Neurobiologically Validated Models of Errors in Decision-Making: Strategies for Remediation and Detection
Final rept. 1 Oct 2011-30 Sep 2014
NEW YORK UNIV NY
Pagination or Media Count:
Despite the predictions of normative choice theories, human decision-making often demonstrates striking inefficiencies. In this proposal, we develop a decision-making framework derived from neurophysiologically-documented computational algorithms. Our specific aims were to 1 characterize the effect of normalization in value circuits using computational models of decision-making, 2 examine whether human decision-making inefficiencies match these predictions, and 3 develop and test normalization-based remediation strategies to reduce inefficient decision-making. For this final report, we report significant findings, published work, and work in submission for Aims 1-2 and continuing progress for Aim 3. We have developed a mathematical model of stochastic choice predicting significant, novel context-dependent inefficiencies arising from normalized value coding, occurring when the value of an irrelevant distracter or number of distracters is manipulated. Furthermore, this context-dependence is biphasic, consistent with an interplay between normalization and stochastic representations in the decision process. Two experimental studies in human subjects confirm normalization model predictions in both trinary choice and set size paradigms one published paper, one manuscript in preparation. In addition, we have developed a neuroeconomic choice model combining the normalization model and a general theoretical framework for the neural decision process this model describes the constraints placed on the decision-making process by our neurophysiology, and demonstrates that the divisive normalization computation implements choice behaviour that is optimal given these constraints. Finally, analysis is ongoing of completed normalization-based remediation experiments aiming to reduce these constraint-induced inefficiencies.
- Anatomy and Physiology
- Statistics and Probability