm-valence (Fig. 2). In

re first submitted to the same acquisition phase (5–10 days), during which the six (threerding to a go/no-go discrimination learning procedure (24 trials/day, 4 trials/arm,enter a negative arm was at least 30% longer than the one to enter a positive arm), thehe location of food within the arms remained unchanged, but the manner of presentingositive one negative arm, and the mouse must make a choice), or non-R/DM (the armsll three positive arms have been visited).

ocampal cholinergic neurons alters memory-dependent measures of/dx.doi.org/10.1016/j.nbd.2013.01.010

Fig. 3. Area delimitation for Fos counting. The number of Fos immunopositive cells by mm2 was measured in the hippocampal subfields (CA1, CA3, DG: 4 slices, coordinates from −1.46 mm to −2.06 mm relative to bregma, Paxinos and Franklin, 2001), septum (medial: M-SEPT; lateral: L-SEPT; 3–4 slices, from +1.18 to +0.74/bregma), amygdala (lateral: LA;basolateral: BLA; 4 slices from −0.8 to −1.46/bregma) and dorsal striatum (lateral: L-STR; medial: M-STR; 4 slices from +0.98 to +0.38/bregma). Photomicrograph illustrations ofFos staining in the dorso-medial striatum, lateral amygdala and CA3 (scale bar: 50 μm).

4 L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

Stage 1, the six armswere successively presented (one at a time, “go/nogo” task) until reaching the acquisition criterion, when median latencyof entry into negative arms was at least 30% longer than that into posi-tive arms over two consecutive sessions. In Stage 2, depending on thegroups, the arms were either (i) presented in pairs of two (of oppositereward valence) or (ii) altogether (all six) at the same time. Thepair-wise arm presentation has been shown to favor the expression ofR/DM while presenting all six arms at once would favor non-R/DMexpression. [See details and explanation in previous experiments(Etchamendy et al., 2003; Mingaud et al., 2007; Touzani et al., 2003)which have shown that mice with hippocampal dysfunction onlyexhibited deficit in the pair discrimination (R/DM) test while go/no-go discrimination or the six-choice performance remained intact.]Animals were subjected to one memory retrieval test session using ei-ther the “R/DM” (two-choice) or the “non R/DM” (six-arm) procedures.The Stage 2 session was limited to 20 min. Performance was measuredby choice accuracy expressed as mean percentage of correct responses(choices of a baited arm) across all trials in the R/DM test (maximum20 trials) or over first three visits on each trial (maximum 8 trials) inthe non-R/DM test (see Mingaud et al., 2007 for more details).

Immunohistochemistry

The trained mice were killed 90 min after the beginning of the Stage2-testing session (70–75 min after the end of the task as the sessionwas limited to 20 min duration). Animals were rapidly killed by a lethali.p. injection of Avertin (0.5 mg/kg of tribromoethanol stabilized in tertia-ry amyl alcohol), followed by intracardiac perfusion with 4% paraformal-dehyde (100 ml/mouse, 30 ml/min) in a 0.1 M phosphate buffer (PB),pH 7.4. The brains were removed, and post-fixed overnight before beingsectioned on a vibratome into three interleaved 50-μm thick coronal sec-tions. Hence, consecutive slices within a given series were 100 μm apart.

Sections were stocked in a cryoprotective solution (ethylene glycol/glycerol/PB) until immunohistochemical processing for the visualizationof Fos or ChAT. Free-floating sections were rinsed in PBS, and then incu-bated in PBS with 0.5% H2O2 for 30 min, followed by incubation in ablocking solution containing 1% bovine serum albumin (BSA), 3% goatserum (GS), and 0.2% Triton X-100 in PBS. Sections were incubated

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:

with the appropriate primary antibody (see below) in a PBS–BSA–GS–Triton X-100 solution for 24 h at 4 °C. After extensive washes in PBS,sections were incubated with the biotinylated secondary antibody in aPBS–BSA–GS–Triton X-100 solution for 2 h at RT. Sections were then in-cubated in avidin–biotinylated horseradish peroxidase complex (ABC,Vectastain Elite kit, Vector Laboratories) for 2 h at RT. Next, the sectionswere rinsed, and the peroxidase reaction end product was visualized in0.05 M PB containing 0.025% 3,3′-diaminobenzidine tetrahydrochloride(DAB, Sigma) and 0.03% H2O2. Finally, the sections were mounted ongelatin-coated glass slides, dehydrated through a graded series of etha-nol, cleared in toluene, and coverslipped with Eukitt mountingmedium.The number of (Fos or ChAT) immunopositive cells in defined regions ofinterest was counted bilaterally using three or four (100 μmdistant) sec-tions within a predetermined area (the same for all brains, coordinatesfrom (Paxinos and Franklin, 2001) as represented in Fig. 3. The numberof positive cells was quantified using an Olympus BX50 microscopeequipped with a computerized imaging analysis system (Visiolab 2000version 4.50; Biocom, Paris, France). The threshold was establishedonce for all brains using “home cage” naïve controls (with nearly noFos positive cells) and “treadmill” active controls. The experimenterwas blind to the experimental condition. The results were expressed asnumber of immunopositive cells by mm2.

ChAT immunohistochemistryThe antibodies used were: anti-ChAT goat polyclonal antibody

(1:1000; Chemicon International) and biotinylated donkey anti-goatIgG (1:2000; Jackson Immunoresearch). For each animal, the numberof ChAT-immunoreactive neurons by mm2 was counted in the medialseptum and bilaterally in the vertical limb of the diagonal band.

Fos immunohistochemistryThe antibodies used were: anti-c-Fos rabbit polyclonal antibody

(1:20,000; Santa Cruz Biotechnology) and biotinylated goat anti-rabbitIgG (1:2000; Jackson Immunoresearch). Fos immunopositive cells (bymm2) were counted in the following structures: CA1, CA3, dentategyrus (DG) of the dorsal hippocampus, the medial and lateral septum;themedial and lateral parts of dorsal striatum; the lateral and basolateralnucleus of the amygdala (cf Fig. 3).

ocampal cholinergic neurons alters memory-dependent measures of//dx.doi.org/10.1016/j.nbd.2013.01.010

A B

Fig. 4. No behavioral impact of the saporin lesion in the radial-maze task (upper part: taskacquisition, lower part: memory retrieval tests).

5L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

Statistical analyses

Analyses of varianceBehavioral and immunohistochemical (Fos and ChAT) data were

submitted to one-way analysis of variance (ANOVA) with either thebetween-subject factor: Age or Lesion, or two-way ANOVA with thefactors: Lesion and Behavior (Treadmill, non-R/DM, R/DM). Post-hoccomparisons were performed using Tukey/Kramer test.

Functional connectivityFos data were submitted to numerical analyses designed to quan-

tify the pattern of inter-regional functional connectivity, which thenallowed us to examine if this might significantly differ between ex-perimental groups. Namely, the pattern of a given group was repre-sented as a whole by the correlative matrix comprising all possiblepairwise correlations between any two structures. In order to comparecorrelationmatrices between groups, we computed two indices: (i) themean of the absolute value of the difference between the lower triangu-lar parts of the two matrices (Md, see Eq. (1)) and (ii) the Euclidean

A

C

Fig. 5. Partial SHC lesion alters the task-specific recruitment of functional activity among the thCA3, DG subfields of dorsal hippocampus; B: the medial and lateral septum), lateral and bmice with prior active engagement of relational/declarative memory (R/DM), only non-Rradial-maze tasks which only differed in the manner of presenting the arms, i.e. by pairsthan the form of memory concerned. The levels of performance were similar between thebetween Lesion conditions; *: pb0.05 between Behavior conditions (Treadmill, non-R/DM, R

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:/

distance between the lower triangular parts of the two matrices (ED,see Eq. (2)). They were computed as follows:

Md ¼2Xkj¼1

Xki¼ jþ1

Aij−Bij

� �

k k−1ð Þ 1

ED ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXkj¼1

Xki¼ jþ1

Aij−Bij

� �2

vuut 2

A and B refer to the two correlationmatrices being compared, and k is thenumber of columns or rows in the matrix. For Md, the numerator ex-presses the sum of the differences (element by element) between thelower triangular part of the two correlationmatrices, and the denomina-tor expresses the number of elements in the lower triangular part of thetwo matrices. The complementary nature of the quadratic (ED) andnon-quadratic (Md) indices has been discussed earlier (Boucard et al.,2009), and a significant difference in either index would be sufficientto illustrate a notable difference in terms of functional connectivity.

Like covariance, Md and ED have no metric. Therefore, statisticaltests are needed to interpret them and to draw conclusions about aputative effect of an experimental factor. Given that the distributionof these two parameters is unknown, a bootstrap procedure wasused to test for statistical significance (Boucard et al., 2009; Efronand Tibshirani, 1991; Horwitz et al., 1987). This was implementedby a script in Octave 2.9.13 (Eaton, 2002). p-Values were determinedafter 10,000 re-samplings (with replacement) based on the specifiedsample sizes of the two groups under comparison. Within each lesioncondition (Sham or Saporin), three separate pairwise comparisonswere made across the three behavioral conditions (Treadmill vsnon-R/DM vs R/DM). Within each behavioral condition, only the com-parison between Sham and Saporin was necessary.

B

D

ree main neural memory systems. Fos levels in the septo-hippocampal system (A: CA1,asolateral amygdala (C) and lateral and medial dorsal striatum (D) were measured in/DM experience or non-mnemonic training (Treadmill). The two memory tests wereor by six, to assess R/DM or non R/DM in conditions identical in every other aspectLesion (Sham vs Saporin) conditions and the two tests (R/DM, non-R/DM). °°: pb0.01/DM).

ocampal cholinergic neurons alters memory-dependent measures of/dx.doi.org/10.1016/j.nbd.2013.01.010

6 L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

This method allowed a global comparison of the two correspondingcorrelation matrices. Only when this yielded a significant differencewouldwe performpair-wise comparisons between individual elementsof the two matrices using the standard procedure described previously(Zar, 1999). This approach thus provided sufficient guard against Type Ierrors and avoided the excessive compromise in statistical power due tocorrection for unplanned multiple explorative comparisons.

Results

ChAT immunohistochemistry

Analyses in young and aged brains (Fig. 1A) showed that aging wasassociated with a mild decrease in the number of cholinergic neuronsby mm2 in the medial septum (−17.5%, SEM: +/−4.5; One-wayANOVA, Age effect: F1,19=5.87; p=0.026) but not in the vertical limbof the diagonal band (−5.4%, SEM: +/−3.3, Fb1).

Fig. 6. Partial SHC lesion alters the task-specific functional connectivity among memory systemsment and behavior condition (Sham-Treadmill, Sham non-R/DM, Sham R/DM, Saporin Treadmareas is represented by one line around the circle. The sign of each correlation is represented bycorrelation. The strength of each correlation is represented by its line proportional length andstatistical comparison of correlation matrices by the method developed by Boucard et al. (bootmatrix of the Sham R/DM group was statistically different from the one in the Sham non-R/DMtween matrices was statistically significant (all p values>0.11).

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:

Analyses in Sham and Saporin groups of young mice (Figs. 2B–C)showed that the saporin injection in the medial septum had induced adecrease in the density of SHC neurons, which was comparable in mag-nitudewith the loss observed in agedmice of the same strain (−19.75%,SEM: +/−5.1; Lesion effect: F1,34=4.99; p=0.032 in medial septum.+3.1%, SEM: +/−6, Fb1 in vertical limb of the diagonal band).

Behavior

Analyses of performance in the radial armmaze test showed that theSHC lesions did not produce any overt deficit in both stages of the ex-periment (Fig. 4). Sham and Saporin groups acquired the “go/no-go”task (Stage 1) with comparable amount of training (Lesion effect:F1,34b1). Similarly, memory retrieval performance (Stage 2) was highlycomparable between groups, whichever thememory retrieval test used(Lesion effect: F1,15=1.05; p=0.32 in the R/DM test; F1,17b1 in the nonR/DM test).

. The figure is a schematic representation of the correlationmatrix for each group of treat-ill, Saporin non-R/DM, Saporin R/DM). In each group, each correlation between two brainthe direction and color of its line: blue, diverging for positive, red, converging for negativethe statistical significance by its line width: thick when pb0.05, thin when p≥0.05. Thestrap test, cf Methods and materials) confirmed the visual impression that the correlationgroup and from the one in the Saporin R/DM group (*: pb0.05). No other difference be-

ocampal cholinergic neurons alters memory-dependent measures of//dx.doi.org/10.1016/j.nbd.2013.01.010

7L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

Thus, partial loss in SHC cells of the same amplitude as that occur-ring in aging did not overtly affect R/DM and non-R/DM performance.

Fos immunohistochemistry

Fos activation induced by the two memory retrieval testsTo analyse the functional activation induced by R/DM and non-R/

DM testing within the three neural memory systems centered on thesepto-hippocampal system, amygdala and striatum, we quantifiedthe protein product of the immediate early gene c-fos in theseareas (Fig. 3 for illustration of area delimitation). We found that: i)the septo-hippocampal and amygdalar systems were differentiallyactivated by the R/DM and non-R/DM retrieval tests in normal condi-tions when compared to non-learning condition (treadmill), and ii)this memory-dependent recruitment of brain activity was alteredin saporin-treated mice (Fig. 5).

i) In Shamanimals (Figs. 5A–D, left parts), the CA1 andDG sub-regionsof the hippocampus, and in addition the lateral amygdala nucleus,exhibited specific R/DM-dependent Fos activation. One-wayANOVAs on Sham Fos data yielded a significant effect of Behavior(R/DM, non-R/DM, Treadmill) in CA1 (F2,23=4.89, p=0.017), DG(F2,23=6.66, p=0.005) and lateral amygdala (F2,23=8.71, p=0.0015), with the R/DM group showing a stronger Fos activationthan non-R/DM and Treadmill groups. The same tendency was alsovisible in the medial septum. Thus, the R/DM retrieval test induceda specific activation of brain area critical for R/DM (the hippocam-pus), and emotional memory (the amygdala) but not of those un-derlying procedural memory (the striatum).

ii) In contrast, Fos expression in the hippocampus and amygdala in thesaporin-treated animals did not differ between the two memorytests (Figs. 5A–D, right parts). In the hippocampus, the lesions ledto abnormal CA3 activation regardless of the memory test condi-tions (two-way ANOVA, Behavior×Lesion interaction: F2,43=3.14,

Fig. 7. Partial SHC lesion alters septo-hippocampo-amygdalar connectivity characteristic of R/Dwithin the matrix of each group (positive in red, negative in orange), as well the correlations sand in the Saporin R/DM matrix (gray dotted lines). Abbreviations. CA1, CA3, DG: subfields ofbasolateral amygdala; LStr: dorsolateral striatum; MStr: dorsomedial striatum.

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:/

p=0.05; Behavior effect in Saporin groups only: F2,19=4.81, p=0.02). The lesions also tended to reduce the difference in Fos activitybetween the R/DM and non-R/DM groups (non-significantly) inCA1, DG, and medial/lateral septum, by augmenting Fos levels inthe non-R/DM test. Hence, if anything, the partial SHC lesionstended to increase activity associated with memory expression inthe septum and hippocampus. In the amygdala and striatum,which are not central structures in R/DM, the SHC lesions unexpect-edly affected Fos activity induced in the R/DM rather than non-R/DM group. Within the R/DM condition, the partial SHC lesionsuppressed Fos activation in lateral amygdala but potentiated it inthe dorso-medial striatum. These observations were supported bythe emergence of a significant Behavior×Lesion interaction in lateralamygdala (F2,43=4.5, p=0.02), and nearly so in the dorso-medialstriatum (F2,43=2.8, p=0.06).

Taken together, these results show that the partial removal of SHCneurons modified R/DM-related activity patterns not only in structuresdirectly targeted by these neurons (viz., hippocampus), but also instructures not directly connected (e.g., amygdala and striatum). Thus,a small SHC lesion can disrupt functional activity among the threemain neural memory systems, especially under R/DM testing.

Functional connectivity among neural systems in the twomemory retrievaltests

Next, we analyzed the functional connectivity among neural sys-tems using a statistical procedure specifically developed by our groupbased on bootstrapping analysis of Fos inter-structure correlation ma-trices (see Methods and materials). The six correlation matrices corre-sponding to each group (Sham, Saporin) and condition (Treadmill,non-R/DM, R/DM) are schematized as radial plots (Fig. 6) to facilitate vi-sual comparison. It is apparent that therewas an increase in the numberof significant positive correlations (blue lines) in the Saporin groups,whereas significant negative correlations (red lines) were found only

M expression. The figure recapitulates statistically significant inter-structure correlationsignificantly different (pb0.05) from the Sham R/DM group in the Sham non-R/DMmatrixdorsal hippocampus; MS: medial septum; LS: lateral septum; LA: lateral amygdala; BLA:

ocampal cholinergic neurons alters memory-dependent measures of/dx.doi.org/10.1016/j.nbd.2013.01.010

8 L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

in Sham animals engaged in a memory retrieval test, especially whentrained under R/DM condition.While the exclusive presence of positivecorrelationsmay be associatedwith artificialmulticollinearity, the addi-tional presence of negative correlations is supportive of a “real” func-tional organization of the structures.

Hence, our observations provide the first suggestion that the func-tional connectivity triggered by memory retrieval amongst the septo-hippocampal, amygdalar and striatal systems differed between R/DMand non-R/DM expressions, and this coordination of multiple neuralsystems involved in memory retrieval was altered by the lesions.Bootstrapping analyses confirmed the visual impression that the R/DMmatrix in the Sham animals was statistically distinct from thenon-R/DM matrix (p=0.025 for Md, p=0.016 for ED) and tentativelyso from the Treadmill one (p=0.071 for Md, p=0.088 for ED). In con-trast, there was little evidence for differences between the behavioraltests among the saporin-treated groups (all p>0.11 for Md and ED).Comparisons between Saporin and Sham groups of the same behavioralcondition revealed a difference only in the R/DM condition (pb0.05 forMd, p=0.057 for ED) but not in Treadmill and non-R/DM conditions(all p>0.2 for Md and ED).

In summary, our analyses lead to two conclusions about functionalconnectivity between brain structures. First, functional connectivitytriggered by memory retrieval differs between the R/DM and thenon-R/DM test conditions in normal subjects, thus further dissociat-ing the neural bases of the two tests at the system level. Second,SHC neurons contribute to the maintenance of the functional connec-tivity pattern seen in the R/DM, which was essentially lost when only~20% of SHC neurons were destroyed.

Functional connectivity disrupted by the SHC lesionTo identify the distinctive structure–structure connectivity trig-

gered by the R/DM retrieval test that had beenmodified significantlyby the lesions, we conducted planned pair-wise comparisons con-trasting (all structure–structure pairs) Sham-R/DM against eitherSham-non-R/DM or Saporin R/DM mice. Fig. 7 illustrates the sta-tistically significant structure–structure correlations within eachgroup (solid lines), and those that distinguished Sham-non-R/DMor Saporin R/DM group from Sham-R/DM group (in gray dottedlines). Seven significant divergences between R/DM and non-R/DMconditions in Sham mice were revealed, and six of them were linkedto correlation between one septo-hippocampal structure and oneamygdalar nucleus. Among these R/DM-specific functional connec-tions, the three between the septum and the amygdala were also sig-nificantly modified by the lesions as captured by the comparisonbetween Saporin and Sham mice in the R/DM condition.

In summary, R/DM testing is uniquely characterized by strong func-tional connections between the septo-hippocampal system and theamygdala that were not engaged by non-R/DM testing. This functionalconnectivity between the septo-hippocampal system and amygdalaunder R/DM testing was disrupted by the partial destruction of SHCneurons.

Discussion

In this novel system-level functional analysis of cholinergic lesionsin youngmice, we demonstrate that a subtotal loss of SHCneurons com-parable in extent to that typically occurring in aging can alter neuronalactivity and functional connectivity pattern across multiple memorysystems, even though the lesions were insufficient to produce overtperformance deficits. Without affecting acquisition or retrieval perfor-mance in either the R/DMand non-R/DM tests, the lesions had nonethe-less altered memory-related Fos activity patterns in brain structurescritical for R/DM (viz., the hippocampus) as well as for non-R/DM, in-cluding structures implicated in emotional and procedural memory(viz., amygdala and striatum). In particular, the subtotal loss of SHCneurons was found to induce abnormal activation of hippocampal CA3

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:

under mnemonic challenge, and disrupt the strong link between thesepto-hippocampal system and amygdala specifically activated by R/DMexpression. Based on similarities between present functional alterationsinduced by SHC lesions and those previously reported in aged humansubjects, ourfindings lend direct support to the current conceptualizationof brain memory systems and provide new insights into the role of SHCneurons in memory deterioration during the aging process. Namely,SHC degeneration may predispose future deterioration in DM by alteringhippocampal function as well as the functional balance among multiplememory systems.

Our novel analytic approach provides direct support to the theory ofmultiple parallel memory systems (White and McDonald, 2002), ac-cording to which hippocampus, amygdala and striatum representthree core independent memory subsystems that process and recorddifferent aspects of experiences in parallel, and can interact in different(either competitive or cooperative) manners to support performance.The present analysis of correlation matrices in control subjects showsfor the first time that memory retrieval is associated with differentialfunctional connectivity between major memory systems as a functionof the form ofmemory processing concerned. Under normal conditions,the R/DM and non-R/DM retrieval tests triggered not only distinct pat-terns of activation but also distinct functional connectivity among thesepto-hippocampal, amygdalar and striatal systems. A functional linkwas reflected in positive/negative between-structure correlation (asopposed to “un-coupling” when no correlation was found). Comparedto the non-R/DM test, we show that R/DM testing in control animalstriggered strong activation in the septo-hippocampal system (especial-ly, CA1 and DG), the lateral amygdala as well as a strong functional con-nectivity between the two systems (negative correlations). While thecompetitive nature of interaction between the (hippocampus-based)R/DM and (striatum-based) procedural memory systems has been pre-viously highlighted (Devan and White, 1999; Mitchell and Hall, 1988;Poldrack and Packard, 2003; Poldrack et al., 2001), the picture is lessclear regarding the functional relationship between R/DMand emotion-almemory, which is based on the amygdala. Nevertheless, recent lesion(Farovik et al., 2011) and unit recording (Rutishauser et al., 2008)studies have shown that familiarity-based recognition relies on theamygdala whereas DM recollection requires the hippocampus, thussuggesting that the two systems sustain complementary functions inDM expression. In line with this suggestion, our present findings showthat a functional link between the septo-hippocampal and amygdalarsystems is engaged by R/DM retrieval.

Our study validates the hypothesis that acetylcholine plays a role inthe coordination of memory systems (Gold, 2004; Micheau andMarighetto, 2011) by showing that (even) a mild loss in SHC neuronsalters the task-specific recruitment of regional activity as well as func-tional connectivity among memory systems. The present findings indi-cate that SHC neurons regulate the recruitment of the non-R/DMsystems centered on the amygdala and striatum,when the test situationpreferentially engages R/DM expression. Specifically, besides aberranthippocampal hyperactivation centering on the CA3 subfield seenunder R/DM and non-R/DM testing, SHC lesions also affected selectivelyR/DM-related activation in the amygdala and striatum. Within theamygdala, partial SHC lesions prevented the activation of the lateralnucleus. The lesions also altered the R/DM-specific functional con-nectivity between amygdala and the septo-hippocampal system.By comparison, the pattern of functional connectivity observedunder non-R/DM retrieval test condition was not substantiallymodified by the lesions. Adding to our previous report on the influ-ence of hippocampal cholinergic transmission on amygdalar activ-ity in Pavlovian fear conditioning (Calandreau et al., 2006), thepresent observations demonstrated that even subtle alterationsof the SHC pathway are sufficient to alter septo-hippocampo-amygdalar functional coordination. Within the striatum, the lesionsled to hyperactivity in the dorsomedial caudate nucleus underthe R/DM test condition. This observation suggests that the SHC

ocampal cholinergic neurons alters memory-dependent measures of//dx.doi.org/10.1016/j.nbd.2013.01.010

9L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

neurons may normally underlie competitive hippocampo-striatalinteraction during R/DM retrieval.

The similarities between the functional alterations induced by the le-sions here and those previously seen in aged subjects readily suggest thatpartial loss of SHC neurons may contribute to the functional imbalanceamong memory systems existing in senescence. In particular, partialSHC depletion may be implicated in the reduction of hippocampo-amygdalar connectivity in aged subjects (Murty et al., 2009; St Jacqueset al., 2009). It may also contribute to the “dedifferentiation” betweenmemory systems as suggested by (i) hippocampus hyperactivationduring both R/DM and non R/DM expression, and (ii) striatal hyper-activation under R/DM testing here, matching a previous study in agedpeople (Dennis and Cabeza, 2011).

The similarities between functional alterations observed in our ani-mal model and those previously reported in cognitively unimpairedaged subjects at risk of AD suggest that SHC degenerationmay be linkedto the preclinical progression of pathological aging. This conclusion isbased on observations that the partial SHC lesions resulted in abnormalhippocampal activation under mnemonic challenge, and it did so with-out affecting acquisition or retrieval performance in the R/DM andnon-R/DM tests. As already mentioned, fMRI studies have revealed aparadoxical increase in hippocampal activity at prodromal stage of ADthat may be predictive of subsequent pathological DM decline (Bassettet al., 2006; Dickerson and Sperling, 2008; Mondadori et al., 2006;Quiroz et al., 2010; Sperling, 2007). Our studypinpointsmore specifical-ly that hippocampal hyperactivation can be part of a wider functionalde-regulation spanning across multiple memory systems; and thecause of which may be traced to the initial degeneration of the SHC.Hence, our findings highlight the potential interest of exploring func-tional connectivity in genetic mouse models of Alzheimer's disease asa critical step towards understanding more precisely the role of cholin-ergic degeneration in the development of this pathology.

The abnormal CA3 activation emerged following partial removal ofSHC neurons also fits the view that cholinergic loss may lead to a shiftin hippocampal computation and imbalance between R/DM encodingand recall processes (Gallagher et al., 2010). According to some compu-tational models (Buzsaki, 1989; Hasselmo, 2006; Hasselmo and Bower,1993), acetylcholine exerts opposite influence over the two comple-mentary functions of memory – encoding and retrieval – mediated bythe hippocampus. While acetylcholine promotes pattern separationnecessary for effective encoding of distinct representations of new de-clarative experiences to protect against interference, cholinergic activitycan also impede CA3-driven pattern completion necessary for efficientretrieval of past memory (Rolls and Kesner, 2006). Thus, loss of SHCinput would shift the balance between external cortical inputs intoDG/CA3 that drive pattern-separation and CA3 autoassociative networkthat mediates pattern-completion (Ikonen et al., 2002; Wilson et al.,2005); and the elevated CA3 activity found in our partial lesions modelsuggests a bias in favor of pattern-completion. It is known that aged peo-ple exhibit an increased tendency to engage in pattern completion at theexpense of pattern separation (Stark et al., 2010; Toner et al., 2009; Yassaand Stark, 2011; Yassa et al., 2011), and this shift is linked to hippocam-pal CA3 hyperactivity (Yassa et al., 2010). Thus, our findings strengthenprevious suggestions that SHCdegeneration contributes to the computa-tional shift in hippocampal function in aging that underlies the progres-sive sensitivity to interference characteristic of early DM degradation.

Finally, the functional alterations induced by the partial SHC lesionsincluding abnormal CA3 activation during memory retrieval might inthemselves constitute remedial compensation against the deleterious ef-fects of the lesions on performance. This is in keeping with the reportedabsence of aberrant CA3 activation among the subset of aged mice thatdid exhibit clear R/DM impairments in the same radial-maze task(Mingaud et al., 2008; Touzani et al., 2003). It follows that overt memoryimpairments would emerge when the hippocampus undergoes addi-tional age-related changes (for a review, see Burke and Barnes, 2006),such that the earlier compensatory changes can no longer sustain normal

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:/

performance. One candidate additional pathology may be the degenera-tion of the entorhinal cortex, which apparently precedes degeneration ofthe hippocampus in early Alzheimer's disease (Pennanen et al., 2004).Indeed, lesions of basal forebrain cholinergic neurons and of the entorhi-nal cortex, which had only minor effects on memory functions when in-troduced separately, can synergistically produce massive memorydeficits when combined (Traissard et al., 2007).

In conclusion, our analytic approach has helped in uncovering thefunctional impact of a mild SHC loss at the memory system level, andprovided new insights into the significance of cholinergic degenerationin aging. SHCdegenerationmay predispose futureDMdegradation by in-ducing a shift in hippocampal computational processes and functionalimbalance between multiple memory systems. Reconsidering the cho-linergic hypothesis in the context of prodromal AD may help the searchfor diseasemodifying andpreventative therapies tominimize the humansuffering and economic burden of AD and related cognitive senescence.

Acknowledgments

This work was supported by the C.N.R.S, Conseil Régionald'Aquitaine, Fondation NRJ-Institut de France and GIS longévité.

Disclosure statement

The authors have not any conflicts of interest related to thismanuscript.

References

Bartus, R.T., Dean, R.L., Pontecorvo, M.J., Flicker, C., 1985. The cholinergic hypothesis: ahistorical overview, current perspective, and future directions. Ann. N. Y. Acad. Sci.444, 332–358.

Bassett, S.S., Yousem, D.M., Cristinzio, C., Kusevic, I., Yassa, M.A., Caffo, B.S., Zeger, S.L.,2006. Familial risk for Alzheimer's disease alters fMRI activation patterns. Brain129, 1229–1239.

Boucard, A., Mons, N., Micheau, J., Nogues, X., 2009. Activating a memory system focus-es connectivity toward its central structure. Behav. Brain Res. 204, 226–234.

Burke, S.N., Barnes, C.A., 2006. Neural plasticity in the ageing brain. Nat. Rev. Neurosci.7, 30–40.

Buzsaki, G., 1989. Two-stage model of memory trace formation: a role for “noisy” brainstates. Neuroscience 31, 551–570.

Calandreau, L., Trifilieff, P., Mons, N., Costes, L., Marien, M., Marighetto, A., Micheau, J.,Jaffard, R., Desmedt, A., 2006. Extracellular hippocampal acetylcholine level con-trols amygdala function and promotes adaptive conditioned emotional response.J. Neurosci. 26, 13556–13566.

Chang, Q., Gold, P.E., 2003. Switching memory systems during learning: changes inpatterns of brain acetylcholine release in the hippocampus and striatum in rats.J. Neurosci. 23, 3001–3005.

Dennis, N.A., Cabeza, R., 2011. Age-related dedifferentiation of learning systems:an fMRI study of implicit and explicit learning. Neurobiol. Aging 32 (2318),e2317–e2330.

Devan, B.D., White, N.M., 1999. Parallel information processing in the dorsal striatum:relation to hippocampal function. J. Neurosci. 19, 2789–2798.

Dickerson, B.C., Sperling, R.A., 2008. Functional abnormalities of the medial temporallobe memory system in mild cognitive impairment and Alzheimer's disease: in-sights from functional MRI studies. Neuropsychologia 46, 1624–1635.

Eaton, J., 2002. Manual Network Theory: GNU Octave (Chapter Chapter).Efron, B., Tibshirani, R., 1991. Statistical data analysis in the computer age. Science 253,

390–395.Etchamendy, N., Desmedt, A., Cortes-Torrea, C., Marighetto, A., Jaffard, R., 2003. Hippo-

campal lesions and discrimination performance of mice in the radial maze: sparingor impairment depending on the representational demands of the task. Hippocam-pus 13, 197–211.

Farovik, A., Place, R.J., Miller, D.R., Eichenbaum, H., 2011. Amygdala lesions selectivelyimpair familiarity in recognition memory. Nat. Neurosci. 14, 1416–1417.

Gallagher, M., Rapp, P.R., 1997. The use of animal models to study the effects of agingon cognition. Annu. Rev. Psychol. 48, 339–370.

Gallagher, M., Bakker, A., Yassa, M.A., Stark, C.E., 2010. Bridging neurocognitive agingand disease modification: targeting functional mechanisms of memory impair-ment. Curr. Alzheimer Res. 7, 197–199.

Gold, P.E., 2004. Coordination of multiple memory systems. Neurobiol. Learn. Mem. 82,230–242.

Grady, C.L., 2008. Cognitive neuroscience of aging. Ann. N. Y. Acad. Sci. 1124, 127–144.Hasselmo, M.E., 2006. The role of acetylcholine in learning and memory. Curr. Opin.

Neurobiol. 16, 710–715.Hasselmo,M.E., Bower, J.M., 1993. Acetylcholine andmemory. Trends Neurosci. 16, 218–222.

ocampal cholinergic neurons alters memory-dependent measures of/dx.doi.org/10.1016/j.nbd.2013.01.010

10 L. Brayda-Bruno et al. / Neurobiology of Disease xxx (2013) xxx–xxx

Hedden, T., Gabrieli, J.D., 2005. Healthy and pathological processes in adult develop-ment: new evidence from neuroimaging of the aging brain. Curr. Opin. Neurol.18, 740–747.

Ho, N.F., Han, S.P., Dawe, G.S., 2009. Effect of voluntary running on adult hippocampalneurogenesis in cholinergic lesioned mice. BMC Neurosci. 10, 57.

Horwitz, B., Grady, C.L., Schlageter, N.L., Duara, R., Rapoport, S.I., 1987. Intercorrelations of re-gional cerebral glucose metabolic rates in Alzheimer's disease. Brain Res. 407, 294–306.

Ikonen, S., McMahan, R., Gallagher, M., Eichenbaum, H., Tanila, H., 2002. Cholinergic systemregulation of spatial representation by the hippocampus. Hippocampus 12, 386–397.

McIntyre, C.K., Pal, S.N., Marriott, L.K., Gold, P.E., 2002. Competition between memorysystems: acetylcholine release in the hippocampus correlates negatively withgood performance on an amygdala-dependent task. J. Neurosci. 22, 1171–1176.

McIntyre, C.K., Marriott, L.K., Gold, P.E., 2003. Cooperation between memory systems:acetylcholine release in the amygdala correlates positively with performance ona hippocampus-dependent task. Behav. Neurosci. 117, 320–326.

Mesulam, M.M., Mufson, E.J., Wainer, B.H., Levey, A.I., 1983. Central cholinergic path-ways in the rat: an overview based on an alternative nomenclature (Ch1–Ch6).Neuroscience 10, 1185–1201.

Micheau, J., Marighetto, A., 2011. Acetylcholine and memory: a long, complex and cha-otic but still living relationship. Behav. Brain Res. 221, 424–429.

Minati, L., Grisoli, M., Bruzzone, M.G., 2007. MR spectroscopy, functional MRI, anddiffusion-tensor imaging in the aging brain: a conceptual review. J. Geriatr. Psychi-atry Neurol. 20, 3–21.

Mingaud, F., Le Moine, C., Etchamendy, N., Mormede, C., Jaffard, R., Marighetto, A.,2007. The hippocampus plays a critical role at encoding discontiguous events forsubsequent declarative memory expression in mice. Hippocampus 17, 264–270.

Mingaud, F., Mormede, C., Etchamendy, N., Mons, N., Niedergang, B., Wietrzych, M.,Pallet, V., Jaffard, R., Krezel, W., Higueret, P., Marighetto, A., 2008. Retinoidhyposignaling contributes to aging-related decline in hippocampal function inshort-term/working memory organization and long-term declarative memoryencoding in mice. J. Neurosci. 28, 279–291.

Mitchell, J.A., Hall, G., 1988. Caudate–putamen lesions in the rat may impair or poten-tiate maze learning depending upon availability of stimulus cues and relevance ofresponse cues. Q. J. Exp. Psychol. B 40, 243–258.

Mondadori, C.R., Buchmann, A., Mustovic, H., Schmidt, C.F., Boesiger, P., Nitsch, R.M.,Hock, C., Streffer, J., Henke, K., 2006. Enhanced brain activity may precede the diag-nosis of Alzheimer's disease by 30 years. Brain 129, 2908–2922.

Moreau, P.H., Cosquer, B., Jeltsch, H., Cassel, J.C., Mathis, C., 2008. Neuroanatomical andbehavioral effects of a novel version of the cholinergic immunotoxin mu p75-saporin in mice. Hippocampus 18, 610–622.

Murty, V.P., Sambataro, F., Das, S., Tan, H.Y., Callicott, J.H., Goldberg, T.E., Meyer-Lindenberg, A., Weinberger, D.R., Mattay, V.S., 2009. Age-related alterations in sim-ple declarative memory and the effect of negative stimulus valence. J. Cogn.Neurosci. 21, 1920–1933.

Parent, M.B., Baxter, M.G., 2004. Septohippocampal acetylcholine: involved in but notnecessary for learning and memory? Learn. Mem. 11, 9–20.

Paxinos, G., Franklin, K.B.J., 2001. The Mouse Brain in Stereotaxic Coordinates. Academ-ic Press, USA (Chapter Chapter).

Pennanen, C., Kivipelto, M., Tuomainen, S., Hartikainen, P., Hanninen, T., Laakso, M.P.,Hallikainen, M., Vanhanen, M., Nissinen, A., Helkala, E.L., Vainio, P., Vanninen, R.,Partanen, K., Soininen, H., 2004. Hippocampus and entorhinal cortex in mild cogni-tive impairment and early AD. Neurobiol. Aging 25, 303–310.

Pepeu, G., Giovannini, M.G., 2009. Cholinesterase inhibitors and beyond. Curr.Alzheimer Res. 6, 86–96.

Pihlajamaki, M., Jauhiainen, A.M., Soininen, H., 2009. Structural and functional MRI inmild cognitive impairment. Curr. Alzheimer Res. 6, 179–185.

Please cite this article as: Brayda-Bruno, L., et al., Partial loss in septo-hippbrain connectivity without overt memory ..., Neurobiol. Dis. (2013), http:

Poldrack, R.A., Packard, M.G., 2003. Competition among multiple memory systems:converging evidence from animal and human brain studies. Neuropsychologia41, 245–251.

Poldrack, R.A., Clark, J., Pare-Blagoev, E.J., Shohamy, D., Creso Moyano, J., Myers, C.,Gluck, M.A., 2001. Interactive memory systems in the human brain. Nature 414,546–550.

Pych, J.C., Chang, Q., Colon-Rivera, C., Gold, P.E., 2005. Acetylcholine release in hippo-campus and striatum during testing on a rewarded spontaneous alternation task.Neurobiol. Learn. Mem. 84, 93–101.

Quiroz, Y.T., Budson, A.E., Celone, K., Ruiz, A., Newmark, R., Castrillon, G., Lopera, F.,Stern, C.E., 2010. Hippocampal hyperactivation in presymptomatic familialAlzheimer's disease. Ann. Neurol. 68, 865–875.

Rolls, E.T., Kesner, R.P., 2006. A computational theory of hippocampal function, and em-pirical tests of the theory. Prog. Neurobiol. 79, 1–48.

Rutishauser, U., Schuman, E.M., Mamelak, A.N., 2008. Activity of human hippocampaland amygdala neurons during retrieval of declarative memories. Proc. Natl. Acad.Sci. U. S. A. 105, 329–334.

Sperling, R., 2007. Functional MRI studies of associative encoding in normal aging, mildcognitive impairment, and Alzheimer's disease. Ann. N. Y. Acad. Sci. 1097, 146–155.

Sperling, R.A., Dickerson, B.C., Pihlajamaki, M., Vannini, P., LaViolette, P.S., Vitolo, O.V.,Hedden, T., Becker, J.A., Rentz, D.M., Selkoe, D.J., Johnson, K.A., 2010. Functional al-terations in memory networks in early Alzheimer's disease. Neuromolecular Med.12, 27–43.

St Jacques, P.L., Dolcos, F., Cabeza, R., 2009. Effects of aging on functional connectivity ofthe amygdala for subsequent memory of negative pictures: a network analysis offunctional magnetic resonance imaging data. Psychol. Sci. 20, 74–84.

St Jacques, P., Dolcos, F., Cabeza, R., 2010. Effects of aging on functional connectivityof the amygdala during negative evaluation: a network analysis of fMRI data.Neurobiol. Aging 31, 315–327.

Stark, S.M., Yassa, M.A., Stark, C.E., 2010. Individual differences in spatial pattern separationperformance associated with healthy aging in humans. Learn. Mem. 17, 284–288.

Toner, C.K., Pirogovsky, E., Kirwan, C.B., Gilbert, P.E., 2009. Visual object pattern separa-tion deficits in nondemented older adults. Learn. Mem. 16, 338–342.

Touzani, K., Marighetto, A., Jaffard, R., 2003. Fos imaging reveals ageing-related changesin hippocampal response to radial maze discrimination testing in mice. Eur. J.Neurosci. 17, 628–640.

Traissard, N., Herbeaux, K., Cosquer, B., Jeltsch, H., Ferry, B., Galani, R., Pernon, A.,Majchrzak, M., Cassel, J.C., 2007. Combined damage to entorhinal cortex and cho-linergic basal forebrain neurons, two early neurodegenerative features accompa-nying Alzheimer's disease: effects on locomotor activity and memory functionsin rats. Neuropsychopharmacology 32, 851–871.

White, N.M., McDonald, R.J., 2002. Multiple parallel memory systems in the brain of therat. Neurobiol. Learn. Mem. 77, 125–184.

Wilson, I.A., Ikonen, S., Gallagher, M., Eichenbaum, H., Tanila, H., 2005. Age-associatedalterations of hippocampal place cells are subregion specific. J. Neurosci. 25,6877–6886.

Yassa, M.A., Stark, C.E., 2011. Pattern separation in the hippocampus. Trends Neurosci.34, 515–525.

Yassa, M.A., Lacy, J.W., Stark, S.M., Albert, M.S., Gallagher, M., Stark, C.E., 2010. Patternseparation deficits associated with increased hippocampal CA3 and dentate gyrusactivity in nondemented older adults. Hippocampus 21, 968–979.

Yassa, M.A., Mattfeld, A.T., Stark, S.M., Stark, C.E., 2011. Age-related memory deficitslinked to circuit-specific disruptions in the hippocampus. Proc. Natl. Acad. Sci. U.S. A. 108, 8873–8878.

Zar, J., 1999. Biostatistical Analysis. Prentice Hall, Upper Saddle River, New Jersey(Chapter Chapter).

ocampal cholinergic neurons alters memory-dependent measures of//dx.doi.org/10.1016/j.nbd.2013.01.010