From optimal to biological theories of choice: identifying neural coding constraints on decision making in the human brain – DECODE

Making decisions, from simple perceptual judgments to complex policy-making orientations, often requires to combine multiple pieces of ambiguous or conflicting information. In such conditions, human choices exhibit a suboptimal variability whose origin remains poorly understood. Dominant theories attribute this choice variability to a mixture of cognitive biases and random noise in an otherwise 'normative' decision process - defined by the Bayes theorem of probabilistic inference. My recent research has taken a different perspective on the issue, by seeking to understand the suboptimality of human decision making from its underlying neurobiology. Here I propose a novel research program which takes this theoretical idea further, by assessing the degree to which choice variability originates from computational constraints imposed by the neural coding of information in the brain. I hypothesize that the recent evolutionary pressure for cognitive flexibility has shaped the human brain to approximate Bayesian inference using a network of parietal and prefrontal circuits capable of dealing with information and decision problems of different kinds. To test this hypothesis, I plan to develop and apply: 1. an original framework for model-based neuroimaging of decision making, 2. a versatile paradigm for studying probabilistic inference and decoding its latent neural correlates in electromagnetic brain signals, and 3. a computational model of choice cast in terms of approximate Bayesian inference. At the methodological level, the key novelty of this proposal rests on relating variability in neural activity and choice behavior through a theoretical model to decompose decision making in terms of effective algorithms implemented in the human brain. By answering several research questions using this approach, the project will shed light on the neural constraints which shape virtually every decision, but also characterize selective alterations observed in psychiatric disorders.