Human Capital Multidisciplinary Research Center

5.1.1 Neurocognitive Decision-making Mechanisms

Boris Gutkin

Project period


Context of Research Project within a Subject of Human Capital

The relevance of the decision-making research in relation to the outstanding discoveries of the international, Russian and Soviet schools on reinforcement learning, operant conditioning, motivation, emotion and functional systems theory is related to the attempt to integrate the outstanding behavioral discoveries of the past into a single system based on the analysis of brain response using modern multimodal methods as well as modern computational theory based on reinforcement learning algorithms. Improved understanding of decision-making mechanisms in the context of increasing dynamic information flow and various social contexts will contribute to complementing and improving theoretical models of decision-making and will open new horizons in developing methods for correcting maladaptive decisions.  Ultimately, this research will contribute to a better understanding of the mechanisms of human potential formation.

The research project Neurocognitive Decision-making Mechanisms is aimed at developing the latest ideas about decision-making mechanisms. It focuses on the study of the role of social influence, cooperation, punishment and other social contexts and norms in decision-making

Project Aim

Developing new approaches and methods in neurocomputational studies of neuromechanisms of decision-making in various social contexts

Project Objectives:

  1. Developing theoretical and methodological justifications to create a new paradigm of decision-making taking into account the brain mechanisms of learning and plasticity, algorithms aimed at the identified patterns of brain activity, ways of integrating cognitive and neurocognitive methodology in the research of experimental neuroeconomics, conducting experimental work on the study of the neurobiological foundations of cooperation, taking into account the role of social contexts and social norms
  2. Modeling decision-making depending on reward scales, the influence of individual psychological and cognitive characteristics in financial markets, the dynamics of decision-making depending on individual financial condition, developing dynamic models of learning through reinforcement taking into account the energy allostasis of the body, conducting behavioral research of learning decision-making depending on changes in rewards, their probability and risk level
  3. Studying the brain mechanisms of learning during decision-making depending on changes in rewards, their probability and risk level, individual financial condition
  4. Developing new approaches and methods of neurocomputational studies of neuromechanisms of decision-making in various social contexts, introducing approaches to instrumental multimodal research using computational models

Key Findings


The theory of decision-making has been developed and supplemented, theoretical and methodological justifications for modeling decision-making in conditions of dynamic changes in rewards and their probabilities have been developed.  Algorithms for decision-making and learning strategies of behavior in continuous time and dynamically changing environment depending on the physiological needs and energy balance of the body have been put forth. A model of a financial multi-agent system (Agent-Based Models, ABM) has been created based on cognitive and neurocognitive reinforcement learning theory


The interface for interaction of the researcher with the model of a financial multi-agent system and experiments has been developed, bounded rationality traits of individual agents (i.e. distortion of the value function) have been implemented and the impact of such deviations on market dynamics has been analyzed.

A study on the distortion of the confirmation of choice has been completed. It is shown that it remains significant when the autocorrelation of choice is included in the computational model of decision–making, thereby proving that the distortion of choice confirmation is a stable feature of human reinforcement learning.  Additional computational processes that can play an important role in the assessment and interpretation of decision-making distortions have been formulated


Based on the developed theoretical framework, several studies of the mechanisms and processes of decision-making depending on changes and properties of the environment have been carried out: using the MIX model (a model of independent contribution of magnitudes and probabilities of remuneration) as a computational algorithm for elections in an uncertain and volatile environment, potential mechanisms of adaptive behavior at two levels of environmental variability have been identified.

It has been revealed that the mechanism of independent weighing of the evaluation of reward probabilities in comparison with the value of utilities is a minimal and sufficient mechanism of adaptation to environmental changes. It is shown that the dependence on the probability of remuneration in comparison with the value of utilities depends in a complex way on the amount of remuneration and the presentation of rewards as wins or losses. In particular, the probability weight was higher when the reward was large, only when the rewards offered were presented as winnings.

The parameters of the MIX model, estimated on the basis of human choices depending on the level of variability of the environment (low or high) and the presentation of the proposed rewards as wins (green) or losses (red). Asterisk (*) and tilde (~) signs indicate statistically significant differences and trends in parameter values

Educational Programs

Integrated track Brain and Cognitive iBrain Sciences

Microdigree Applied Cognitive Neuroscience


  1. A. Ghambaryan, B. Gutkin, V. Klucharev, E.Koechlin  Additively Combining Utilities and Beliefs: Research Gaps and Algorithmic Developments // Frontiers in Neuroscience, 2021
  2. Palminteri S, Lebreton M. Context-dependent outcome encoding in human reinforcement learning. // Current Opinion in Behavioral Sciences, Vol. 41, 2021, Pages 144-151


International School of Social Neuroscience Intersubjective Correlation Analysis of fMRI data: Practical Application (RU) (Moscow, Russia, June 21-23, 2021)