The experiments reported in the present thesis investigated the effects of lesions to the anterior nuclei of the thalamus (ATN) on the acquisition of two spatial, working memory tasks performed on the eight-arm radial maze. In the task used in Experiment 1 and 2, the animals were required to discriminate and remember all the eight arms of the maze simultaneously. Lesions of the ATN produced impairments in the acquisition of this task, but the degree of impairment depended on the amount of damage within this region. In the task used in Experiment 3, the animals were required to discriminate and remember only two arms at once. Lesions of the ATN were shown to impair its acquisition even though performance was facilitated by the addition of visual intra-arms cues. These experiments suggest that the ATN may be involved in spatial learning and in the retention of non-specific information over time.

This thesis examined the involvement of the lateral, central and basolateral nuclei of the amygdala in both appetitive and aversive affective behavior. In Experiment I, using electrolytic lesions, it was found that damage to the lateral but not central or basolateral nuclei blocked a Conditioned Cue Preference (CCP) to food (Froot Loops) in rats that were not deprived of food. In Experiment II, also using electrolytic lesions, it was found that damage to the basolateral but not central or lateral nuclei blocked a Conditioned Cue Aversion (CCA) produced by a lithium chloride injection (42 mg/kg). In Experiment III results similar to those in Experiments I and II were obtained using axon-sparing NMDA lesions. The results of Experiments I-III demonstrate a double dissociation of affective memory with respect to the amygdala. The lateral nucleus of the amygdala mediated the memory of an appetitive affective experience and the basolateral nucleus mediated memory for an aversive affective experience. / In Experiment IV the contributions of appetitive and aversive affective states to a food CCP were examined. Lesions of the lateral but not the basolateral nucleus were found to attenuate but not completely eliminate a food CCP when the rats were food deprived in the Paired compartment and sated in the Unpaired compartment. Food deprivation alone produced a CCA and lesions of the basolateral but not the lateral nucleus blocked this effect. The possibility that both the appetitive and aversive behaviours are mediated through connections from the dopamine-reward centres in the ventral striatum is discussed.

The pedunculopontine tegmental nucleus (PPTg) has been proposed to be critical in drug and food reward. It is a major source of cholinergic inputs to the substantia nigra and ventral tegmental area, areas important in reward, and is believed to modulate activity of dopamine neurons through a cholinergic mechanism. The firing rate of cholinergic PPTg cells is regulated by muscarinic autoreceptors that can be blocked to increase cell firing. If the PPTg modulates dopamine neurons involved in reward, then a muscarinic antagonist microinjected into the PPTg should be rewarding. To test this hypothesis, bilateral microinjections of scopolamine methyl bromide (5 $ mu$g or 20 $ mu$g) or 0.9% saline were used as reward treatments in the conditioned place preference test. On the test day, rats from both doses of drug avoided the drug-paired chamber, which suggests that cholinergic PPTg cells are not involved in reward.

The peripheral and central mechanisms mediating the modulatory effects of glucose and acute stress in rats were investigated using two versions of an appetitive win-stay task. Post-training injections of glucose, but not fructose, enhanced retention on the closed maze task. Acquisition of this task was found to be impaired by lesions of the fimbria-fornix (FF). Further experiments showed that while the celiac ganglion and the FF normally participate in suppressing the memory-enhancing effect of an acute stressor, neither structure is involved in mediating the effect of glucose on memory. Post-training injections of glucose, but not fructose, enhanced retention on the open maze task. Although acquisition of this task was not affected by FF lesions, both the celiac ganglion and the FF participate in mediating the memory-enhancing effect of glucose. Together, the results suggest that the peripheral and central mechanisms studied here are both substance- and task-specific. The modulatory effects of different types of stress, and the issue of whether task acquisition and memory modulation are anatomically distinct are discussed.

The effect of post-training self-stimulation on associations formed during classical conditioning was studied. In the first three experiments, rats experienced tone-shock, tone-water and tone-light pairings, respectively. Control groups experienced the same situations except that the temporal relationship between the stimuli excluded the possibility of an association being formed. Following this training period, the experimental animals pressed a bar for electrical self-stimulation of the brain; control animals did not self-stimulate. Retention of the association between the training stimuli was evaluated by measuring conditioned suppression of drinking (experiments 1 and 3), or extinction of secondary reinforcement. The results showed that self-stimulation can retroactively and non-contingently improve memory for associations in various classical conditioning situations, independently of the presence of a natural reinforcer or of the responses they elicit. In another experiment, the effects of varying amounts of tone preexposure (latent inhibition) and of varying amounts of post-training self-stimulation on retention of a tone-shock association, measured by the amount of tone-induced suppression of drinking, were studied. Increasing amounts of tone pre-exposure produced decreased retention over all levels of self-stimulation, and increasing amounts of self-stimulation produced increased retention over all levels of pre-exposure. The interaction of these two factors suggests that they both acted on the strength of the tone-shock association: the pre-exposure acted proactively to weaken the association; the self-stimulation acted retroactively to strengthen it. The effect of delaying the post-training stimulation for various times after the tone-shock pairings was also investigated. Retention was improved by self-stimulation delayed for 15 or 30 minutes, but was not affected by self-stimulation delayed for an hour. These experiments suggest that direct activation

A series of experiments was carried out to clarify the effects of morphine (0.3 - 10 mg/kg) on intracranial self-stimulation (ICS) and to compare these with the effects of the stimulant amphetamine on this behavior. It was shown that the enhancement of ICS by morphine requires repeated drug exposure, is prevented by pre-exposure to the drug in a non-ICS context, is mimicked by administration of vehicle, and is not reliably reversed by naloxone. In contrast, facilitation of ICS by amphetamine was immediate and remained stable over repeated days of testing. It was concluded that ICS facilitation induced by morphine, but not by amphetamine, is largely the outcome of a learned association between the drug effect and the ICS procedure, and does not appear to be a direct, opiate receptor-mediated effect. Finally, 6-OHDA lesions of ventral tegmental dopamine neurons were shown to block the facilitation of ICS by morphine but not by amphetamine. These lesions were also shown to delay the development of tolerance to morphine-induced catalepsy.

This thesis explored whether interpretations of the reinforcing effect of stimulation trains used in the self-administration of brain-stimulation (SABS) paradigm were artifacts of the reinforcement schedule chosen or whether it represented a genuine attempt by animals to maintain optimal levels of reward. Results demonstrate that stimulation trains used in SABS are reinforcing and that animals regulate pulse frequency to optimize the level of reward. The thesis then explored whether pedunculopontine tegmental nucleus (PPTg) lesions blocked the acquisition or maintenance of SABS, and the acquisition of eight-arm radial maze learning. Results showed that lesions confined to the PPTg block acquisition and maintenance of SABS, suggesting that the PPTg mediates the positive reinforcing effects of BSR. Further, PPTg lesions blocked win-shift and win-stay radial maze learning. However, results indicate that animals were not impaired in "shifting" or "staying" behavior. It is speculated that PPTg lesions block the reinforcing effects of food, which produce inefficient performance on both memory tasks.

It has been proposed that analgesia in the formalin test is mediated through forebrain systems associated with reinforcement, whereas motor responses necessary for the expression of pain are organized at the level of the brainstem. Because it is located in the brainstem and connected with both limbic reward systems and motor structures, the pedunculopontine tegmental nucleus (PPTg) is a site where reward signals might influence the expression of pain. Experiment 1 confirmed that NMDA-induced lesions of the PPTg block the development of a conditioned place preference to morphine. Subsequently, morphine-induced analgesia was found to be reduced, but not eliminated. The reduction of reward was not significantly correlated with loss of choline acetyltransferase containing neurons in the PPTg. In Experiment 2, PPTg lesions did not affect morphine analgesia in drug naive animals, but produced motor abnormalities and blocked the morphine-induced depression of spontaneous motor activity and catalepsy.

The work reported here was aimed at extending previous findings suggesting that the medial and lateral substantia nigra (SN) are functionally different with regard to intracranial self-stimulation (ICSS) and circling. It was found that the effects of systemic D- and L-amphetamine on ICSS were different for medial and lateral SN sites. These differences were attributed to the presence of two subtypes of dopamine (DA) neurons in the SN. The circling results indicate that circling in opposite directions can be elicited from the medial and lateral parts of the nigrostriatal DA system from the same hemisphere suggesting that medial and lateral nigrostriatal DA have opposing roles in the expression of circling. It was further demonstrated that the superior colliculus is critical for the expression of medial SN derived circling and the midbrain reticular formation is critical for both medial and lateral SN derived circling.

This study examined the contributions of different brain areas to several of the behavior effects of amphetamine. The drug was micro-injected into each of six discreet brain sites in rats and the effects on behavior were examined. Amphetamine's rewarding effects were studied using the conditioned place preference (CPP) paradigm. Animals that had received injections into the nucleus accumbens showed a CPP, suggesting a rewarding effect of the drug. No effect was produced by injections into the medial frontal cortex, medial or lateral parts of the caudate nucleus, amygdala or the region around the area postrema. A conditioned taste aversion (CTA) towards a flavour that had been paired with the drug was produced by injections into the region around the area postrema, but not from the other sites. Anorexia and adipsia were both producted only by injections into the nucleus accumbens and amygdala. In the open field, increased activity was produced by intra-accumbens amphetamine injections, with smaller effects from the medial frontal cortex and medial caudate. Stereotyped behavior was not produced by any intra-cranial injection. The CPP, anorexia, adipsia and increase in activity that were produced by the intra-accumbens injections were interpreted as suggesting that the drug had stimulated approach behavior towards all stimuli, as if they had all become rewarding. The observation of a CPP from the accumbens and a CTA from the region around the area postrema suggests that the rewarding and apparently aversive effects of systemically injected amphetamine result from actions of the drug on different neuroanatomical substrates. Other hypotheses of the behavioral function of the neural substrates of the observed effects are presented.