We successfully navigate the world by making decisions based on what we have learned. In the brain, two prominent learning systems have been identified and each is likely to guide decisions in different ways. Research on decision making has primarily focused on a reward learning system in the striatum. These studies have illuminated the how repeated choices and rewards build representations that guide choices and actions when encountering the same situation again. However, in a constantly changing environment, choices may not repeat themselves. Further, the environment may have more structure than simple reward learning can navigate. In these situations, decisions may be guided by a different learning system, namely a flexible learning system in the hippocampus which encodes episodes, or more broadly, relations between stimuli. However, investigations into the role of a reward learning system and a relational learning system in decision making have developed largely independently of each other. In the studies described below, I explore the function of these learning systems in value-guided decision making. Complementarily, I also explore how ongoing reward learning may modulate memory formation in the hippocampal system. In these studies, I demonstrate that reward learning and decision making is influenced by relational learning, and that these effects are predicted by hippocampal-striatal connectivity during learning. Separately, I establish that episodic memory is, in turn, influenced by ongoing reward learning. Successful memory is predicted by modulations of reward and memory regions including the striatum and hippocampus. Overall, these results provide novel insights into the learning systems encoding memories for future adaptive behavior.

Naturally occurring apoptosis in the developing nervous system is an important event for the proper shaping of the system, for eliminating precursor cells with mutations and for curbing the population of post-mitotic neurons so that appropriate target innervation can take place. In the latter case, neurons are usually subjected to a competition for a limited supply of target-derived growth factor. A classic example includes neurons of the superior cervical ganglia that compete for nerve growth factor (NGF). Cell death arising from lack of NGF stimulation is a topic of intense research and remains enigmatic. Pro-apoptotic molecules that have been characterized so far are partially responsible for inducing death and furthermore, it is not fully understood how the homeostatic balance of the neuron is disrupted by NGF deprivation. In this thesis, a novel neuronal pro-apoptotic protein Trib3 is defined. It is shown to be required for developmental neuron death and to employ a cellular mechanism involving Akt and its substrates, the FoxO transcription factors. It is also shown that Trib3 is transcriptionally regulated by the FoxO, JNK and apoptotic cell-cycle pathways. The data further demonstrate that Trib3 functions in a feed-forward loop with the FoxOs and deactivates Akt in a self-propelled pathway that amplifies the apoptotic cascade resulting from NGF deprivation. Moreover, Trib3 is found to be induced and necessary for neuron death in cellular models of Alzheimer's and Parkinson's diseases. The latter preliminary findings have encouraged other researchers to begin exploring Trib3's mechanism and regulation in neurodegenerative disease models; this will advance our knowledge and understanding of the cellular processes of neuron death and may lead to development of novel therapeutic targets.

One of the most remarkable features of the central nervous system is the exquisite specificity of its synaptic connections, which is crucial for the functioning of neuronal circuits. Thus, understanding the cellular and molecular mechanisms leading to the precise assembly of neuronal circuits is a major focus of developmental neurobiology. The structural organization and specific connectivity of neuronal circuits arises from a series of morphological transformations: neuronal differentiation, migration, axonal guidance, axonal and dendritic arbor growth and, eventually, synapse formation. Changes in neuronal morphology are driven by cell intrinsic programs and by instructive signals from the environment, which are transduced by transmembrane receptors on the neuronal cell surface. Intracellularly, cytoskeletal rearrangements orchestrate the dynamic modification of neuronal morphology. A central question is how the activation of a neuronal cell surface receptor triggers the intracellular cytoskeletal rearrangements that mediate morphological transformations. A group of proteins linked to the regulation of cytoskeletal dynamics are the small GTPases of the Rho family. Small RhoGTPases are regulated by GTPase exchange factors (GEF) and GTPase activating proteins (GAP), which can switch GTPases into "on or off" states, respectively. It is thought, that GEFs and GAPs function as intracellular mediators between transmembrane receptors and RhoGTPases, to regulate cytoskeletal rearrangements. During my dissertation I identified the GAP α2-chimaerin as an essential downstream effector of the axon guidance receptor EphA4, in the assembly of neuronal locomotor circuits in the mouse. Furthermore, I identified two novel neuronal GAPs, mSYD-1A and mSYD-1B, which interact with components of the presynaptic active zone and which may contribute to presynaptic assembly downstream of synaptic cell surface receptors.

Neighboring neurons in cat primary visual cortex (V1) have similar preferred orientation, direction, and spatial frequency. How diverse is their degree of tuning for these properties? Are they also clustered in their tuning for the spatial phase of a flashed grating ("absolute spatial phase") or the temporal phase of a drifting grating ("temporal response phase")? To address these questions, we used tetrode recordings to simultaneously isolate multiple cells at single recording sites and record their responses to flashed and drifting gratings of multiple orientations, spatial frequencies, and spatial/temporal phases. We recorded the responses of 761 cells presented with drifting gratings and 409 cells presented with flashed gratings. We found that orientation tuning width, spatial frequency tuning width and direction selectivity index all showed significant clustering. Absolute spatial phase and temporal response phase, however, showed no clustering. We also present an algorithm that improves the performance of spike-sorting algorithms, for use in analyzing cells recorded using tetrodes. A cluster of spikes corresponding to a putative cell obtained through automatic or manual spike sorting algorithms may contain spikes from other cells with similarly-shaped waveforms. Our algorithm preferentially removes contaminating spikes from other cells, thereby decreasing the level of contamination of each unit. We call this procedure "pruning", as it entails removing portions of the cluster that are determined to be more likely to contain contaminating spikes than the cluster as a whole. Testing of the algorithm on data in which "ground truth" is known shows excellent performance, for example on average giving a percentage reduction in false positive spikes 8.2 times the percentage reduction in true positive spikes, and reducing the degree of contamination by an average of about 13%.

Both placebo and distraction have long been used clinically to relieve pain. The present series of experiments examined the neural and cognitive processes that constitute these two psychological forms of analgesia. Study 1 provides evidence that overlapping cognitive resources are involved in both pain and executive attention and working memory. Study 2 provides evidence that these same executive attention and working memory resources are not involved in placebo analgesia, and that placebo analgesia and distraction provide separate routes to pain relief. Study 3 suggests that while distraction-based analgesia reduces the neural signature of pain, expectancy-driven placebo analgesia may not.

The knowledge of the maps of neuronal interactions is key for system neuroscience, but at the moment we possess relatively little of it . The recent development of experimental methods which allow a simultaneous recording of the spiking activity, but not the intracellular voltage, of thousands of neurons gives us an opportunity to start filling that gap. In Chapter 2, I present a method for the inference of the parameters of the leaky integrate-and-fire (LIF) model featuring time-dependent currents and conductances based only on the extracellular recording of spiking in the network. The fitted parameters can describe the functional connections in the network, as well as the internal properties of the cells. The method can also be used to determine whether a single-compartment model of a neuron should include conductance- or current-based synapses, or their mixture. In addition, because the same mathematical model describes some of the flavors of the Drift Diffusion Model (DDM), popular in the studies of decision making process, the presented method can be readily used to fit their parameters. Making the proposed inference procedure -- based on the expectation-maximization (EM) algorithm -- accurate and robust, necessitated a development of a new numerical adaptive-grid (AG) method for the forward-backward (FB) propagation of the probability density, which is required in the computation of the sufficient statistic in the EM algorithm. These topics are covered in Chapter 3. Another issue which had to be addressed in order to obtain a usable inference algorithm is the well known slow convergence of the EM algorithm in the flat regions of the loglikelihood. Two complementary approaches to this issue are presented in this dissertation. In Chapter 4, I present a new framework for the acceleration of convergence of iterative algorithms (not limited to the EM) which unifies all previously known methods and allows us to construct a new method demonstrating the best performance of them all. To make the computations even faster, I wrote a Matlab package which allows them to be done in parallel on several machines and clusters. As one can see, all the aforementioned projects were sprouted up from one "head" project on the inference of the LIF model parameters. At the end of the dissertation, I briefly describe a disconnected project which is devoted to the development of a flexible experimental setup (software and hardware) for behavioral experiments, with a specific application to a particular type of the virtual Morris water maze experiment (VMWM).

Neurons are divided into functional compartments: the soma houses the DNA and transcriptional machinery, the axons conduct action potentials, presynaptic boutons transmit synaptic signals, and dendrites receive and integrate synaptic information. These functional differences are achieved by localizing different complements of proteins to these compartments; axons contain a high density of voltage-gated channels, presynaptic boutons house the machinery for synaptic vesicle-mediated neurotransmitter release, while dendrites contain ligand-gated and voltage-gated ion channels that generate and integrate postsynaptic responses. These primary compartments are further divided in some cell types, allowing the dendrite to expand its information processing power. For example, in CA1 pyramidal neurons of the hippocampus, a brain region crucial for memory formation, the apical dendrite is divided into a proximal and a distal compartment. These two dendritic compartments receive synaptic inputs from different sources and contain different proteins, making them functionally distinct. While many motifs and molecules have been identified that regulate the trafficking of proteins to axons versus dendrites, little is known about the mechanisms of protein targeting to compartments within a dendrite. We have investigated the regulation of protein composition of the distal dendritic compartment in the CA1 neurons. This distal compartment is enriched in specific ion channels, including the hyperpolarization-activated HCN1 cation channel, which we have focused on because of its striking distal localization and its importance in hippocampal function. Using dissociated and organotypic hippocampal culture systems, we show that HCN1 surface expression is activity dependent. However, the axons that innervate the distal compartment are not required for HCN1 localization. Our data suggest that while activity plays a role in HCN1 channel regulation, it is not sufficient for the distal dendritic targeting. We show that proper distal localization of HCN1 channels requires a non-cell autonomous factor. We provide evidence that the extracellular matrix protein reelin acts as this non-cell autonomous factor regulating ion channel composition of the distal dendrites. Blocking reelin signaling in organotypic culture results in reduction of HCN1 in the dendrites and distal HCN1 levels are reduced in reeler mice. In vivo viral knockdown of dab1, the intracellular signaling partner of reelin, leads to loss of HCN1 and other distally enriched ion channels specifically from the distal dendrites, but does not alter ion channel composition of the proximal dendritic compartment. Viral knockdown of dab1 increases a physiological indicator of HCN channels at the soma, indicating that some of HCN1 channels are redistributed. Blockade of src family kinases that are activated by reelin signaling likewise leads to a loss of distally enriched ion channels. These results define a novel role for reelin signaling in dendritic compartmentalization by regulating ion channel composition.

We develop new methods of Bayesian inference, largely in the context of analysis of neuroscience data. The work is broken into several parts. In the first part, we introduce a novel class of joint probability distributions in which exact inference is tractable. Previously it has been difficult to find general constructions for models in which efficient exact inference is possible, outside of certain classical cases. We identify a class of such models that are tractable owing to a certain "low-rank" structure in the potentials that couple neighboring variables. In the second part we develop methods to quantify and measure information loss in analysis of neuronal spike train data due to two types of noise, making use of the ideas developed in the first part. Information about neuronal identity or temporal resolution may be lost during spike detection and sorting, or precision of spike times may be corrupted by various effects. We quantify the information lost due to these effects for the relatively simple but sufficiently broad class of Markovian model neurons. We find that decoders that model the probability distribution of spike-neuron assignments significantly outperform decoders that use only the most likely spike assignments. We also apply the ideas of the low-rank models from the first section to defining a class of prior distributions over the space of stimuli (or other covariate) which, by conjugacy, preserve the tractability of inference. In the third part, we treat Bayesian methods for the estimation of sparse signals, with application to the locating of synapses in a dendritic tree. We develop a compartmentalized model of the dendritic tree. Building on previous work that applied and generalized ideas of least angle regression to obtain a fast Bayesian solution to the resulting estimation problem, we describe two other approaches to the same problem, one employing a horseshoe prior and the other using various spike-and-slab priors. In the last part, we revisit the low-rank models of the first section and apply them to the problem of inferring orientation selectivity maps from noisy observations of orientation preference. The relevant low-rank model exploits the self-conjugacy of the von Mises distribution on the circle. Because the orientation map model is loopy, we cannot do exact inference on the low-rank model by the forward backward algorithm, but block-wise Gibbs sampling by the forward backward algorithm speeds mixing. We explore another von Mises coupling potential Gibbs sampler that proves to effectively smooth noisily observed orientation maps.

This thesis details two projects at the interface of chemistry and neuroscience. Part I focuses on the development and characterization of fluorescent optical tracers of monoamine neurotransmitters for use in the study of monoamine transporter activity and distribution, as well as synaptic transmission in the brain. The second section details early studies concerning experimental therapeutics recently identified by our group to induce the synthesis and release of Glial Derived Neurotrophic Factor (GDNF) from a model astroglial cell line. Part I consists of Chapters 1-5. Chapter 1 provides a brief introduction to the relationship between chemistry and neuroscience, while explaining the motivation and background behind the development and study of Fluorescent False Neurotransmitters (FFNs). Chapter 2 discusses the characterization of new fluorescent substrates for the Vesicular Monoamine Transporter 2 (VMAT2) for applications in high throughput screening of VMAT activity and inhibition, as well as for imaging aminergic synaptic transmission in brain tissue. Chapter 3 describes the discovery of fluorescent probes FFN201 and AGH093 as dual substrates for the norepinephrine transporter (NET) and VMAT, with potential applications in the imaging of noradrenergic modulation of cortical signaling networks. NG54, a FFN with substrate activity at each of the monoamine transporters was also identified, and represents a new structural class of FFNs. In Chapter 4, APP+, a known fluorescent substrate for the monoamine transporters, is evaluated as a potential tracer of monoamines in the brain. It has been determined that while this molecule may have utility as a marker for monoaminergic neurons, APP+ is not an appropriate FFN due to high background labeling of mitochondria and other intracellular structures. Chapter 5 discusses the discovery of fluorescent substrates of the Organic Cation Transporter 3 (OCT3) and Plasma Membrane Monoamine Transporter (PMAT). Mounting evidence suggests that these transporters play a significant role in clearing extracellular space of monoamines and are likely involved in neurological disease. As such, fluorescent substrates of OCT3 or PMAT would be useful as imaging agents in the brain, as well as for the development of fluorescence-based quantitative inhibition assays. Part II of this thesis (Chapter 6) discusses N-arylethyl isoquinuclidines as releasers of glial cell line-derived neurotrophic factor (GDNF) from a model glial cell line. Initial characterization of GDNF release and its dependence on protein synthesis and MAPK/ERK signaling is described. Preliminary studies indicate that at least two experimental compounds described herein modulate fibroblast growth factor receptor (FGFR) signaling.

Stress is an anticipatory or actual disruption of homeostasis which triggers physiological responses to prepare for the threat or to reestablish the system's internal equilibrium. Other than physical stress (e.g. blood loss, infection, pain, and cold) and psychogenic stress (e.g. novel environment (rodent) and immobilization (rodent)), nutrient (metabolic) stress such as high fat diet and fasting can also trigger stress responses. Central to the activation and regulation of the stress response in mammals is the hypothalamus, where signals from the periphery and other regions of the brain are integrated and processed. Insulin and leptin are two important hormones that provide information about the energy status of the body to the hypothalamus. The hypothalamic insulin- and leptin- sensing circuits integrate these signals to coordinate metabolic outcomes such as food intake, energy expenditure, fuel partitioning, etc. Given the tight correlation between stress axis activity and nutrient status, it raises the possibility that hypothalamic insulin- and leptin- sensing circuits may also be involved in coordinating responses of the stress axis, particularly during stress pertaining changes in energy homeostasis. My thesis research focused on understanding the roles and the interactions of hypothalamic insulin and leptin signals in regulating and coordinating metabolic and hypothalamic-pituitary-adrenal (HPA) axis functions. My first project involved understanding the role of hypothalamic insulin signals in hypothalamic leptin receptor deficient animals (L^2.1 KO) in regulating energy metabolism (Chapter 2). We observed an increase in body weight and adiposity in D^2.1 KO mice that lack both hypothalamic insulin receptor (InsR) and leptin receptor (LepRb) signals. These changes were accompanied by reduced energy expenditure, rather than an increase in food intake. Unexpectedly, there was a drastic loss in body temperature in D^2.1 KO during fasting that was significantly exacerbated by single-housing. These results suggest that interactions between hypothalamic insulin and leptin signals are important for regulating energy expenditure and body temperature. Furthermore, this study also highlights the influence of housing conditions on the evaluation of thermogenic defects in mice. My second project involved the study of hypothalamic insulin signals in regulating stress-related functions by using the non-obese hypothalamic InsR-deficient mice (I^2.1 KO)(Chapter 3). We showed that I^2.1 KOs have elevated baseline hypothalamic Avp synthesis, increased activity of the HPA axis after restraint, as well as increased anxiety-like behaviors. This study demonstrated that hypothalamic InsR signals suppress the stress response to restraint, possibly by influencing AVP release to the median eminence and decreasing hypothalamic glucocorticoid receptor (GR) signals, and may also modulate anxiety-like behaviors. My current research focuses are: to remove InsR signals by using more restrictedly expressed Cre lines in order to identify brain nuclei mediating insulin's effect on stress response and/or anxiety-like behaviors; to identify the hypothalamic AVP neuronal population affected by the loss of InsR and potential extra-hypothalamic downstream targets of these neurons; to pinpoint the hypothalamic nuclei where GR signaling is altered in I^2.1 KO. By identifying the critical anatomical and functional components mediating insulin's effects, we hope to provide more understanding of the contribution of central insulin resistance to the development of HPA axis dsyregulation and anxiety disorders in humans.