Pharmacological inhibition of NMDA receptor activity by ketamine is accompanied by pyschotomimetic side-effects; however, the Alzheimers disease therapeutic memantine blocks NMDA receptor activity without debilitating side-effects. This dissertation provides electrophysiological and behavioral characterizations of these two NMDA receptor antagonists in an attempt to understand the unique therapeutic utility of memantine. The following work explores memantine and ketamine inhibition at NMDA receptors, their main site of action, with a focus on the mechanism of inhibition and receptor subtype selectivity in physiologically relevant conditions. This research shows NMDA receptors possess a second binding site at which memantine, but not ketamine, can inhibit activity. The research also shows the dramatic effect physiological concentrations of magnesium has on the ability of these drugs to inhibit NMDA receptor activity. Behavioral and cognitive effects of memantine and ketamine are also assessed and compared directly in rat. The effects of memantine and ketamine in rat were found to be similar at the low doses tested and more divergent as dose increased. Furthermore, memantines effects appeared to be more pronounced and longer-lasting than those of ketamine. These findings demonstrate the importance of considering the physiological environment in which a drug acts, as well as the principles of drug action, when examining the effects of a drug on central nervous system activity.

The uniquely human ability to rapidly learn novel tasks from instructions is extremely important in everyday life, and yet its evolutionary origin and basis in the brain remain mysteries. In order to address these gaps in scientific knowledge, comparative human-monkey studies were consulted to predict the human brain areas involved in rapid instructed task learning (RITL). These predictions were tested using functional MRI (fMRI), magnetoencephalography (MEG), and a novel cognitive paradigm developed to systematically investigate the neural basis of RITL for the first time. In accordance with cross-species neuroanatomical differences, anterior prefrontal cortex (aPFC), anterior temporal lobe (aTL), dorsolateral prefrontal cortex (DLPFC), and posterior parietal cortex (PPC) were found to be involved in RITL. DLPFC and PPC formed a network involved in loading individual task semantics into working memory, while aPFC and aTL formed a network involved in integrating semantics in preparation for task performance. Both networks supported novel task set formation, which occurred in a bottom-up manner (semantic loading, then integration), and practiced task set retrieval, which occurred in a top-down manner (integration retrieval, then semantic loading). These findings suggest that RITL relies upon semantic loading by DLPFC and PPC, but that aPFC and aTL support semantic integration both dynamically during RITL and from long-term memory after extensive practice. More broadly, the findings suggest RITL is enabled in humans via a combination of enhanced symbolic processing (language), enhanced working memory manipulation (aPFC), and enhanced integrated semantic representation (aTL). The present document begins with a broad overview of RITL and related topics, such as its relation to animal cognition, other forms of learning, and cognitive control. These topics support several novel hypotheses regarding RITL and its likely basis in the brain. The fMRI study is then presented, verifying many of the hypotheses developed in the previous section. The MEG study is reported next, clarifying many of the questions about timing and causality suggested by the fMRI results. Finally, a general discussion integrates the results from both studies, expanding conclusions with an overview of brain connectivity findings, cross-species differences, and the role of neural hierarchies in RITL and cognition generally.

Ih is an important contributor to the subthreshold membrane properties of various mammalian neurons, including interneurons. Here I characterize the properties of Ih in a subpopulation of hippocampal area CA3 interneurons with somata in stratum radiatum and stratum lacunosom moleculare. As shown in previous studies, Ih in these cells has sigmoidal voltage dependence of activation with kinetics characterized by two exponential components for both channel activation and deactivation. Interestingly, the activation and deactivation kinetics were most aptly described by distinct functions of voltage. These results were incorporated into a novel biophysical model of Ih that was applied in single compartment model simulations and dynamic clamp experiments. Finally, I assessed the functional consequences of Ih by examining the effects of this current on subthreshold temporal summation of mossy fiber EPSPs as well as frequency dependent neuronal responses. My results show that Ih decreases temporal summation of mossy fiber EPSPs but does not impart resonance in CA3 interneurons at potentials where Ih is active.

Our eyes are constantly moving yet our perception remains stable. Neurons in lateral intraparietal cortex (LIP) update spatial representations by remapping visual information at the time of an eye movement. In order for remapping to occur over a wide range of eye movements, neurons must have access to visual information from the entire visual scene. The forebrain commissures appear to be the primary pathway for the transfer of visual information across hemispheres but they are not necessary. If the forebrain commissures are transected, behavior dependent on accurate spatial updating is impaired, but recovers. In three sets of experiments we examined different mechanisms of spatial updating in split brain monkeys. First, we studied the relationship between neural activity in LIP and the behavior of the monkey. We found across the population a small but significant relationship between the activity in LIP and the performance of the split brain monkey on the double-step task. This result showed that information about the opposite visual field still reaches LIP, and this activity contributes to the overall behavior of the monkey. Second, we determined if LIP neurons in the split brain monkeys have bilateral receptive fields. One explanation for the observed across-hemifield remapping is that information from both visual fields are represented in a single hemisphere. We found no neurons in the split brain monkeys with ipsilateral representations. We concluded that there must be a subcortical source for the across-hemifield remapping observed in the split brain monkeys. Third, we examined the difference in spatial updating between intact and split brain monkeys in the superior colliculus (SC). In both the intermediate layers of the SC and LIP, neural activity is selectively reduced for the across-hemifield condition in split brain compared to intact animals. This suggest that remapping activity is passed from LIP to the intermediate layers of the SC. In contrast, remapping activity in the superfical layers did not differ between the intact and split brain monkeys. It may be that the superfical neurons contribute to recovered remapping activity observed in LIP.

In humans, cocaine use has long been associated with poor attentional control and decreased efficiency in goal-oriented behavior. Animal models of these stereotypic drug effects, however, have thus far failed to produce quantifiable data sets in part because of a lack of species differences and analysis techniques. Recent work (Hervey et al. 2006) has successfully quantified attentional lapses in disorders such as ADHD through the analysis of response time variations in simple tasks, but this analysis has yet to be applied to the drug abuse scenario. To determine the effects of chronic cocaine administration on response time variability, 14 rhesus macaque monkeys (8 cocaine administering and 6 performance-matched controls) were subjected to a 50 trial simple attention task. This task was performed W-F prior to cocaine self-administration sessions in the test group. Treatment groups were compared to both each other and to baseline task sessions recorded prior to beginning the administration paradigm. In addition to typical measures of variability, an ex-Gaussian response time analysis was performed to quantify the contribution of attentional lapses to overall variability. The cocaine-administering group had a significantly higher response time standard deviation than their pre-administration sessions (p<0.05). No difference was observed between pre- and post-administration sessions for the control group. When ex-Gaussian methods were applied to the response time datasets, no differences were observed between groups in the normal mean (mu), suggesting that the variability increase in the cocaine group was due to an increased skew in the right tail of the response time distribution. Indeed, the cocaine group showed a significant increase in the value of tau(exponential value representing the distribution tail magnitude) post-administration versus tau pre-administration (p<0.05). These data suggest that cocaine administration leads to increased behavioral variability in simple response time tasks, and that this variability increase is primarily due to the prevalence of abnormally long responses. Similar results have been demonstrated in clinical disorders such as ADHD, suggesting both the relevance of the primate model in studies of attentional processing and the possible similarity in affected brain regions or transmitter systems.

Alterations in GABA-related Transcripts in the Dorsolateral Prefrontal Cortex of Subjects with Schizophrenia Harvey M. Morris, Ph.D. University of Pittsburgh, 2009 Besides the financial burden upon society, families undergo a substantial emotional burden when presented with a loved one affected by schizohprenia. Elucidation of the pathophysiology underlying the core features of schizophrenia is necessary for the development of more effective treatment targets. Cognitive deficits are regarded as a core feature of schizophrenia and are thought to arise from alterations in ã-aminobutyric acid (GABA)-containing interneurons in the dorsolateral prefrontal cortex (DLPFC). Specifically, postmortem studies have demonstrated decreased levels of the mRNA encoding the 67 kDa isoform of glutamic acid decarboxylase (GAD67), an enzyme that synthesizes GABA, and this alteration seems to be specific to certain subsets of GABA neurons. For example, parvalbumin and somatostatin mRNAs, which are expressed in separate subsets of GABA neurons, were decreased, whereas calretinin mRNA, expressed in a third subset of GABA neurons, was unchanged in schizophrenia. The studies in this thesis examined the compartmental and cellular expression of and the potential causal mechanisms of reductions in SST mRNA expression; furthermore, the disease and cellular specificity of and post-synaptic consequences of reductions in SST mRNA expression were examined. We found that reductions in the levels of SST mRNA appear to be restricted to SST interneurons that do not contain NPY mRNA in the gray matter and are due to reductions in expression per neuron. These alterations appear to be a consequence of impaired neurotrophin signaling through the trkB receptor. Also, the profile of alterations in GABA-related mRNA expression is specific to schizophrenia. Finally, a post-synaptic receptor of SST, SST receptor subtype 2 (SSTR2), mRNA is reduced in schizophrenia. Since the SST protein is putatively inhibitory and SST-containing interneurons target the distal dendrites of pyramidal neurons, these data suggest reduced inhibition of pyramidal neurons and may represent a compensatory mechanism to increase excitatory drive. We conclude that reductions in SST and SSTR2 mRNA represent a downstream consequence of a neuropathological entity in the DLPFC of individuals with schizophrenia and contribute to cognitive dysfunction in schizophrenia.

During fast chemical synaptic transmission, neurotransmitter release is triggered by calcium (Ca2+) influx through voltage-gated Ca2+ channels (VGCCs) opened by an action potential (AP) at the nerve terminal. The magnitude and time course of neurotransmitter release is critically determined by the coupling between Ca2+ channels and synaptic vesicles. Studies of the quantitative dependence of transmitter release on the number of VGCCs provide important information for our understanding of the mechanisms that underlie the control and modulation of presynaptic release probability and kinetics. Using high-resolution calcium imaging techniques and variance analysis, I have determined the number of functional VGCCs within individual active zones (AZs) of the adult frog neuromuscular junction (NMJ) and their opening probability in response to single AP stimulation. The results have shown that the average number of VGCCs within individual active zones was relatively small (~28) and the average opening probability of individual Ca2+ channels during a presynaptic AP was very low (~0.24). Therefore, it is predicted that an action potential induced opening of relatively few Ca2+ channels in a single active zone. Furthermore, by combining pharmacological channel block, calcium imaging, postsynaptic recording, and 3D Monte Carlo diffusion-reaction simulations, I have studied the coupling of single Ca2+ channel openings to the triggering of vesicle fusion. I have provided evidence that Ca2+ entry through single open Ca2+ channels at the nerve terminal can be imaged directly and that such Ca2+ flux is sufficient to trigger synaptic vesicle fusion. I have further shown that following a single AP, the Ca2+ influx through a single open channel plays the predominant role in evoking neurotransmitter release, while Ca2+ ions derived from a collection of open Ca2+ channels are rarely required for vesicle exocytosis at this synapse.

In the sensorimotor domain, adaptation to distorted sensory input has been well-characterized and is largely attributed to learning mechanisms in the cerebellum that adjust motor output to achieve the same desired sensory outcome. Our interest in the role of the cerebellum in cognitive processes has led us to question whether it also contributes to adaptation in tasks that do not require voluntary motor output. Speech perception is a domain where there exist many examples of adaptation that are guided by both sensory and cognitive processes, without intentional motor involvement. Thus, we investigated behavioral and neural characteristics of speech perception adaptation to spectrally distorted words using a sophisticated noise-vocoded speech manipulation that mimics cochlear implants. We demonstrated that adaptation to spectrally distorted words can be achieved without explicit feedback by either gradually increasing the severity of the distortion or by using an intermediate distortion during training. We identified regions in both the cerebellar and cerebral cortex that showed differences in neural responses before and after training. In the cerebellum, this included regions in lobes V and VI, and Crus I. In the cerebrum, this included regions in the inferior frontal gyrus, the superior temporal sulcus, and the posterior inferior/middle temporal gyrus. In some of these regions, we further found changes in the magnitude of the neural responses that corresponded to the degree of behavioral improvements in performance. To gain some insight into the nature of the interactions between cerebral and cerebellar cortices and the types of representations involved in speech perception adaptation, we conducted a simple functional connectivity analysis using cerebellar seed regions of interest. We found interactions between the cerebellum and cerebral cortex that were dependent on the location of the cerebellar region. Overall, our behavioral and functional neuroimaging results point to cerebellar involvement in speech perception adaptation, and we conclude with a discussion of the learning mechanisms and neuroanatomical pathways that may support such plasticity.

The hippocampus is a brain structure thought to be critically important for the formation and maintenance of memories. In order do this, the region must process information as both a linear sequence and as discrete events or objects such that representations can be recalled even when the stimulus is incomplete. This is thought to be accomplished in the hippocampus through several streams of serial and parallel processing kept separate and intact by inhibition. These streams of processing take the anatomical form of the three major glutamatergic pathways of the hippocampus: the perforant path, the mossy fibers and the Schaffer collaterals. Inhibition is provided by specialized groups of GABAergic interneurons. Though the hippocampus has been the subject of intense study for decades, there remain populations of cells that are not well understood in terms of their role in the network. Within area CA3, one of these populations is a group of interneurons with soma residing in the str. lacunosum moleculare that provide feedforward inhibition onto CA3 pyramidal cells. The goal of this thesis was to understand the synaptic physiology of mossy fiber (MF) input to str. lacunosum moleculare interneurons (L-Mi), a connection that has been largely overlooked due to skepticism that an interneuron population ~ 250 µm away could be interacting with the MF pathway. The data presented here describe the functional anatomy of the connection, and define the short term synaptic physiology of MF input to L-Mi including an estimate of the quantal amplitude. I have also documented the modulation of this connection by a presynaptic metabotropic glutamate receptor not previously thought to have a role in MF physiology, thus expanding the known repertoire of MF target specificity. Most importantly, however, these data provide a mechanism through which feedforward inhibition onto CA3 pyramidal cells is regulated in the short term, under physiologic stimulus patterns, and contribute to our knowledge of the function of the CA3 network.

The role of dopamine neurons in reward processing is well-established, as is the observation of reward-related responses in the striatum, a region to which these midbrain dopamine neurons project. The reward-prediction error signals generated in the midbrain may play a role in the striatum in learning, as they help to shape expectations about future events based on prior experiences. The goal of the current experiment was to use principles of striatal function in order to optimize learning in an arithmetic domain. We created a training program that we believed would lead to increased arithmetic fluency by maximally engaging the striatum, through the use of contingent feedback, uncertainty regarding performance, and incentives for correct responses. Both experimental and control participants, who completed training focusing on arithmetic calculation and digit-entry respectively, showed improvement on a task involving the addition of a double-digit and a single-digit number following training, as successful performance on the task required accurate computations and entry of the solution within a narrow response window. We conducted functional magnetic resonance imaging before and after training while participants performed this task, in order to examine the effect of feedback on activity in the caudate nucleus and to determine if learning signals generated by the striatum during arithmetic training are able to modify quantity representations in parietal cortex. Results indicated activation of both the caudate nucleus and the hIPS region. Activation of the caudate nucleus replicated previous work, as it showed the prototypical pattern of activity that distinguished between positive and negative feedback. Activation of the hIPS region was not surprising, given the focus on arithmetic calculation, but this region also exhibited feedback-sensitive activation that differed between sessions and groups, possibly indicating the common influence of a reinforcement learning system.