Reward Value Coding Distinct From Risk Attitude-Related Uncertainty Coding in Human Reward Systems

Department of Physiology, Development and Neuroscience, University of Cambridge, Cambridge, England, United Kingdom
Journal of Neurophysiology (Impact Factor: 2.89). 03/2007; 97(2):1621-32. DOI: 10.1152/jn.00745.2006
Source: PubMed


When deciding between different options, individuals are guided by the expected (mean) value of the different outcomes and by the associated degrees of uncertainty. We used functional magnetic resonance imaging to identify brain activations coding the key decision parameters of expected value (magnitude and probability) separately from uncertainty (statistical variance) of monetary rewards. Participants discriminated behaviorally between stimuli associated with different expected values and uncertainty. Stimuli associated with higher expected values elicited monotonically increasing activations in distinct regions of the striatum, irrespective of different combinations of magnitude and probability. Stimuli associated with higher uncertainty (variance) elicited increasing activations in the lateral orbitofrontal cortex. Uncertainty-related activations covaried with individual risk aversion in lateral orbitofrontal regions and risk-seeking in more medial areas. Furthermore, activations in expected value-coding regions in prefrontal cortex covaried differentially with uncertainty depending on risk attitudes of individual participants, suggesting that separate prefrontal regions are involved in risk aversion and seeking. These data demonstrate the distinct coding in key reward structures of the two basic and crucial decision parameters, expected value, and uncertainty.

Download full-text


Available from: Raymond J Dolan, Apr 03, 2015
21 Reads
  • Source
    • "This secondary result provides additional correlational evidence for a relationship between the time allotted to a decision and the risk preference exhibited for the decision. Other studies have also found that participants respond faster to higher value decisions (e.g., Madan et al., 2012; Shenhav & Buckner, 2014; Tobler et al., 2006). This connection between reward value and RT could also be thought of as a potentiation of motivated movements (Madan, 2013). "

    Journal of Cognitive Psychology 08/2015; DOI:10.1080/20445911.2015.1055274 · 1.20 Impact Factor
  • Source
    • "It is possible that different activation patterns in the two regions bias the weight of the aspects of gamble information and subsequently result in different choices. In addition , as the caudate is associated with a broad range of information processing in a gamble task such as anticipatory risk (Preuschoff et al., 2006; Venkatraman et al., 2007), reward magnitude, expected value (Kable and Glimcher, 2007; Tobler et al., 2007), and action selection (Mullette-Gillman et al., 2011), one interpretation of our result is that the caudate integrates various information, including pre-stimulus endogenous fluctuations, to generate a choice tendency between risk gamble and certainty outcome. Finally, after the gamble information appeared , the right NAcc was found to reflect the gamble choices, suggesting that people were anticipating the gains of a gamble after choice. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Rational decision-making models assume that people resolve an economic problem based on its properties and the underlying utility. Here we challenge this view by examining whether pre-stimulus endogenous neuronal fluctuations can bias economic decisions. We recorded subjects' pre-stimulus neural activity patterns with fMRI before they chose between pairs of certain outcomes and risky gambles. Our results indicate that activities in the left nucleus accumbens and medial frontal gyrus can bias subsequent risky decision making, showing that neuronal activities in regions associated with uncertainty and reward processing are involved in biasing subsequent choice selection. This finding challenges theories which propose that choices merely reveal stable underlying distributions of hedonic utility. Endogenous brain states of this sort might originate from a systematic cause or a stochastic type of neural noise, which can be construed as contextual factors that shape people's decision making.
    NeuroImage 07/2014; 101. DOI:10.1016/j.neuroimage.2014.07.036 · 6.36 Impact Factor
  • Source
    • "During decision processing, neuroimaging methods, such as functional magnetic resonance imaging (fMRI), have identified the role of prefrontal areas (such as the dorsolateral, lateral and ventromedial prefrontal cortex, orbitofrontal and the anterior cingulate cortex) and subcortical areas (such as the striatum and the amygdala) in encoding the two main components of a subjective value signal, i.e., its risk probability and expected value (Fukunaga et al., 2012; Li et al., 2010; Lighthall et al., 2012; Rao et al., 2008; Schonberg et al., 2012). Typically, increasing blood-oxygen-level-dependent (BOLD) signals in these areas encode risk with respect to uncertainty, variance, or volatility (Huettel et al., 2006, 2005; Knutson et al., 2005; Preuschoff et al., 2006; Tobler et al., 2007) and encode expected value with respect to magnitude, probability, and their combination (Knutson et al., 2005; Preuschoff et al., 2006; Tobler et al., 2007). "
    [Show abstract] [Hide abstract]
    ABSTRACT: The study measured cortical hemodynamic signals and peripheral correlates of decision makers during a dynamic risky task, the Just One More task (JOM), in which the risky decision entails choosing whether to incrementally increase accumulated earnings at the risk of ruin (going bust ending up with nothing). Twenty subjects participated in multiple instantiations of this task in which the probability of ruin and size of the stakes varied. Physiological correlates were simultaneously quantified by functional near-infrared spectroscopy (fNIRS) over dorsolateral prefrontal cortex (DLPFC) and electrodermal activity (EDA). First, in the task decision phase (i.e., when subjects are contemplating options before making a choice) probability of ruin had a dissociating effect on fNIRS and EDA. FNIRS derived DLPFC hemodynamic signals reflected a subjective value signal, correlating positively with individual risk attitude. Contrary, EDA reflected the probability of ruin in terms of a common affective measure, irrespective of individuals’ risk attitude. Second, during the task outcome phase (i.e., the time after subjects have made a choice and observed the outcomes) fNIRS and EDA revealed opposite patterns. While fNIRS derived DLPFC hemodynamic signals were larger in response to gains, EDA signals were larger in response to losses; both patterns were statistically independent of individual risk attitude. Lastly, fNIRS derived DLPFC hemodynamic signals in the decision phase correlated positively with the mean round earnings, providing a measure of the quality of individual decision-making performance. Together with the positive correlation with individual risk attitude, our findings indicate that fNIRS signals, but not EDA, could be taken as a useful method for studying individual risk attitude and task performance in dynamic risky decision-making.
    Brain research 04/2014; 1557. DOI:10.1016/j.brainres.2014.02.013 · 2.84 Impact Factor
Show more