Camillo Padoa-Schioppa - US grants

Economic choice is the behavior observed when individuals make choices solely based on subjective preferences[unreadable]for example out of a restaurant menu. Choice behavior is a traditional focus in economics and psychology. However, economic choice is also critically important from a medical perspective, as choice deficits are often observed in patients with neurological disorders of the frontal lobe and in addiction. Behaviorally, evidence indicates that economic choice entails assigning values to the available options[unreadable]for example, to different items on the menu. The underlying neurobiological mechanisms remain, however, poorly understood. Evidence from lesions, functional imaging and electrophysiology suggest that the orbitofrontal cortex (OFC) may play an important role in economic choice. The overarching goal of this proposal is to examine the contribution of this cortical area in detail. This will be achieved by recording the activity of individual neurons in awake monkeys while they engage in economic choice tasks. In our experiments, monkeys choose between different beverages offered in variable amounts. Their choices provide a behavioral measure for the relative value of the two juices. Our preliminary studies indicate that neurons in OFC encode the value monkeys assign to offered and chosen juices, suggesting that choices are based upon the activity of this neuronal population. This observation raises several critical questions that will be addressed in the experiments proposed here. First, does the neuronal representation of value vary depending on the behavioral context? For example, the neuronal valuation system might adapt to the contextual value range to exploit maximal discriminatory power, in a way similar to how the visual system adapts to ambient light. Second, do OFC neurons encode economic utility or pure value? Choices are often made between options that involve risks and costs, both reflected in the expected utility. If choices are based on the activity of OFC, we expect neurons in this area to encode economic utility, as distinguished from pure value. Finally, in spite of a substantial literature on the anatomy of the frontal lobe, what areas receive anatomical projections from value-encoding neurons remains unclear. To address this issue, we will combine anatomical tracings with neurophysiological recordings. This study will also provide the PI an opportunity to learn histological and anatomical techniques that will complement his portfolio of expertise.

DESCRIPTION (provided by applicant): Economic choice (EC) is the behavior observed when individuals make choices solely based on subjective preferences - for example out of a restaurant menu. As a mental function, EC entails assigning values to the available options;a decision is then made by comparing these values. Several lines of evidence implicate the orbitofrontal cortex (OFC) as a neural substrate for EC. First, lesions or malfunction of the OFC (e.g., in frontotemporal dementia and drug addiction) typically disrupt EC behavior. Second, imaging studies in humans show enhanced OFC activation when subjects make choices compared to when choices are made for them. Third, individual neurons in the primate OFC encode the value subjects assign to different options when they choose between them. Taken together, these elements suggest that OFC plays a role in EC. However, a causal relationship between values encoded in the OFC and behavioral choices has not yet been established. To begin addressing this fundamental question, we will perform a secondary analysis of existing data (Aim 1) and we will conduct two pilot studies using electrical microstimulation and pharmacological inactivation (Aim 2). In our experiments monkeys choose between two juices offered in variable amounts. We previously found that neurons in the OFC encode the subjective values monkeys assign to different juices while choosing. More precisely, neuronal responses in the OFC encode 3 variables: offer value (the value of one of the two juices), chosen value (the value of the chosen juice) and taste (the chosen juice type, independently of the amount). In a computational sense, offer value responses (which are most prevalent immediately after the offer) seem to encode the values necessary to make a choice. Conversely, chosen value responses and taste responses seem to encode the result of the choice process. Thus our working hypothesis is that choices may be based on offer value responses. To test this hypothesis and establish a causal link between the encoding of offer values and choices, we will address three key questions. (1) Are variables offer value and chosen value encoded by distinct and perhaps anatomically clustered populations of cells? This question will be addressed with a secondary analysis of existing data. (2) Can choices be biased by "artificially" activating a population of neurons encoding the offer value? This question will be addressed using electrical microstimulation. Specifically, we will identify sites where neurons encode the offer value of a particular juice and inject small current as a monkey chooses between that juice and another juice. (3) Are choices disrupted by "silencing" this neuronal population? This question will be addressed using reversible pharmacological inactivation. Demonstrating a causal relationship between the encoding of value in OFC and choice behavior would represent an important breakthrough. Thus this research can potentially shape how we understand the neural mechanisms that underlie EC and that malfunction in mental illness. PUBLIC HEALTH RELEVANCE: When individual make choices - for example between different dishes on a restaurant menu - neurons in the orbitofrontal cortex (OFC) compute the subjective value of the different options. This research will test whether choices are indeed caused by values computed in the OFC.

ABSTRACT Mental and neurological disorders such as frontotemporal dementia and drug addiction can be characterized as patients making ?poor choices.? Specifically affected are choices based on subjective preferences, also referred to as economic choices. Thus to better understand the origins of these disorders and to pave the way for treatments, it is critical to understand the neural underpinnings of this behavior. Economic choices are thought to involve two mental stages ? values are assigned to the available options and a decision is made by comparing values. Evidence from clinical data, lesion studies, functional imaging and neurophysiology links economic decisions to the orbitofrontal cortex (OFC). In particular, work in my lab examined the activity of OFC neurons in monkeys choosing between different juices. We thus identified three populations intimately related to the decision: offer value cells encoding the value of individual juices; chosen value cells encoding the value of the chosen juice; and chosen juice cells encoding the identity of the chosen juice independent of quantity. Notably, these groups of cells capture both the input (offer value) and the output (chosen juice, chosen value) of the decision process, suggesting that decisions might be generated within OFC. The overarching goal of this proposal is to shed light onto the organization and mechanisms of this neural circuit. Using a combination of behavioral, computational and neurophysiology techniques, we will pursue two Specific Aims. Aim 1 focuses on choices under sequential offers. The vast majority of previous studies examined choices between goods offered simultaneously. Exp.1 will establish whether current notions generalize to choices under sequential offers, which are arguably more relevant to real-life decisions. Furthermore, by dissociating value computation from value comparison, Exp.1 will address outstanding questions on the decision circuit. Aim 2 focuses on the fundamental question of causality. Despite the enormous advances of recent years, our understanding of the mechanisms underlying economic decisions remains tentative. This is primarily because causal links between OFC activity and decisions have not yet been established. In principle, such links may be demonstrated using electrical stimulation to bias choices in a predictable way, but the absence of columnar organization in OFC makes this approach more difficult. To obviate this issue, we designed three experimental protocols. In Exp.2, monkeys will perform choices under sequential offers. By injecting current during one of the offers, we will bias the animal's decision. Exp.3 and Exp.4 will focus on specific groups of cells. The experiments will take advantage of phenomena recently discovered in my lab, namely the range adaptation of offer value cells (Exp.3), the predictive activity of chosen juice cells and choice hysteresis (Exp.4). The first cycle of this grant has generated multiple breakthroughs, and the current Specific Aims build on our previous results. By fulfilling them, the research described here will significantly advance our understanding of the neural mechanisms that underlie economic decisions and that malfunction in mental illness.

? DESCRIPTION (provided by applicant): Many mental and neurological disorders - including frontotemporal dementia, obsessive-compulsive disorder and drug addiction - are broadly characterized by patients making poor choices. More specifically, deficits are observed when subjects make decisions based on subjective preferences - a behavior termed economic choice. Thus to better understand the origins of these disorders, and to pave the way for treatments, it is critical to understand the neuronal underpinnings of economic choice. This proposal builds on recent results showing that neurons in the primate orbitofrontal cortex (OFC) encode three variables intimately related to the decision. In the experiments, monkeys chose between different juices offered in variable amounts. Different groups of cells encoded the offer value (the value of one of the two juices, independent of the animal choice), the chosen value (the value of the chosen juice, independent of its identity) and the chosen juice (the identity of the chosen juice, independent of its value). In principle, these groups of cells could be sufficien to generate a decision. Furthermore, an established literature shows that OFC lesions selectively impair economic choices. We thus put forth the hypothesis that good-based decisions take place within the OFC. We will pursue three specific aims. In Aim 1, we will assess whether and how the identity of individual neurons is maintained across behavioral contexts. Monkeys will choose between two juices (A and B) in a first trial block and between different juices (C and D) in a second trial block. Preliminary results indicate that a given cell typically encodes the same variable in the two blocks. At the same time, cells are re-assigned to the juices available in any given block. In Aim 2, we will examine the neuronal origins of choice variability. We recently discovered several phenomena relating the decision made by the animal in any given trial with trial-by-trial fluctuations in the activity of individual neurons in OFC. Theoretical considerations indicate that these phenomena critically depend on how trial-by-trial fluctuations in the activity of different neurons are correlated with each other. Under Aim 2, we will measure noise correlations in the OFC using high-density U-probes. In addition, we will assess whether different groups of cells are anatomically organized in clusters or mini-columns. In Aim 3, we will examine the role of the amygdala (AMY). With the OFC, the AMY is the only area where lesions disrupt performance in reinforcer devaluation tasks. Previous work suggests that value-relevant information flows from the AMY to the OFC. However, the precise role of the AMY in economic decisions remains unclear. While lesion studies long implicated the OFC in choice behavior, the hypothesis that economic decisions take place within the OFC is highly innovative and potentially transformative. Throughout this proposal, experiments and computational work will proceed hand-in-hand. In our estimation, this research is the most advanced ever conducted to understand the neuronal mechanisms underlying economic decisions.

Many mental and neurological disorders ? including frontotemporal dementia, schizophrenia and drug addiction ? can be broadly characterized as patients making ?poor choices.? While these disorders disrupt choices in a variety of domains, deficits can be unitarily described as affecting decisions based on subjective preferences. This behavior is referred to as ?economic choice?. Thus to better understand the origins of these disorders, and to pave the way for treatments, it is critical to understand the neural underpinnings of this behavior. Evidence from neurophysiology in non-human primates, functional imaging in humans and lesions studies in multiple species establishes a link between economic decisions and the orbitofrontal cortex (OFC). In particular, previous research in the PI's lab examined the activity of single neurons in the OFC of monkeys choosing between different liquid rewards. This work identified three groups of cells encoding the identity and subjective values of offered and chosen goods. The three groups of cells identified in the primate OFC capture both the input and the output of the choice process, suggesting that decisions are formed in a circuit composed by these neurons. Recent results support this hypothesis. However, most aspects of this neural circuit remain poorly understood. For example, it is unclear whether the three groups of cells correspond to different morphologically identified cell types, whether they reside in different cortical layers, whether they use different neuro-transmitters and/or whether they can be manipulated genetically. Also, a causal link between neurons in the OFC and the decision made by the animal has not yet been established. Addressing these questions using traditional neurophysiology is technically difficult. However, all of these questions can in principle be addressed using genetic approaches developed in recent years. The overarching goal of this proposal is to take this field of investigation to a genetically tractable model ? the mouse. Specific aims are (1) to develop such animal model, (2) to recapitulate in mice the main physiological findings from the monkey work, and (3) to exploit this model to address some of the outstanding questions. Specifically, we will combine optogenetics and neuronal recordings to assess whether different groups of cells identified in relation to behavior are inhibitory or excitatory, and whether they are located in different cortical layers. If successful, this research will provide the opportunity for a quantum leap and open new avenues to address numerous fundamental questions. This is a collaboration between Dr Padoa-Schioppa (PI), an expert of economic decisions in primates, and Dr Holy (co-I), an expert of olfaction in mice. The experiments will break new ground and extend previous work in a new and promising direction. While preliminary results argue for feasibility, it remains unclear whether the critical experiments will provide interpretable results. Thus this research is high-risk high-reward. Finally, our ultimate goal is to establish a new model that will significantly impact future research on decision making. For all these reasons, the proposal is closely aligned with the purposes of the R21 funding mechanism. 1

Over the past 30 years, the scope of Systems Neuroscience has expanded enormously. Functional brain imaging has provided the opportunity to study the neural mechanisms of complex cognitive functions in humans. Concurrently, powerful techniques such as optogenetics have made it possible to dissect neural circuits with unprecedented resolution. In parallel, computational approaches such as deep networks and Bayesian models are increasingly central to the field as a whole, putting a premium on advanced quantitative skills and literacy. Clearly Systems Neuroscience is now a trans-disciplinary field, integrating theoretical frameworks and techniques from molecular biology, neurophysiology, cognitive science, ethology, computer science, statistics, and more. In the face of this remarkable expansion, PhD programs face three challenges. First, although the body of knowledge relevant to Systems Neuroscience has increased in breadth and depth, students also face increased pressure to conduct research, publish, and get independent funding early in their career. Second, systems-level research is conducted in multiple departments and PhD programs. At Washington University, these include the PhD programs in Neuroscience, Psychology, and Biomedical Engineering. Yet, students coming from different disciplines often do not speak each other?s language. Third, a successful career in science requires a broad portfolio of professional skills ? writing papers and grant proposals, collaborating with colleagues with different scientific backgrounds, presenting results in scientific venues and to wider audiences, navigating the academic job market ? that exceed the normal coursework. The Cognitive, Computational and Systems Neuroscience (CCSN) pathway was developed in response to these challenges. CCSN is an elite pathway available for graduate students in years 3-4, with eligibility from multiple PhD programs relevant to Systems Neuroscience broadly defined. The emphasis of CCSN is on trans- disciplinary training and professional skills development. To access the pathway, students must complete (in years 1-2), three pre-requisite and foundational courses on systems neuroscience, cognitive science and animal behavior, and computational neuroscience. In year 3, CCSN students take two additional courses ? one providing foundational knowledge and hands-on training with advanced quantitative methods and data-science tools, and the other allowing them to develop a trans-disciplinary grant proposal ? which often becomes an actual NRSA application ? shaped by peer, instructor, and committee feedback. In year 4 (and throughout the pathway), CCSN students take part in multiple Career Development activities, including mentoring junior students, organizing scientific events, interacting with external speakers, participating in informal dinners with CCSN faculty, and taking part in community outreach. The CCSN pathway has existed for ~15 years and has a demonstrated history of remarkable success. Here we request funds for 5 fellowships. Contingent on the success of this application, Washington University will provide matching funds for an additional 5 slots.