Symmonds, M.; (2011) The neurobiology of decision making under risk. Doctoral thesis , UCL (University College London).

PDF Symmonds_Thesis_word2007_Dec2011FINAL_REDACTv2.pdf Download (10MB)

Abstract

Risk is a highly salient psychological decision variable, and sensitivity to risk is an evolutionarily ancient attribute. In this thesis I address the neurobiological foundation of risk assessment, and show that behaviour is driven by an underlying distributed neural representation of different elements of risk in the brain. In particular, I show using fMRI (in Chapter 4) and MEG (in Chapter 8) that variance (dispersion) and skewness (asymmetry) of gambles evokes anatomically separable neural responses in a parietal, prefrontal and insula cortical network. I discuss possible theoretical neurobiological mechanisms by which preferences could be imbued to choice, and show that subjective tastes for risk, in terms of behavioural sensitivity to each of these risk dimensions, influences the encoding of risk and subsequent anticipatory responses. In Chapter 5, I show that a representation of prospective outcomes several trials into the future is supported by a dissociated encoding of the statistical information of future states in medial prefrontal cortex; furthermore that this encoding is contingent upon overarching goals or constraints. In Chapter 6, I demonstrate that economic choice is highly susceptible to exogenous biological influences, namely the effect of metabolic state, whilst in Chapter 7 I provide evidence that the encoding of risk is not affected by dopaminergic disruption, suggesting that dopamine might mediate effects on risk-taking via its role in reward feedback representation. In summary, the studies in this thesis elaborate the neural mechanisms underlying how humans make both single-shot and sequential decisions under risk, central elements in decision-making scenarios ranging from foraging to financial investment. This demonstrates that phylogenetically ancient circuitry subserving valuation and reward decompose choice into their salient statistical features, enabling the sophisticated representation of the future and its alternatives.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item