Contextual modulation of conditioned responses: role of the ventral subiculum and nucleus accumbens

The performance of conditioned responses (CRs) is diminished when trained subjects are tested in a novel context. This study tested the hypothesis that the flow of contextual information along the disynaptic "ESA" (entorhinal cortex-ventral subiculum-nucleus accumbens) pathway is responsible for context-related modulation of CRs. Rabbits received electrolytic or sham lesions of the ventral subiculum followed by discriminative avoidance conditioning and counterbalanced extinction sessions in the original training context, a novel context, and the original training context with a novel cue. Neuronal activity was recorded simultaneously in the nucleus accumbens, cingulate cortex, and basolateral amygdala. The lesions did not affect the acquisition of avoidance behavior or prevent the reduction of CRs in response to a novel cue. However, the lesions did reduce CR incidence during extinction, and they did eliminate a further novel-context-induced CR reduction found in controls. In addition, lesions disrupted context-dependent neuronal responses in the nucleus accumbens but not in the cingulate cortex or amygdala. These findings are interpreted as supportive of the hypothesis that the ESA pathway mediates contextual modulation of CRs during extinction.     

Neuronal encoding of conditional stimulus duration in the cingulate cortex and the limbic thalamus of rabbits

Neuronal activity in cingulate cortex was recorded during discriminative active avoidance conditioning of rabbits. In one subpopulation of neurons, brief (200 and 500 ms) conditional stimuli (CSs) elicited greater average cingulate cortical training-induced neuronal discharges during conditioned response acquisition than did a long (5,000 ms) CS, and the amount of neuronal discrimination between CS+ and CS- was greater in response to the brief CSs than to the long CS. Neurons in a different subpopulation did not encode CS duration per se but were sensitive to the novelty of the CS duration. Medial dorsal and anteroventral thalamic neurons were suppressed by novel CS durations that activated novelty-sensitive neurons in related cingulate cortical areas. These results are discussed in relation to a theoretical model of the neural mediation of avoidance conditioning.     

Hippocampal control of cingulate cortical and anterior thalamic information processing during learning in rabbits

Summary Past studies of the neural determinants of discriminative avoidance conditioning in rabbits have fostered a theoretical model that describes the interactive functioning of the cingulate cortex (Brodmann's Areas 24 and 29), the anterior ventral and medial dorsal thalamic nuclei (AVN and MDN) and the hippocampus. Here we test hypotheses of the model concerning the influence of the hippocampus on cortical and thalamic information processing. The rabbits learned to perform locomotory conditioned responses (CRs) in an activity wheel in response to an acoustic (pure tone) positive conditional stimulus (CS+). A shock unconditional stimulus (US) was given 5 s after CS+ onset, but locomotion during the CS+ -US interval prevented the US. The rabbits also learned to ignore a second tone (a negative conditional stimulus, CS-) of different auditory frequency than the CS+, that did not predict the US. Multi-unit activity and intracranial macropotentials were recorded in the cingulate cortex and the AVN uring acquisition, overtraining, extinction, reacquisition and reversal training. Data were obtained in intact rabbits and in rabbits with bilateral lesions of the subicular complex, the origin of projections of the hippocampal formation to the cingulate cortex and AVN. In addition, the activity in the AVN was recorded in a separate group of rabbits with posterior cingulate cortical (Area 29) lesions. Subicular and Area 29 lesions were associated with an enhancement of the training-induced CS+ elicited neuronal response in the AVN. The frequency of CRs was enhanced in animals with subicular lesions. CS elicited unit responses in the cingulate cortices were attenuated in rabbits with subicular lesions. Both of the lesions were associated with significantly increased amplitudes of the CS elicited average cortical and thalamic macropotentials. These results suggested the following conclusions: a) subiculocortical afferents provide an enabling influence that is essential for CS elicited excitation in the cingulate cortex; b) the cingulate cortical excitatory response in intact animals exerts a limiting influence on the activity in the AVN; c) the enhanced AVN neuronal response in rabbits with lesions is due to the absence of the limiting influence and it contributes to the increased CR frequency in those animals. It is hypothesized that the hippocampus via subiculocortical projections, governs the flow of CR-inducing thalamocortical excitatory volleys. This governance determines the timing of CR output. The results of hippocampal processing of contextual information acting through the subiculocortical projection determines the moment most appropriate for the CR.     

Anterior thalamic lesions and neuronal activity in the cingulate and retrosplenial cortices during discriminative avoidance behavior in rabbits

Bilateral electrolytic lesions of the anteroventral (AV) nucleus of the thalamus given after training impaired retention performance (extinction and reacquisition) of rabbits in a differential avoidance conditioning task. In addition, the lesions abolished the excitatory, discriminative multiple-unit discharges that had developed in the cingulate and retrosplenial cortices to the auditory conditional stimuli (CSs) during the course of behavioral acquisition, prior to the induction of the lesions. The excitatory discharges were supplanted in the subjects with lesions by CS-elicited reduction of neuronal firing to levels below the prestimulus baseline. Lesions given before training did not disrupt behavioral acquisition, but they did eliminate the excitatory tone-elicited neuronal discharges that normally occur in the cortex before and during training. The CS-elicited reduction of neuronal firing did not occur at the beginning of training in the subjects given lesions before training, but it developed during the course of training. The lesions did not eliminate the excitatory and discriminative neuronal activity of the prefrontal cortex. These results demonstrate that excitatory and discriminative neuronal discharges in the cingulate and retrosplenial cortices are critically dependent on the connections of these areas with the anterior thalamic nuclei. Also the lesion-induced disruption of performance during extinction and reacquisition but not during original learning confirms a prediction from past electrophysiological studies, that the AV thalamic nucleus is involved in the mediation of the maintenance and retention of the conditioned avoidance behavior, but not in its original acquisition.     

Cingulate cortical coding of context-dependent latent inhibition

Neuronal activity of the auditory thalamus, amygdala, cingulate cortex, and substantia nigra was recorded during the administration of a behavioral test for latent inhibition (LI) or the retardation of behavioral conditioning because of preexposure of the conditional stimulus (CS). Following CS preexposure, both the preexposed CS and a control CS predicted avoidable footshock. LI occurred as significantly fewer avoidance conditioned avoidance responses after the preexposed CS than after the control CS. Attenuation of neuronal responses to the preexposed CS, or neural LI, occurred in all monitored areas. One group of subjects (Oryctolagus cuniculus) then received context extinction, and additional groups experienced novel context exposure or handling. Context extinction enhanced behavioral responding to the preexposed CS, eliminating LI. Context extinction also eliminated cingulate cortical neural LI by enhancing posterior cingulate cortical responses to the preexposed CS and attenuating anterior cingulate cortical responses to the control CS. Present and past results are interpreted to indicate that LI is (a) a failure of response retrieval and/or expression mediated by interfering CS-context associations and (b) a product of interactions of the posterior cingulate cortex and the hippocampus.     

Anterior and medial thalamic lesions,discriminative avoidance learning,and cingulate cortical neuronal activity in rabbits

Summary Four groups of male albino rabbits were trained to perform a conditioned response (CR, stepping in an activity wheel) to an acoustic (pure tone) conditional stimulus (CS+). A 1.5–2.0 mA shock unconditional stimulus (US) delivered through the grid floor of the wheel was administered 5 s after CS + onset, but stepping during the CS-US interval prevented the US. The rabbits were also trained to ignore a second tone (a negative conditional stimulus or CS-) of different auditory frequency than the CS+, that was presented in an irregular order on half of the conditioning trials but never followed by the US. One group had bilateral electrolytic lesions in the medial dorsal (MD) thalamic nucleus, a second group had combined bilateral lesions in the MD and the anterior thalamic nuclei, and a third group had no lesions. The fourth group was composed of rabbits with combined lesions that resulted in only partial damage in the anterior and MD nuclei. In all rabbits, multi-unit activity and field potentials were recorded from the cingulate cortical projection targets of the MD and anterior nuclei. The average rate of acquisition in rabbits with MD and partial lesions was not significantly different from that in controls, but the asymptotic performance in rabbits with lesions was significantly impaired, relative to that in controls. None of the rabbits that had the combined MD and anterior thalamic lesions reached the acquisition criterion. The average proportion of trials in which these rabbits performed avoidance responses during their final training sessions was 0.3, compared to 0.8 in controls. The unconditioned response was not significantly affected by the lesions, nor was there any indication that the lesions impaired the sensory processing of the CSs. These results and the massive traininginduced neuronal discharges shown in past studies to occur in the limbic thalamic neurons indicate that these neurons are importantly involved in the circuitry that mediates discriminative avoidance conditioning in rabbits. The training-induced neuronal activity in cingulate cortex was dramatically attenuated in rabbits with lesions. Differences in the degree of this attenuation between lesion conditions and with respect to training stages were discussed in relation to a theoretical working model of limbic thalamic and cingulate cortical associative functions.     

Renewal of an extinguished instrumental response: neural correlates and the role of D1 dopamine receptors

Contexts play an important role in controlling the expression of extinguished behaviors. We used an ABA renewal design to study the neural correlates, and role of D1 dopamine receptors, in contextual control over extinguished instrumental responding. Rats were trained to respond for a sucrose reward in one context (A). Responding was then extinguished in the same (A) or different (B) context. Rats were tested for responding in the original training context (A). Return to the original training context after extinction (group ABA) was associated with a return of responding. Three distinct patterns of Fos induction were detected on test: 1) ABA renewal was associated with selective increases in c-Fos protein induction in basolateral amygdala, ventral accumbens shell, and lateral hypothalamus (but not in orexin- or melanin-concentrating hormone (MCH)-hypothalamic neurons); 2) being placed in the same context as extinction training (AAA or ABB) was associated with a selective decrease in c-Fos induction in rostral agranular insular cortex; 3) being placed in any context on test was associated with the up-regulation of c-Fos induction in anterior cingulate, dorsomedial accumbens shell, accumbens core, lateral septum, and substantia nigra. The return of responding in ABA renewal was prevented by pre-treatment with the D1 dopamine receptor antagonist SCH23390 (10 microg/kg; s.c.). SCH23390 also suppressed basal and renewal-associated c-Fos protein induction throughout accumbens, and, selectively suppressed renewal-associated c-Fos induction in lateral hypothalamus. These results suggest that renewal of extinguished responding for a sucrose reward depends on a distributed neural circuit involving basolateral amygdala, ventral accumbens shell, and lateral hypothalamus. D1 dopamine receptors within this circuit are essential for renewal. The results also suggest that rostral agranular insular cortex may play an important role in suppressing reward-seeking after extinction training.     

Enhanced metabolic capacity of the frontal cerebral cortex after Pavlovian conditioning

While Pavlovian conditioning alters stimulus-evoked metabolic activity in the cerebral cortex, less is known about the effects of Pavlovian conditioning on neuronal metabolic capacity. Pavlovian conditioning may increase prefrontal cortical metabolic capacity, as suggested by evidence of changes in cortical synaptic strengths, and evidence for a shift in memory initially processed in subcortical regions to more distributed prefrontal cortical circuits. Quantitative cytochrome oxidase histochemistry was used to measure cumulative changes in brain metabolic capacity associated with both cued and contextual Pavlovian conditioning in rats. The cued conditioned group received tone-foot-shock pairings to elicit a conditioned freezing response to the tone conditioned stimulus, while the contextually conditioned group received pseudorandom tone-foot-shock pairings in an excitatory context. Untrained control group was handled daily, but did not receive any tone presentations or foot shocks. The cued conditioned group had higher cytochrome oxidase activity in the infralimbic and anterior cingulate cortex, and lower cytochrome oxidase activity in dorsal hippocampus than the other two groups. A significant increase in cytochrome oxidase activity was found in anterior cortical areas (medial, dorsal and lateral frontal cortex; agranular insular cortex; lateral and medial orbital cortex and prelimbic cortex) in both conditioned groups, as compared with the untrained control group. In addition, no differences in cytochrome oxidase activity in the somatosensory regions and the amygdala were detected among all groups. The findings indicate that cued and contextual Pavlovian conditioning induces sustained increases in frontal cortical neuronal metabolic demand resulting in regional enhancement in the metabolic capacity of anterior cortical regions. Enhanced metabolic capacity of these anterior cortical areas after Pavlovian conditioning suggests that the frontal cortex may play a role in the retention and regulation of learned associations.     

The neural correlates and role of D1 dopamine receptors in renewal of extinguished alcohol-seeking

We used an ABA renewal design to study the neural correlates, and role of D1 dopamine receptors, in contextual control over extinguished alcohol-seeking. Rats were trained to respond for 4% beer in one context (A), extinguished in a different (B) context, and then tested for responding in the original training context (A) or the extinction context (B). ABA renewal was mediated by D1 dopamine receptors because it was prevented by SCH23390. ABA renewal of alcohol-seeking was associated with selective increases in c-Fos protein induction in basolateral amygdala, ventral accumbens shell, and lateral hypothalamus (renewal-associated Fos). By contrast, being tested was associated with increased c-Fos induction in anterior cingulate, prelimbic and infralimbic cortex, rostral agranular insula, dorsomedial accumbens shell, and accumbens core (test-associated Fos). Renewal-associated Fos in ventral accumbens shell and lateral hypothalamus, but not basolateral amygdala, was D1 dopamine receptor dependent. Double immunofluorescence showed that renewal-associated Fos was expressed in orexin-negative lateral hypothalamic neurons. However, c-Fos induction in either lateral hypothalamic orexin-negative or orexin-positive neurons was positively and significantly correlated with alcohol-seeking. Test-associated c-Fos induction was observed in orexin-positive perifornical neurons. In both regions, c-Fos expression was dependent on D1 dopamine receptors. These results suggest that renewal of extinguished alcohol-seeking depends on a distributed neural circuit involving basolateral amygdala, ventral accumbens shell, and lateral hypothalamus that involves D1 dopamine receptors. Comparison with our previous results [Hamlin AS, Blatchford KE, McNally GP (2006) Renewal of an extinguished instrumental response: Neural correlates and the role of D1 dopamine receptors. Neuroscience 143:25-38] permits identification of similarities and differences in the correlates of renewal of extinguished drug- and natural-reward seeking.     

Entorhinal cortex stimulation modulates amygdala and piriform cortex responses to olfactory bulb inputs in the rat

The rodent olfactory bulb sends direct projections to the piriform cortex and to two structures intimately implicated in memory processes, the entorhinal cortex and the amygdala. The piriform cortex has monosynaptic projections with the amygdala and the piriform cortex and is therefore in a position to modulate olfactory input either directly in the piriform cortex, or via the amygdala. In order to investigate this hypothesis, field potential signals induced in anesthetized rats by electrical stimulation of the olfactory bulb or the entorhinal cortex were recorded simultaneously in the piriform cortex (anterior part and posterior part) and the amygdala (basolateral nucleus and cortical nucleus). Single-site paired-pulse stimulation was used to assess the time courses of short-term inhibition and facilitation in each recording site in response to electrical stimulation of the olfactory bulb and entorhinal cortex. Paired-pulse stimulation of the olfactory bulb induced homosynaptic inhibition for short interpulse interpulse intervals (20-30 ms) in all the recording sites, with a significantly lower degree of inhibition in the anterior piriform cortex than in the other structures. At longer intervals (40-80 ms), paired-pulse facilitation was observed in all the structures. Paired-pulse stimulation of the entorhinal cortex mainly resulted in inhibition for the shortest interval duration (20 ms) in anterior piriform cortex, posterior piriform cortex and amygdala basolateral but not cortical nucleus. Double-site paired-pulse stimulation was then applied to determine if stimulation of the entorhinal cortex can modulate responses to olfactory bulb stimulation. For short interpulse intervals (20 ms) heterosynaptic inhibition was observed in anterior piriform cortex, posterior piriform cortex and amygdala basolateral but not cortical nucleus. The level of inhibition was greater in the basolateral nucleus than in the other structures. Taken together these data suggest that the entorhinal cortex exerts a main inhibitory effect on the olfactory input via the amygdala basolateral nucleus and to a lesser extent the piriform cortex. The potential role of these effects on the processing of olfactory information is discussed.     

Effects of pre- or post-training entorhinal cortex AP5 injection on fear conditioning

Fear conditioning is one of the most studied paradigms to assess the neural basis of emotional memory. The circuitry involves NMDA receptor activation in the amygdala and, in the case of contextual conditioning, in the hippocampus. Entorhinal cortex is one of the major input/output structures to the hippocampus and also projects to the amygdala, both through glutamatergic transmission. Other learning tasks involving hippocampus and amygdala, such as inhibitory avoidance, require entorhinal cortex during acquisition and consolidation. However, the involvement of NMDA receptors mediated transmission in entorhinal cortex in fear conditioning acquisition and consolidation is not clear. To investigate that issue, rats were trained in fear conditioning to both contextual and tone conditioned stimulus. Immediately before, immediately, 30 or 90 min after training they received NMDA antagonist AP5 or saline injections bilaterally in the entorhinal cortex (AP-6.8 mm, L +/-5.0 mm DV-6.8 mm). Contextual fear conditioning was measured 24 h after training, and tone fear conditioning 48 h after training. AP5 injections selectively impaired contextual fear conditioning only when injected pre-training. Post-training injections had no effect. These findings suggest that entorhinal cortex NMDA receptors are necessary for acquisition, but not for consolidation, of contextual fear conditioning. On the other hand, both acquisition and consolidation of tone fear conditioning seem to be independent of NMDA receptors in the entorhinal cortex.     

Complementary roles for the amygdala and hippocampus in aversive conditioning to explicit and contextual cues.

Experiment 1 investigated the effects of catecholaminergic deafferentation or cell body lesions of the amygdala on fear conditioning to explicit and contextual cues. Bilateral infusions of quinolinic acid mainly damaged neurons within the basolateral region of the amygdala. 6-Hydroxydopamine infusions at the same coordinates resulted in an 86% depletion of noradrenaline and a 63% depletion of dopamine from the amygdala, but had no effect on the concentration of 5-hydroxytryptamine. After recovery from surgery, lesioned rats and controls were exposed to pairings of an auditory (clicker) conditioned stimulus and (foot shock) unconditioned stimulus in a distinctive environment. During testing, rats with both 6-hydroxydopamine and cell body lesions showed severely impaired conditioning to explicit cues, compared with controls, indicated by their reduced suppression of drinking when the conditioned stimulus was introduced into a separate, lick-operant chamber. Neither lesion affected fear conditioning to contextual cues, measured as preference for a "safe" environment over the one in which they were shocked. In Experiment 2, rats received bilateral, ibotenic acid-induced lesions of the hippocampal formation. Lesioned rats and controls were again tested for aversive conditioning to explicit and contextual cues. Rats with cell body lesions of the hippocampus showed normal suppression of drinking in the presence of the conditioned stimulus, but were severely impaired in choosing the safe environment based on contextual cues alone. These results suggest a double dissociation of the effects of amygdala and hippocampal damage on fear conditioning to explicit and contextual cues.     

Direct projections of non-pyramidal neurons of Ammon's horn to the amygdala and the entorhinal cortex.

When WGA-HRP (wheat germ agglutinin-horseradish peroxidase conjugate) was injected into the amygdala (lateral and basolateral amygdaloid nuclei) or entorhinal cortex of the cat, a number of nonpyramidal neurons in Ammon's horn were retrogradely labeled. The results indicate that some non-pyramidal neurons in Ammon's horn send projection fibers to the amygdala and entorhinal cortex.     

Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning.

The contribution of the amygdala and hippocampus to the acquisition of conditioned fear responses to a cue (a tone paired with footshock) and to context (background stimuli continuously present in the apparatus in which tone-shock pairings occurred) was examined in rats. In unoperated controls, responses to the cue conditioned faster and were more resistant to extinction than were responses to contextual stimuli. Lesions of the amygdala interfered with the conditioning of fear responses to both the cue and the context, whereas lesions of the hippocampus interfered with conditioning to the context but not to the cue. The amygdala is thus involved in the conditioning of fear responses to simple, modality-specific conditioned stimuli as well as to complex, polymodal stimuli, whereas the hippocampus is only involved in fear conditioning situations involving complex, polymodal events. These findings suggest an associative role for the amygdala and a sensory relay role for the hippocampus in fear conditioning.     

Infralimbic cortex activation increases c-Fos expression in intercalated neurons of the amygdala

Recently, it was reported that stimulation of the infralimbic cortex produces a feedforward inhibition of central amygdala neurons. The interest of this observation comes from the fact that the central nucleus is the main output station of the amygdala for conditioned fear responses and evidence that the infralimbic cortex plays a critical role in the extinction of conditioned fear. However, the identity of the neurons mediating this infralimbic-evoked inhibition of the central nucleus remains unknown. Likely candidates are intercalated amygdala neurons. Indeed, these cells receive glutamatergic afferents from the infralimbic cortex, use GABA as a transmitter, and project to the central amygdala. Thus, the present study was undertaken to test whether, in adult rats, the infralimbic cortex can affect the activity of intercalated neurons. To this end, disinhibition of the infralimbic cortex was induced by local infusion of the non-competitive GABA-A receptor antagonist picrotoxin. Subsequently, neuronal activation was determined bilaterally within the amygdala using induction of the immediate early gene Fos. Infralimbic disinhibition produced a significant increase in the number of Fos-immunoreactive intercalated cells bilaterally whereas no change was detected in the central nucleus. In the basolateral amygdaloid complex, increases in the number of Fos-immunoreactive cells only reached significance in the contralateral lateral nucleus. These results suggest that glutamatergic inputs from the infralimbic cortex directly activate intercalated neurons. Thus, our findings raise the possibility that the infralimbic cortex inhibits conditioned fear via the excitation of intercalated cells and the consequent inhibition of central amygdala neurons.     

Conditional stimulus probability and activity of hippocampal, cingulate cortical, and limbic thalamic neurons during avoidance conditioning in rabbits

Past studies of the neuronal correlates of avoidance conditioning in rabbits have led to a model of information flow among structures of the limbic system. A hypothesis of the model is that unexpected stimuli activate certain hippocampal and cingulate cortical neurons. This activity in turn suppresses or "limits" the firing of limbic thalamic neurons. This hypothesis is tested in relation to stimuli classified as unexpected or expected on the basis of their incidence or "probability." Multi-unit and field potential responses in the anterior and posterior cingulate cortices (AC and PC), the dentate gyrus (DG), and the anterior ventral (AV) and medial dorsal (MD) thalamic nuclei were recorded during the acquisition and performance of a locomotor conditioned response (CR). The CR, stepping in an activity wheel in response to a 0.5-s tone (CS+), prevented the occurrence of a shock US scheduled 5 s after CS+ onset. The rabbits also learned to ignore a different tone (CS-), not predictive of the US. Training was given daily (120 trials, 60 with each CS in an irregular sequence) until behavioral discrimination reached criterion. After criterion, asymmetric probability (AP) sessions were given, in which the CS+/CS- proportions were .2/.8 or .8/.2. The AP sessions were the same as conditioning sessions except for the probability manipulation. A significant discriminative response, i.e., a greater neuronal discharge to the CS+ than to the CS-, developed in all regions during behavioral acquisition. The unit response in the AP sessions was enhanced in all areas by rare presentation of the CS-, compared with the equal and frequent CS- conditions. Rare presentation of the CS+ enhanced the unit response in the cortical areas (AC, PC, and DG), but it suppressed the firing of limbic thalamic (AV and MD) neurons. These results were supportive of the model. Rare CS+ presentations did not alter AV and PC neuronal activity in rabbits with subicular lesions, a result suggesting that an intact hippocampus is essential for normal neuronal responses to stimulus probability in the cingulate cortex and limbic thalamus.     

Impairment of cingulothalamic learning-related neuronal coding in rabbits exposed to cocaine in utero: general and sex-specific effects

Neuronal activity was recorded in the cingulate cortex and the limbic thalamus in Dutch-belted rabbits (Oryctolagus cuniculus) exposed to cocaine (8 mg/kg/day i.v.) or saline in utero during acquisition and reversal learning of a discriminative avoidance response. Anterior cingulate cortical excitatory training-induced activity (TIA) was attenuated in cocaine-exposed female rabbits during acquisition and reversal learning, but only during reversal learning in male rabbits. Posterior cingulate cortical excitatory TIA was lessened in cocaine-exposed rabbits during acquisition, whereas discrimination between the positive and negative cues was enhanced. Neuronal firing was attenuated in the anterior ventral thalamus in cocaine-exposed rabbits during acquisition and reversal learning. Behavioral learning was normal in cocaine-exposed rabbits. Other data suggest that rabbits exposed to cocaine in utero exhibit a learning deficit when trained with nonsalient cues.     

Local norepinephrine depletion and learning-related neuronal activity in cingulate cortex and anterior thalamus of rabbits

Summary Multi-unit neuronal activity was recorded in posterior cingulate cortex (area 29) and the anterior ventral (AV) thalamic nucleus during discriminative instrumental avoidance learning wherein a response (stepping in an activity wheel) to a 0.5-s tone (CS⁺) prevented a foot-shock 5 s after CS⁺ onset. Presentations of a different tone (CS^-) on 50% of the conditioning trials in an irregular sequence with the CS⁺ did not predict shock and thus required no response. Two groups of rabbits received intracranial micro-injections of 6-hydroxydopamine (6-OHDA) to locally deplete the NE in area 29 or the AV nucleus. Vehicle was injected in the non-depleted area in each group and a third group received vehicle injections in both areas. Dopamine neurons in subjects that received 6-OHDA were protected by pre-treatment with GBR-12909. Neuronal data were collected during two pre-training sessions in response to the tones only and when the tones and shock were presented unpaired. Thalamically depleted rabbits made more, and cortically depleted rabbits made fewer, avoidance responses than controls during the early stages of behavioral acquisition, and cortically depleted rabbits made fewer responses than controls and thalamically depleted rabbits during extinction testing administered after the completion of acquisition. One effect of NE depletion on neuronal activity was entirely local: elimination of neuronal sensitization effects (enhanced discharges elicited by tones during the unpaired tone-shock pre-training treatment relative to pre-training with tones only). Other neuronal effects of NE depletion were system-wide, i.e., they occurred whether the depletion was cortical or thalamic. These were: attenuation of area 29 tone-elicited neuronal discharges and enhancement of AV thalamic discharges before and during training; elimination in area 29 of neuronal discrimination between CS⁺ and CS^-, induced in controls by CS⁺-shock pairings in the first conditioning session; induction of this neuronal discrimination, not present in controls, in the AV nucleus during the first conditioning session; attenuation of discharge enhancements elicited in controls by unexpected stimuli (presentation of auditory stimuli different in quality and incidence from the CS⁺). Excepting the noted losses at the outset of training, the results did not support an involvement of NE in the production of cingulate cortical or AV thalamic excitatory and discriminative training-induced neuronal activity. The system-wide alterations due to NE depletion implicated NE in the processing of unexpected events and in the production of dynamic neuronal patterns relevant to mnemonic retrieval. Several of the depletion-related neuronal changes were similar to the effects of hippocampal formation (subicular) lesions, suggesting that NE-dependent functions in area 29 and the AV nucleus are governed by hippocampal efferents, which may control the release of NE in these areas.     

Prefrontal cortical contributions during discriminative fear conditioning, extinction, and spontaneous recovery in rats

This research examined the roles played by the ventromedial orbital prefrontal cortex (OPFC) and the infralimbic/prelimbic prefrontal cortex (I/P PFC) during discriminative fear conditioning. The first experiment included nine rats with bilateral lesions to the I/P PFC, an additional nine with OPFC lesions, and eight sham lesion controls. Behavioural analysis was conducted using a discriminative fear conditioning to context task 10 days after surgery. Results indicate that lesions to ventromedial orbital prefrontal cortex result in generalized fear and impaired extinction. In contrast, infralimbic/prelimbic cortical lesioned animals exhibit appropriate fear response patterns and extinction, but show a specific impairment in spontaneous recovery. To ascertain why I/P PFC lesion rats did not exhibit spontaneous recovery, a second experiment was conducted. All procedures in the second experiment were identical to the first except a decay period was employed in place of extinction training. Results from the second experiment indicate that the difficulty retrieving the extinguished association is related to extinction processes and not decay. Taken together, these findings suggest that OPFC and I/P PFC have distinct roles in associative processes necessary for discriminative fear conditioning, extinction, and spontaneous recovery. These results further implicate OPFC and I/P PFC in the pathology underlying generalized anxiety disorder.     

Bidirectional synaptic plasticity in intercalated amygdala neurons and the extinction of conditioned fear responses

Classical fear conditioning is believed to result from potentiation of conditioned synaptic inputs in the basolateral amygdala. That is, the conditioned stimulus would excite more neurons in the central nucleus and, via their projections to the brainstem and hypothalamus, evoke fear responses. However, much data suggests that extinction of fear responses does not depend on the reversal of these changes but on a parallel NMDA-dependent learning that competes with the first one. Because they control impulse traffic from the basolateral amygdala to the central nucleus, GABAergic neurons of the intercalated cell masses are ideally located to implement this second learning. Consistent with this hypothesis, the present study shows that low- and high-frequency stimulation of basolateral afferents respectively induce long-term depression (LTD) and potentiation (LTP) of responses in intercalated cells. Moreover, induction of LTP and LTD is prevented by application of an NMDA antagonist. To determine how these activity-dependent changes are expressed, we tested whether LTD and LTP induction are associated with modifications in paired-pulse facilitation, an index of transmitter release probability. Only LTP induction was associated with a change in paired-pulse facilitation. Depotentiation of previously potentiated synapses did not revert the modification in paired pulse facilitation, suggesting that LTP is associated with presynaptic alterations, but that LTD and depotentiation depend on postsynaptic changes.Taken together, our results suggest that basolateral synapses onto intercalated neurons can express NMDA-dependent LTP and LTD, consistent with the possibility that intercalated neurons are a critical locus of plasticity for the extinction of conditioned fear responses. Ultimately, these plastic events may prevent conditioned amygdala responses from exciting neurons of the central nucleus, and thus from evoking conditioned fear responses.