Many daily decisions are instances of value-based decision-making. Examples include what to buy at a store, or what to eat for lunch. An essential computation that the brain needs to perform to make these kinds of decisions is to assign value to each of the options under consideration. An important open question in the field is whether the human amygdala encodes signals that affect value-based choices, perhaps by influencing activity in the medial orbitofrontal cortex. This proposal describes an extremely cross-disciplinary collaboration between a computational neuroscientist (Rick Jenison), a neurosurgeon and human electrophysiologist (Matthew Howard), and a neuroeconomist (Antonio Rangel) that seeks to combine tools from all of these fields to address this question. The proposal has four key aims:
Specific Aim 1. To apply computational and statistical modeling tools from point process theory and Bayesian particle filters to analyze the encoding of subjective value signals in single-unit recordings from awake behaving human patients with pharmacologically intractable epilepsy and in fMRI experiments.
Specific Aim 2. To investigate the extent to which the human amygdala is involved in value-based appetitive decision-making using single-unit recordings, microstimulation, fMRI, and the computational tools from Specific Aim 1.
Specific Aim 3. To investigate the extent to which the human amygdala is involved in value-based aversive decision-making using single-unit recordings, microstimulation, fMRI, and the computational tools from Specific Aim 1.
Specific Aim 4. To implement an intensive and unique cross-disciplinary education and training program for three post-docs in order to prepare them for further work at the intersection of the three disciplines represented in this proposal. Intellectual merit: First, the combination of point process and Bayesian particle filters tools from computational neuroscience, with human electrophysiology techniques from neurobiology, and with experimental paradigms from model-based fMRi taken from neuroeconomics has the potential to generate a quantum-leap in our understanding of the role of the human amygdala in value-based choice. Second, the methods that we propose will be able to look not only for correIations between brain activity and value computations, but also for the causality of such signals. Third, the computational approaches that we propose entail an extension of tools to domains in which the variables encoded by the brain are subjective (e.g., valuations) instead of objective (e.g., a location in space), which could have a myriad of applications in various areas of decision, social, and cognitive neuroscience, as well as psychiatry. Fourth, the combination of human single-unit data with related fMRI analyses will provide as a sidebenefit insights about the relationship between the BOLD signal and the underlying action potentials. Broader Impacts: The proposal also has a strong educational and training component. Future advances in neuroeconomics and decision neuroscience are likely to come at the crossdisciplinary intersection of computational neuroscience, neurobiology, and neuroeconomics. Unfortunately, it is difficult for young neuroscientists to get the necessary expertise in more than one of these areas. In order to address this shortcoming we will implement an intensive and quite unique crossdisciplinary education and training program for three post-docs to prepare them for further work at the intersection of the three disciplines represented in this proposal. Not only that, in order to increase diversity in science, applications for the post-doctoral position from women and members of disadvantaged minorities will be strongly encouraged, and such applicants will be favored. Many psychiatric disorders can be characterized as diseases of decision making. Consider, as extreme examples, the cases of addiction and OCD. The role of the amygdala in these diseases it is not well-understood. This research will contribute to the nascent field of computational psychiatry by advancing our understanding of the interaction between the amygdala computations and the value-based decision-making systems.
Showing the most recent 10 out of 11 publications
