C Daniel Salzman - US grants

DESCRIPTION (provided by applicant): This project seeks to elucidate how the brain can form predictions about impending reinforcement in a variety of behavioral contexts and learning paradigms, which is a fundamental goal in neuroscience. The project involves obtaining neurophysiological recordings in the amygdala and OFC simultaneously so that one can understand the relationship and timing of activity between the two brain areas. Recent work from the Salzman lab has shown that the amygdala provides a representation of the positive or negative value of visual stimuli during a classical conditioning procedure in which there is a one-to-one mapping between a sensory stimulus and a particular reinforcement outcome. This proposal involves extending this recent work to now examine simultaneously the interrelated neurophysiology of the amygdala and orbitofrontal cortex (OFC) during both simple and more complex forms of reinforcement learning. The amygdala and OFC are central nodes in neural circuitry commonly assumed to link sensory stimuli with affective values so as to drive adaptive cognitive, behavioral, and physiological responses. Dysfunction of these neural circuits likely plays a role in many psychiatric diseases, such as mood, anxiety, addictive and other disorders. The first aim examines amygdala and OFC single neuron activity local field potentials (LFPs) during learning induced by classical conditioning in order to understand the physiological properties and relative timing dynamics of activity in the two brain areas. The second aim extends this work by studying the physiology of these brain areas in conditions when motivationally significant stimuli have different meanings depending on the moment-to-moment context in which they are presented. If neural processing in the amygdala and OFC can switch rapidly as the value of a stimulus changes from trial to trial depending upon a contextual cue, it will indicate that rapid context-dependent mechanisms can facilitate the switching between representations of value. The third aim investigates whether neurons in the amygdala and OFC represent the ?absolute? or the ?relative? value of conditioned stimuli by using a reinforcer revaluation paradigm that manipulates the relative value of a stimulus but holds constant its absolute value. This task requires the integration of information about the overall context of the task in order to judge the relative value of a particular stimulus. For both the second and third aims, we hypothesize that OFC either encodes contextual information about a stimulus (Aim 2), or integrates information about an overall task context (Aim 3) in part to help govern and update neural representations of value. In this scenario, OFC would be part of a cortical mechanism that integrates high-level information in order to help control neural representations of value and the emotional processes that are based on such representations.

Neuromodulatory treatments, such as repetitive transcranial magnetic stimulation (rTMS) applied to dorsolateral prefrontal cortex (DLPFC), have shown efficacy in reducing cocaine craving. This grant seeks to understand the neurophysiological mechanisms underlying rTMS efficacy in reducing cocaine induced behaviors at a circuit level, and, more broadly, the effects of rTMS on single cell physiology. Two rhesus monkeys will perform a decision-making task in which they choose between iv cocaine and a natural reward. Parametric manipulations of reward amounts or probability will enable quantification of how monkeys value cocaine in relation to natural rewards. As the monkey develops a preference for cocaine, we will record simultaneously in DLPFC and two brain areas known to represent the values of rewards during economic decision-making, the orbitofrontal cortex (OFC) and amygdala. We will test the hypothesis that neuronal activity recorded in DLPFC, OFC and amygdala is correlated with how monkeys value cocaine. As monkeys escalate cocaine self-administration, reflecting an increase in how cocaine is valued, we hypothesize that OFC and amygdala will update neural representations of the value of cocaine (Aim 1). Monkeys will then enter a cross-over rTMS vs. sham trial design. They will receive either rTMS at 15Hz, 100% RMT (resting motor threshold) or sham, every other day for 20 sessions in left DLPFC (area 9/46d). The days without rTMS, monkeys will undergo a recording session with the same choice paradigm. We hypothesize that rTMS will cause a decline in preference for cocaine, a parallel shift in the encoding of relative value in OFC and amygdala, and restoration of baseline activity in DLPFC for monkeys undergoing rTMS vs. sham monkeys (Aim 2). Recordings will continue for 3 months after the last treatment to assess the duration of rTMS induced electrophysiological changes. Different frequencies and different networks will be tested if needed. This grant will be the first to characterize directly neuronal responses in relation to rTMS hours and days after applications. This will advance rTMS treatment of substance abuse by elucidating the circuits and mechanisms responsible for conferring clinical benefits, which in turn may provide key insights for developing and targeting improved therapeutics in the future.

As animals move about in the environment, they must be able to rapidly assess and update the meaning of cues around them in order to approach reward and avoid danger. Reversal learning, a paradigm commonly used to test behavioral flexibility, involves switching the association between a stimulus and its outcome. For example, a tone that initially predicted a sweet tastant can come to predict a bitter tastant. Two brain regions, the orbitofrontal cortex (OFC) and basolateral amygdala (BLA), are involved in tasks that demand flexible responses to the changing significance of stimuli during reversal learning. While several studies have provided evidence that the OFC and BLA interact during reversal learning, little is known about the specific mechanisms of this interaction. In this grant, we propose to use a combination of genetic and cellular imaging techniques in mice to identify and characterize the physiological response properties during reversal learning of BLA neurons that project to OFC. The reversal learning task we employ involves stimuli of multiple modalities that predict outcomes of both positive and negative value to thoroughly elucidate the differential response properties of BLA neurons to stimuli that change their associated reinforcement outcome (Aim 1). In these experiments, the physiological response properties of neurons identified as projecting from BLA to OFC will be compared to the response properties of the overall population. Then, using an optogenetic approach, we will test the causal role of OFC input onto BLA neurons during reversal learning by inhibiting the specific projections from OFC to BLA to test their causal role on behavioral flexibility and on the nature of representations in BLA (Aim 2). The data analysis for both aims involves collaboration with theoretical neuroscientists to use linear classifiers and other sophisticated techniques to understand the nature and dynamics of neural encoding in BLA and its relation to behavior. These experiments promise to shed light on the mechanisms by which cortico-amygdalar circuits mediate flexibility critical for adaptive emotional responses and behavior, an ability that is impaired in individuals with psychiatric disorders.