Memory Processing by the Limbic System: Role of Specific Neurotransmitter Systems

BBR 1509

Memory processing by the limbic system: role of specific neurotransmitter systems

Ivan Izquierdo ~'*, Jorge H. Medina b, Marino Bianchin a, Roger Walz a, Marilene S. Zanatta a, Ricardo C. Da Silva a, Marcia Bueno E Silva a, Anelise C. Ruschel a and Natalia Paczko a

a Centro de Memoria, Departamento de Bioquimica, lnstituto de Biociencias, Universidade Federal do Rio Grande do Sul, 90046-900 Porto Alegre, RS (Brazil), b Laboratorio de Neurorreceptores, Instituto de Biologia Celular, Facultad de Medicina, Universidad de Buenos Aires, Paraguay 2155, 1121 Buenos Aires (Argentina)

(Received 15 January 1993) (Accepted 11 October 1993)

Key words: Long-term potentiation and memory; Glutamatergic mechanisms and memory; GABAergic mechanisms and memory; Amygdala and memory; Hippocampus and memory; Medial septum and memory; Entorhinal cortex and memory

Experiments using localized infusions into selected brain structures of agonists and antagonists of various synaptic receptors, given before or after behavioral training, have led to the following conclusions: (1) Memory is processed shortly after training in the amygdala, medial septum and hippocampus by glutamatergic NMDA and AMPA receptors activated in that sequence. Cholinergic muscarinic receptors are activated concurrently with the former. GABAA receptors modulated by brain benzodiazepines and by beta-noradrenergic receptors inhibit the process. (2) The sequential involvement of NMDA and AMPA receptors suggests that long-term potentiation (LTP) of the synapses activated by the learning experiences in the hippocampus and/or amygdala and medial septum is the crucial event. Expression of this LTP at the time of testing is necessary for retrieval: AMPA receptor blockade in the hippocampus and amygdala at the time of testing hinders retrieval. This suggests that the LTP underlies the memory process itself. (3) The amygdala, medial septum and hippocampus mediate different types of memory and/or different components of memories. The entorhinal cortex, through mechanisms that require intact NMDA receptors and are inhibited by GABAA receptors, intervenes in post-training memory processing 90-180 rain after the other limbic regions. The entorhinal cortex integrates consecutively acquired memories; this role could be maintained by the LTP that is generated after training in the amygdala, hippocampus and medial septum. Post-training intervention of the entorhinal cortex does not occur if this region is inhibited at the time of training.

INTRODUCTION is the case with long-term memories, the maintenance phase of LTP can persist for very long periods even if Like long-term potentiation (LTP) 8'44'54'56,long-term its expression is withheld or prevented 56. The early, memories are initially labile and after a few minutes labile phase of long-term memories is often called the become stable for very long periods 25'33'46. The initial consolidation phase, because it is believed that it in- labile phase of LTP is called the induction phase. It is volves a process of strengthening of the underlying neu- mediated by N-methyl-D-aspartate (NMDA) gluta- ral processes 25"46. LTP, of course, consists of the matergic receptors 8, can be modulated by cholinergic strengthening of synaptic transmission 8'27'56. In LTP 56, muscarinic and beta-noradrenergic receptors (see refs. as in long-term memory 25'61'67, re-iteration of the stimuli 30, 42) and is inhibited by gamma-amino-butyrate type that brought them into being will trigger their expres- A (GABAA) receptors (see ref. 8). The late, stable phase sion. The expression of memories is called re- of LTP is expressed through 7-amino-3-hydroxy-5- trieval25,21,67. methyl-4-isoxazole propionate (AMPA) glutamatergic Indeed, the two most widely accepted current hy- receptors and is called the maintenance phase 8'56. As potheses of memory storage are that it involves LTP 1"54"55 and/or the operation of neural networks 39'43. * Corresponding author. Centro de Memoria, Departamento de Bio- These hypotheses are not mutually exclusive. The ex- quimica, Instituto de Biociencias, UFRGS (centro), 90050 Porto pression of LTP in the particular synapses that had Alegre, RS, Brazil. Fax: (55)(51)227-2343. been specifically activated during training could sub- 92

serve important functions in amplifying the signals that The agonist, glutamate has opposite effects to those are to enter the neural networks 27'37. In addition, LTP of AP5 and the effects of both are not restricted to the can underlie the function of components of the net- amygdala. In fact, the two drugs affect memory retro- works 1. Network theory explains the continued func- actively when infused into the hippocampus and the tioning of neural circuits through alternative pathways medial septum as well, depending on the task. Post- in spite of lesions 43, which may in turn explain the training infusion of AP5 into the amygdala, medial sep- persistence of memories after extensive brain damage turn or hippocampus blocks the consolidation of step- as is seen in the early stages of Alzheimer's disease or down inhibitory avoidance 3235. The infusion of AP5 of other organic brain syndromes 21. into the hippocampus but not into the amygdala or Lesion and drug infusion studies have shown that the medial septum causes retrograde amnesia for habitua- main areas involved in consolidation are the amygdala, tion to a novel environment32'35. The agonist, glutamate, the hippocampus, the medial septum and the entorhinal has effects exactly opposite to those of AP5. Its imme- cortex, which are parts of the so-called limbic diate post-training infusion into the amygdala, medial system 16'17"24'25"28"32-35'46. Evidence suggests that other septum or hippocampus causes retrograde facilitation regions of the brain are involved in parallel 25"64 or al- of inhibitory avoidance; its intrahippocampal but not ternatively62 to the limbic areas of the temporal lobe. its intraseptal or intra-amygdala infusion causes retro- These parallel or alternative areas include the cerebel- grade facilitation of habituation 3~. lum 64, the caudate nucleus 2"68 and various regions of Recent evidence suggests a role for cholinergic mus- the cerebral cortex 1'62. It is possible that these extra- carinic receptors in consolidation processes in addition limbic circuits specialize in types of memory not to glutamatergic receptors. A role for cholinergic handled by the limbic structures referred to above 2'62. mechanisms in memory had long been suspected mainly The present article reviews recent data from our because of the fact that systemic administration of the laboratories on the neurotransmitter mechanisms in- cholinergic muscarinic receptor antagonist, scopola- volved in memory processing by the amygdala, hippoc- mine, causes amnesia in humans and animals (see refs. ampus and medial septum, and on the interaction of 26, 38). these structures with the entorhinal cortex in the for- The comparative effect on memory consolidation of mation of long-term memories. As will be seen, the data cholinergic muscarinic antagonists or agonists given indicate that LTP in the amygdala, hippocampus and into specific brain structures was studies for the :first medial septum is at the core of the memory process. time by our group. In a step-down inhibitory avoidance task, post-training intra-amygdala, intraseptal or intrahippocampal scopolamine administration causes retro- NEUROTRANSMITTERS INVOLVED IN MEMORY CON- SOLIDATION grade amnesia, whereas that of the agonist, oxotremo- rine, causes instead retrograde facilitation 32"35. In the Previous work in several laboratories had shown that habituation task, the intrahippocampal infusion of sco- glutamatergic synapses play a key role in memory. This polamine causes retrograde amnesia and that of ox- is not surprising since glutamatergic synapses are by far otremorine causes retrograde facilitation; intra- the most abundant excitatory synapses of the brain, amygdala or intraseptal infusions of these substances and can generate LTP. The systemic administration of was ineffective32"3s. indirect (i.e. Ca 2+ channel) blockers of NMDA (N- It is possible that cholinergic muscarinic transmis- methyl-D-aspartate) receptors to glutamic acid causes sion may act by facilitating the induction of LTP and amnesia for a variety of tasks in rats 59'65. Intracere- neighboring glutamatergic synapses 42. The amnesia broventricular infusion of the indirect NMDA receptor caused by scopolamine in humans can be alleviated by antagonist, dizolcipine, or of the direct antagonist, D-2- D-cycloserine, a modulator of the glycine site of NMDA amino-5-phosphonopentanoic acid (AP5), hinders re- receptors 38. tention of spatial learning54'55. AP5 and its heptanoic acid analogue, AP7, disrupt retention of a conditioned startle response when infused into the amygdala prior THE ROLE OF GABA A RECEPTORS to training 51. D-cycloserine, a partial agonist at the glycine modulatory site of NMDA receptors 57, enhances The immediate post-training infusion of the indirect memory in rats 53 and has been proposed as a useful (C1- channel) blocker of GABAA receptors, picro- drug in Alzheimer's disease TM. Sensitivity of the NMDA toxin, into the amygdala, medial septum or hippocam- receptor complex to glycine is reduced in Alzheimer's pus causes retrograde memory facilitation of inhibitory disease s7. avoidance behaviour and counteracts the amnesic ac- 93 tion of AP5 and/or scopolamine32'35. In the habitation amygdala25'28 and, at least in the amygdala, are possi- task, similar effects are found but only in the hippoc- bly mediated by beta-noradrenergic synapses 46. Recent ampus 32'35. The effect of systemic or intra-amygdala data suggest an involvement of dopaminergic synapses picrotoxin on inhibitory avoidance behavior is shared in the caudate nucleus in post-training memory pro- by another C1- channel blocker. R05-4864 (4'- cesses 68. It is not known whether dopaminergic syn- chlordiazepam) 11. Post-training systemic picrotoxin apses in the septum, amygdala or hippocampus are also administration has long been known to cause memory involved in memory; there are dopaminergic (and se- facilitation 3'47. The GABA A receptor agonist, musci- rotonergic) terminals in these structures, and they have mol, causes retrograde amnesia when given systemi- been implicated in a number of functions relevant to cally7 or when infused into the amygdala, medial sep- brain psychopathology9'19. tum or hippocampus in the case of inhibitory avoidance, or when infused into the hippocampus but not the amygdala or septum in the case of habituation to a THE MAJOR INTERACTIONS AMONG NEUROTRANSMIT- novel environment32'35. Post-training infusion of the TERS INVOLVED IN MEMORY PROCESSING GABAA receptor antagonist, bicuculline, into the amygdala causes retrograde facilitation for aversive be- The major interactions among neurotransmitter haviors when given post-training either systemically4. mechanisms in the amygdala, medial septum and hip- The GABAA receptors involved in memory modu- pocampus commented upon above are: lation in the medial septum, amygdala and hippocam- (1) GABA A receptors inhibit the cells that are acti- pus are in turn regulated by benzodiazepines or vated by glutamatergic NMDA and cholinergic mus- benzodiazepine-like molecules released in the same carinic receptors 3z'35. brain structures 33"7°. The benzodiazepines are released (2) The GABAergic synapses are positively modu- in relation to the degree of anxiety and/or stress asso- lated by beta-noradrenergic synapses 32'35 and by ben- ciated with each task ~2. This topic has been recently zodiazepines, possibly of endogenous origin, released reviewed in extenso 31"33'34. by the training experiences 29'33'34"49'7°. (3) Administered norepinephrine may excite the cells that are excited by glutamatergic and cholinergic mus- ROLE OF BETA-NORADRENERGIC RECEPTORS AND carinic receptors 3~ and/or promote LTP at the OTHER SUBSTANCES glutamatergic receptors 3°. The postulation of these neurotransmitter interac- The systemic, intra-amygdala, intraseptal or intra- tions is supported by histochemical and electrophysi- hippocampal adminstration of beta-adrenoceptor an- ological studies on the three regions4°'5°; see ref. 29. tagonists usually has no effect on memory of its own but Interestingly, the interactions are similar in the three hinders the memory enhancing effect of picrotoxin 32. structures despite their anatomical and functional dif- This suggests that beta-noradrenergic receptors modu- ferences. This suggests that similar synaptic mecha- late the influence of GABAergic synapses on memory nisms might develop ontogeneticaUy in structures that consolidation32'3~. Post-training intra-amygdala 32'41, specialize in the processing of one or other kind of intraseptal or intrahippocampa132'35 norepinephrine in- memory35. The amygdala processes alerting 5 or aver- fusion causes memory facilitation, suggesting that, in sive 13 memories or components of memory. The medial addition to their influence on GABAergic terminals, septum and hippocampus process working memory and noradrenergic receptors may also stimulate memory on spatial and olfactory information52'54, but by virtue of their own 29. As mentioned above concerning cholin- their different input-output connections they probably ergic muscarinic receptors, it is possible that noradr- process different data pertaining to these domains, or energic receptors may also act by promoting the gen- the same data differently 16. This specialization of the eration of LTP at glutamatergic receptors (see ref. 30). amygdala, medial septum and hippocampus explains Other neurotransmitter systems may also participate their differential involvement in the consolidation of in memory consolidation. The systemic or intra- different behaviors mentioned above 29'32'33'35. amygdala administration of the GABAB receptor Stress hormones (epinephrine, adrenocorticotropin, blocker, baclofen, impairs retention6. This suggests an vasopressin) modulate memory consolidation possibly involvement of GABAB receptors in the amygdala in through influences on central beta-noradrenergic syn- consolidation. Much evidence suggests a role of beta- apses 25'46. The mapping of the effect of the stress hor- endorphin in post-training memory processes 25'28'46. Its mones, or of the opioid, dopaminergic and serotoner- effects may be exerted in the medial septum and in the gic systems onto the glutamatergic, cholinergic and 94

GABAergic synapses discussed above requires further It is possible that the underlying mechanism of LTP investigation. Surely regional differences are to be ex- may persist in the medial septum, amygdala and hip- pected. For example, beta-endorphin-containing termi- pocampus beyond the initial 180 min after training dur- nals are found in the amygdala and medial septum but ing which it is expressed (see ref. 56), and that this not in the hippocampus; the distribution of dopamin- persistence underlies memory storage. If this were so, ergic and serotonergic terminals in these areas is dif- memory would be the LTP, and its retrieval would ferent (Ref. 25, 29; etc.). depend merely on the reactivation (or the renewed ex- For a review of interactions among neurotransmitter pression) of the LTP at the time of testing. A case for systems involved in memory processing, see refs. 29, the identity of consolidation and retrieval processes 35. was made by Spear and Mueller 6~ and Izquierdo 2s with particular reference to the need to reiterate stimuli that THE ROLE OF LTP IN MEMORY PROCESSES had been crucial to consolidation, at the time of testing, in order that retrieval may occur > . If consolidation were to be defined in synaptic terms, Recent findings from the present authors support this it should be a process whereby responses at the syn- idea. If retrieval depends on the reactivation of the apses involved in each particular experience are expression of hippocampal or amygdala LTP by a re- strengthened while in a labile state 29. The best known iteration of the stimuli that had caused in the first place, process whereby synaptic responses are strengthened is then blockade of AMPA receptors at the cells originally LTP at glutamatergic synapses; LTP is indeed labile activated by the learning experience in the hippocam- during its induction phase because of its susceptibility pus and amygdala should block both the renewed ex- to inhibition by GABAA receptors 8. pression of the LTP and, as a consequence, retrieval. Units in the amygdala, medial septum and hippoc- In recent experiments from our laboratories, it was ob- ampus respond to different sensory modalities, and served that the bilateral administration of CNQX each cell has a particular pattern of response 2°'45. Le- (0.5 #g) into the amygdala and hippocampus 10 min sion studies by Mishkin and his coworkers indicate that prior to inhibitory avoidance testing, in rats, completely information pertinent to learning experiences is relayed blocked expression of the memory of this task. Two from sensory areas onto the amygdala and hippocam- hours later, when the effect of the drug had worn olT*~, pus via the perirhinal and the entorhinal cortex, during retention test performance became fully normal again. or very shortly after acquisition (see refs. 2, 29). Thus, Similar injections of 0.5 or 1.25 #g of CNQX into ei- each learning experience should elicit a pattern of unit ther the amygdala or hippocampus reduced retention response in the amygdala, medial septum and/or hip- test performance of this task only by approximately 40 pocampus that is conceivably unique for each experi- or 500/o . It had been previously shown that memory of ence29,37. the inhibitory avoidance task depends both on the LTP induction is mediated by glutamatergic NMDA amygdala and the hippocampus (and medial septum) receptors and is therefore sensitive to blockade by AP5, (see above and 32,33,3s). In the habituation task (free and is maintained through AMPA receptors sensitive exploration of the training apparatus for 1 rain), which to the antagonist, 6-cyano-7-nitroquinoxaline-2,3-dione appears to depend on the hippocampus but not the (CNQX) TM. A role of hippocampal LTP in memory amygdala or septum 32'33'3s, the bilateral intrahippoc- has been suggested by Morris, LynchTM and oth- ampal administration of CNQX (0.5 #g) 10 min prior ers (e.g., refs. 1, 19, 27, 64). to testing blocked the expression of memory completely; The retention of inhibitory avoidance is hindered by the intra-amygdala pre-training infusion of 0.5 or the infusion of AP5 into the amygdala, hippocampus 37 1.25/~g of CNQX had no effect. Davis ~4 showed an or the medial septum 66 immediately but not 90 min amnestic effect of 0.375/~g of CNQX given bilaterally after training, and by CNQX infused into any of these before testing in the amygdala, in a fear-potentiated three structures 0, 90 or 180 but not 360 min after train- startle task in which previous data had suggested that ing 37'66. This strongly suggests that consolidation is me- the amygdala was crucially involved in retention13. In diated by LTP in synapses of the amygdala, medial both our and Davis's experiments on pre-test CNQX septum and hippocampus specifically activated by each the training-test interval was 24 h. training experience, and that this LTP needs to be ex- Thus, these findings support the hypothesis that me- pressed during at least 180 min after training in order mory is (a consequence of) LTP in limbic structures up for consolidation to Occur 29'37'66. CNQX specifically to at least 24 h after training. They do not, however, blocks the expression of LTP which is mediated by provide any hint as to the mechanism of retention or of AMPA receptors 8'56. retrieval more than 24 h after training. Many data, in- 95 cluding the lack of retrograde amnesia seen in patient grative role: the infusion of AP5 or muscimol into this H.M. for memories stored weeks or years prior to bi- region 90 min after inhibitory avoidance training pre- lateral temporal lobectomy, indicate that after some vents the summation of the trace left by this training time temporal lobe structures are no longer important with that of a subsequent training session carried out for the storage or retrieval of many if not most memo- 30 mi later 16'17. Such a role could be of great impor- ries 25'62. Storage and retrieval elsewhere in the brain at tance for the formation of memory files25 or complex long times after acquisition, may or may not involve memories and, if lacking, could conceivably lead to a LTP ~, but are believed to depend on the operation of severe disintegration of cognitive processes, such as is neural networks 27'39"43. seen in Alzheimer's disease (see refs. 21, 23, 36). Recent data from our laboratories, obtained in col- laboration with Paul Willner of the University of THE ROLE OF THE ENTORHINAL CORTEX Swansea, show that if the entorhinal cortex is inhibited during and immediately after training, it does not par- The entorhinal cortex has two-way monosynaptic ticipate in the memory processing 90 for 100 min after connections with the amygdala, hippocampus and me- training. The infusion of muscimol into discrete brain dial septum, and is also interconnected, through the regions produces a localized inhibition measurable by neighboring perirhinal region, with sensory and asso- a reduction of 2-deoxyglucose uptake extending for ciative areas of the neocortex 23"36'69. Therefore, it is about 30 mm 3 and lasting for about 60 min 48. In rats strategically located both to convey signals from these that received a bilateral infusion of muscimol into the cortical areas to the amygdala, septum and hippocam- entorhinal cortex, a second infusion of muscimol pus (see ref. 2), and to handle memory-relevant infor- 100 min after training has no amnestic effect on inhibi- mation after it has been processed by these struc- tory avoidance in animals trained with a high footshock tures 16'17. Entorhinal lesions disrupt various types of level (0.5 mA). Further, in animals trained with a low spatial and non-spatial learning in different spe- intensity footshock (0.2 mA) in which two training ses- cies 60'63'71. The most prominent and most typical lesions with a 120 min interval between sessions are sions of Alzheimer's disease are in the entorhinal cor- needed in order to obtain good memory (see above), the tex 23'36. The early appearance of such lesions signals pre-training infusion of muscimol into the entorhinal the onset of the disintegration of memory and cognition cortex prevents the summation of the two training ses- typical of this disease 15. sions. Thus, in the absence of a normally functioning We have recently obtained evidence for a delayed entorhinal cortex at the time of training or in the im- post-training role of the entorhinal cortex in memory, mediate post-training period, animals can learn, pre- secondary to amygdala, hippocampal or septal activa- sumably using other brain areas29; they are, however, tion. AP5 or muscimol infused bilaterally into the en- incapable of integrating consecutive memories. torhinal cortex 90 or 180, but not 0 or 360 min after These findings argue in favor of an early role of the training in habituation or in inhibitory avoidance cause entorhinal cortex in memory, during or very shortly full retrograde amnesia for both tasks. Thus, the en- after training, in addition to its delayed post-training torhinal cortex is not only important as a relay station role referred to above. It is reasonable to think that this between the sensory cortex and the amygdala and hip- early role consists of conveying learning-related signals pocampus 2, but also essential for memory after the to the hippocampus, amygdala and medial septum from amygdala, medial septum and hippocampus had inter- sensory and polysensory regions of the cortex 2. These vened in it, and for a limited period of time: 90 to signals could be those that trigger the LTP in the cells 180min from training. Its intervention relies upon of those limbic structures that had been specifically glutamatergic NMDA receptors and is inhibited by activated by each training experience35'37'66, which in GABAA receptors: the intra-entorhinal infusion of AP5 turn maintains the entorhinal cortex active late after or muscimol 90 or 180 (but not 0 or 60) min after training 16,17. training causes amnesia for inhibitory avoidance and In addition, these findings on the effect ofpre-training for habituation to a novel environment ~6'17. The late muscimol given into the entorhinal cortex indicate that role of the entorhinal cortex in post-training memory a form of memory (inhibitory avoidance) that is nor- processing might be secondary to the LTP in the mally processed after training by the entorhinal cortex, amygdala, hippocampus 37 and medial septum 66 which and for which this processing is normally indispens- is expressed for up to 180 min after training. able ~6'17, can be also processed, although defectively, The late intervention of the entorhinal cortex in post- by other structures when the entorhinal cortex is not training memory processing apparently plays an inte- operative during and very shortly after training. This 96 point may be relevant to the observations of Thomp- 6 Castellano, C., Brioni, J.D., Nagahara, A. and McGaugh, J.L.. Post-training systemic and intra-amygdala administration of the son 64 that eye-blink conditioning, although normally GABA-b agonist baclofen impairs retention, Behav. ,~2,ural Biol., processed or modulated by the hippocampus can occur 52 (1989) 170-179. in the absence of this structure, and to the more gen- 7 Castellano, C. and McGaugh, J.L., Effect of p~st-training bicu- eral issue of the formation of memories in the absence culline and muscimol on retention: Lack of state-dependency, of the temporal lobe, which is limited, defective, and/or Behav. Neural Biol., 54 (1990) 156-164. restricted to the more simple, primitive and/or gradu- 8 Collingridge, G.L., Bashir, Z.I., Blake, J .F., Davies, C.H., Davies, ally acquired forms of memories 2"62. S.N., Irving, A.J., Randall, A.D. and Schofield, J.G., Amino acid receptors and LTP in area CAI of the hippocampus. In L.R. Mishkin and his group 2 have argued in favor of the Squire and E. Lindenlaub (Eds.), The Biology q/Memory, Schat- position that procedural memories or 'habits' are pro- tauer, Stuttgart, 1990, pp. 233-250. cessed by cortico-striatal systems, whereas the tempo- 9 Costall, B., Domeney, A.M., Kelly, M.E. and Naylor, R.J., Pre- ral lobe structures discussed above are involved rather clinical evidence for the anxiolytic effect of E-4424, Behav. Phar- macol., 3 (1992) 22-23. in the processing of declarative or explicit memories. 10 Cotman, C.W., Nieto-Sampedro, M. and Harris, W., Synapse Squire and his coworkers 6~'71 support a similar replacement in the nervous system of adult vertebrates, Physiol. dichotomy, but suggest a diencephalic locus for the Rev. 6l (1981) 684-784. processing of procedural or implicit memories. These 11 Da Cunha, C., Huang, C.H., Walz, R., Dias, M., Koya, R., authors have restricted their analysis mostly to the type Bianchin, M., Pereira, M.E., lzquierdo, 1. and Medina, J.H,, Memory facilitation by post-training intraperitoneal, intraven- of memories (or components of memories) that humans tricular and intra-amygdala injection of Ro5-4864, Brain Res. 544 or animals with temporal lobe lesions can acquire, how- (1991) 133-136. ever, and neural regeneration and reorganization, and 12 Da Cunha C., Wolfman, C., Izquierdo, 1. and Medina, J.H., the entry into play of vicarious circuits are problems Anxiety and brain benzodiazepine-like molecules. In I. Izquierdo intrinsic to lesion studies 25'58. In fact, entorhinal lesions and J.H. Medina (Eds.), Endogenous Benzodiazepines: Structure, Distribution and Function, Ellis Horwood, Chichester, 1993, are followed after a few days by a reactive synaptoge- pp. 81-88. nesis in all surviving afferent systems to the dentate 13 Davis, M., The role of the amygdala in fear-potentiated startle: gyrus 1°'58, which complicates the interpretation of the implications for animal models of anxiety, Trends Pharmacol. Sci., behavioral effects of such lesions. 13 (1992) 35-41. Further research on the role in memory of circuits 14 Davis, M., Neural systems involved in fear and anxiety, Behav. Brain Res., in press. parallel and/or alternative to the temporal lobe struc- 15 De Leon, M.J., George, A.E., Golomb, J., Convit, A. and De tures and the medial septum is desirable, if possible Santi, S., Hippocampal atrophy: an early brain marker for using temporary inhibition of restricted brain regions by Alzheimer's disease, Behav. Pharmacol., 3 (1992) p. 31. drugs, (e.g., refs. 16, 17, 32, 35, 37, 48, 66) or correla- 16 Ferreira, M.B.C., da Silva, R.C., Medina, J.H. and Izquierdo, I., tional studies measuring electrical activity concomi- Late post-training memory processing by entorhinal cortex: role of NMDA and GABA-A receptors, Pharmacol. Biochem. Behav., tantly with behavior (e.g., ref. 55). The so-called 'ex- 41 (1992) 767-771. tended amygdala', which includes cell groups 17 Ferreira, M.B.C., Wolfman, C., Walz, R., da Silva, R.C., Zanatta, ontogenetically and neuroanatomically related to the M.S., Medina, J.H. and Izquierdo, 1., NMDA-dependent, amygdala in the striatal complex 22 seems particularly GABA-A-sensitive role of the entorhinal cortex in post-training worthy of investigation as part of a possible vicarious memory processing, Behav. Pharmacol., 3 (1992) 387-391. 18 Francis, P.T., Procter, A.W. and Bowen, D.M., A glycine site or alternative circuit in memory processing. as therapeutic target, Ann. NY Acad. Sci., 640 (1991) 184- 188. 19 Gray, J., The Neuropsychology of Anxiety: an Enquiry inta the Sep- tohippocampal System, Oxford University Press, Oxford, 1982, REFERENCES 548 pp. 20 Green J.D. and Machne, X., Unit activity of rabbit hippocampus, 1 Asanuma, H. and Keller, A., Neurobiological basis of motor Am. J. Physiol., 181 (1955) 219-224. learning and memory, Concepts Neurosci., 2 (1991) 1-30. 21 Growdon, J.H. and Cole, D.G., Clinical features of Alzheimer's 2 Bachevalier, J., Ontogenetic development of habit and memory disease, Behav. Pharmacol. 3 (1992) p. 36. formation in primates, Ann. NYAcad. Sci., 608 (1990) 457-477. 22 Heimer, L. and Alheid, G.F., Perspectives on basal forebrain 3 Breen, R.A. and McGaugh, J.L., Facilitation of maze learning organization and psychotic disorders, Behav. Pharmacol., 3 (1992) with post-trial injection of picrotoxin, J. Comp. Physiol. Psychol., p. 20. 54 (1961) 498-501. 23 Hyman, B.T., Van Hoesen, G.W. and Damasio, A.R., Memory- 4 Brioni, J.D., Nagahara, A. and McGaugh, J.L., Involvement of related neural systems in Alzheimer's disease: an anatomic study, the amygdala GABAergic system in the modulation of memory Neurology, 40 (1990) 1721-w1730. storage, Brain Res., 47 (1989) 105-112. 24 lntroini-Collison, I.B. and McGaugh, J.L., Interaction of hor- 5 Cahill, L. and McGaugh, J.L., Amygdaloid complex lesions dif- mones and neurotransmitter systems in the modulation of me- ferentially affect retention of tasks using appetitive and aversive mory storage. In R.C.A. Fredrickson, J.L. McGaugh and D.L. reinforcement, Behav. Neurosci., 104 (1990) 532-543. 97

Felten (Eds.), Peripheral Signaling of the Brain, Hogrefe and 44 Lynch, G.S., What memories are made of, The Sciences, 215 Huber, Toronto, 1990, pp. 275-301. (1985) 38-43. 25 Izquierdo, I., Different forms of post-training memory processing, 45 Machne, X. and Segundo, J.P., Unitary responses to afferent Behav. Neural Biol., 51 (1989) 171-202. volleys in amygdaloid complex, J. Neurophysiol., 19 (1956) 232- 26 Izquierdo, I., Mechanism of the amnestic action of scopolamine, 240. Trends Pharmacol. Sci., 10 (1989) 175-177. 46 McGaugh, J.L., Modulation of memory storage processes. In 27 Izquierdo, I., Role of NMDA receptors in memory, Trends Phar- P.R. Solomon, G.R. Goethals, C.M. Kelley and B.R. Stephens macol. Sci., 12 (1991) 128-129. (Eds.), Perspectives of Memory Research, Springer, New York, 28 Izquierdo, I., Opioids and memory. In T.W. Stone (Ed.), Aspects 1988, pp. 33-64. of Synaptic Transmission, Taylor and Francis, London, 1991, 47 McGaugh, J.L., Castellano, C. and Brioni, J.D., Picrotoxin en- pp. 162-178. hances latent extinction of conditioned fear, Behav. Neurosci., 104 29 Izquierdo, I., The neurobiology of memory consolidation, Neu- (1990) 264-267. rosciences, 18 (1992) 1-11. 48 Martin, J.H., Autoradiographic estimation of the extent of re- 30 Izquierdo, I., Acetylcholine and noradrenaline modulate long- versible inactivation produced by microinjection of lidocaine and term potentiation, Ciencia Cult., 44 (1992) 12-13. muscimol in the rat, Neurosci. Lett., 127 (1991) 160-164. 31 Izquierdo, I., Endogenous benzodiazepine-like molecules and 49 Medina, J.H., Pena, C., Piva, M., Wolfman, C., Stein, M.L. de, memory processes. In I. Izquierdo and J.H. Medina (Eds.), En- Wasowski, C., Da Cunha, C., Izquierdo, I. and Paladini, A.C., dogenous Benzodiazepines: Structure, Distribution and Function, Benzodiazepines in the brain. Their origin and possible biologi- Ellis Horwood, Chichester, 1993, pp. 65-80. cal roles, Mol. Neurobiol., 6 (1992) 1-10. 32 Izquierdo, I., Da Cunha, C., Rosat, R., Jerusalinsky, D., Ferreira, 50 Milner, T.A. and Bacon, C.E., GABAergic neurons in the rat M.B.C. and Medina, J.H., Neurotransmitter receptors involved hippocampal formation: ultrastructure and synaptic relationship in memory processing by the amygdala, medial septum and hip- with catecholaminergic terminals, J. Neurosci., 9 (1989) 3410- pocampus of rats, Behav. Neural Biol., 58 (1992) 16-26. 3427. 33 Izquierdo, I. and Medina, J.H., GABA-A receptor modulation of 51 Miserendino, M.J.D., Sananes, C.B., Melia, K.R. and Davis, M., memory: The role of endogenous benzodiazepines, Trends Phar- Blocking of acquisition but not expression of conditioned fear- macol. Sci., 12 (1991) 260-265. potentiated startle by NMDA antagonists in the amygdala, Na- 34 Izquierdo, I., Medina, J.H., Da Cunha, C., Wolfman, C., Jeru- ture, 345 (1990) 716-718. salinsky, D. and Ferreira, M.B.C., Memory modulation by brain 52 Mizumori, S.J.Y., Perez, G.M., Alvarado, M.C., Barnes, C.A. benzodiazepines, Braz. J. Med. Biol. Res., 24 (1991) 865-881. and McNaughton, B.L., Reversible inactivation of the medial 35 Izquierdo, I., Medina, J.H., Jerusalinsky, D. and Da Cunha, C., septum differentially affects two forms of learning in rats, Brain Post-training memory processing in amygdala, septum and hip- Res., 528 (1990) 12-20. pocampus: Role of benzodiazepine/GABA-Areceptors and their 53 Monahan, J.B., Handelmann, G.E., Hood, W.F. and Cordi, A.A., interaction with other neurotransmitter systems, Rev. Neurosci., 3 D-Cycloserine, a positive modulator of the NMDA receptor, en- (1992) 11-23. hances learning and memory in the rat, Pharmacol. Biochem. 36 Jellinger, K., Braak, H., Braak, E. and Fisher, P., Alzheimer Behav., 34 (1989) 649-665. lesions in the entorhinal region and isocortex in Parkinson's 54 Morris, R.G.M., The role of NMDA receptors in certain kinds and Alzheimer's disease, Ann. NY Acad. Sci., 640 (1991) 203- of learning and memory. In L.R. Squire and E. Lindenlaub 209. (Eds.), The Biology of Memory, Schattauer, Stuttgart, 1990, 37 Jerusalinsky, D., Ferreira, M.B.C., Walz, R., Da Silva, R.C., pp. 299-316. Bianchin, M., Ruschel, A., Median, J.H. and Izquierdo, I., Am- 55 Morris, R.G.M., Anderson, E., Lynch, G.S. and Baudry, M., nesia by infusion of glutamate receptor blockers into the amygdala, Selective impairment of learning blockade of long-term potentia- hippocampus and entorhinal cortex, Behav. NeuralBiol., 58 (1992) tion by an N-methyl-D-aspartate receptor antagonist, Nature, 319 76-80. (1986) 774-776. 38 Jones, R.W., Wesnes, K.A. and Kirby, J., Effects of NMDA 56 Muller, D., Arai, A. and Lynch, G.S., Factors governing the modulation in scopolamine dementia, Ann. NY Acad. Sci., 640 potentiation of NMDA receptor-mediated responses in hippoc- (1991) 241-244. ampus, Hippocampus, 2 (1992) 29-38. 39 Leng, X., McGrann, J.V., Quillfeldt, J.A., Shaw, G.L. and 57 Procter, A.W., Stratmann, G.C., Francis, P.T., Lowe, S.L., Ber- Shenoy, K.V., Learning and memory processes and the modu- tolucci, P.H.F. and Bowen, D.M., Characterization of the Gly- larity of the brain. In J. Delacour (Ed.), Neural Bases of Learn- cine modulatory site of the N-methyl-D-aspartate receptor iono- ing and Memory, World Scientific, Singapore, in press. phore complex in human brain, J. Neurochem., 56 (1991) 299- 40 Leranth, C. and Frotscher, M., Organization of the septal region 310. in the rat brain: cholinergic-GABAergic interconnections and the 58 Reeves, T.M. and Smith, D.C., Reinervation of the dentate gyrus termination of hippocampo-septal fibers, J. Comp. Neurol., 289 and recovery of alternation behavior following entorhinal cortex (1989) 304-314. lesions, Behav. Neurosci., 101 (1987) 179-186. 41 Liang, K.C., McGaugh, J.L. and Yao, H-Y., Involvement of 59 Robinson Jr., G.S., Crooks Jr., G.B., Shinkman, P.G. and Gal- amygdala pathways in the influence of post-training intra- lagher, M., Behavioral effects of MK-801 mimic deficits associ- amygdala norepinephrine and peripheral epinephrine in memory ated with hippocampal damage, Psychobiology, 17 (1989) 156- storage, Brain Res., 508 (1990) 225-233. 164. 42 Lin, Y. and Phillis, J.W., Muscarinic agonist-mediated induction 60 Skelton, R.W. and McNamara, R.K., Bilateral knife cuts to the of long-term potentiation in rat cerebral cortex, Brain Res. 551 perforant path disrupt spatial learning in the Morris water maze, (1991) 342-347. Hippocampus, 2 (1992) 73-80. 43 Little, W.A., The evolution of non-Newtonian views of brain 61 Spear, N.E. and Mueller, C.W., Consolidation as a function of function, Concepts Neurosci., 1 (1990) 149-164. retrieval. In H. Weingarten and E.S. Parker (Eds.), Memory con- 98

solidation - Psychobiology of Cognition, Lawrence Erlbaum, Hills- 68 White, N.M. and Viaud, M., Localized intracaudate dopaminc dale, 1984, pp. 111-147. D2 receptor activation during the post-training period improves 62 Squire, L.R., Memory and Bra#t, Oxford University Press, Ox- memory for visual or olfactory conditioned emotional responses ford, 1987, 315 pp. in rats, Behav. Neural Biol., 55 (1991) 255-269. 63 Thompson, R.F., Entorhinal-subicular lesions: amnestic effects 69 Witter, M.P., Groenewegen, H.J., Lopes da Silva. F'.H. and Lo- on an assortment of learned responses in the white rat, Bull. hman, A.H.M.. Functional organization of the extrinsic and in- Psvchol. Soc., 8 (1976) 433-434. trinsic circuitry of the parahippocampal region, Pro e. Neurohiol. 64 Thompson, R.F., Are memory traces localized or distributed? 33 (1989) 161-253. Neuropsychologia, 29 (1991) 571-582. 70 Wolfman, C., Da Cunha, C., Jerusalinsky, D., Levi de Stein, M.. 65 Walker, D.L. and Gold, P.E., Effects of the novel NMDA an- Viola, H., lzquierdo, I. and Medina, J.H., Habituation and in- tagonist, NPC 12626, on long-term potentiation, learning and hibitory avoidance training alter brain regional levels of benzo- memory, Brain Res. 549 (1991) 213-221. diazepine-like molecules and are affected by intracerebral fluma- 66 Walz, R., Da Silva, R.C., Bueno e Silva, M., Medina, J.H. and zenil microinjection, Brain Res. 548 (1991) 74-80. Izquierdo, I., Post-training infusion &glutamate receptor antago- 71 Zola-Morgan, S., Squire, L.R., Amaral, D.G. and Suzuki, W. nists into the medial septum of rats causes amnesia, Ciencia Cult., Lesions of perirhinal and parahippocampal cortex that spare the 44 (1992) 339-341. amygdala and hippocampal formation produce severe memory 67 Warrington, E.K. and Weiskrantz, L., Amnesia: consolidation or impairment, J. Neurosci., 9 (1989) 4355-4370. retrieval? Nature, 228 (1970) 628-630.