Flexible modulation of risk attitude during decision-making under quota
Introduction
Risk attitude is one of the critical factors that intensely bias one's decision-making. Although numerous studies have examined the decision-making process associated with risk (i.e., variance of potential outcome), how an individual arranges his/her risk attitude has remained elusive.
Past studies claimed two contrasting perspectives regarding risk attitude. A large body of neuroscience, psychology, and economics studies has considered risk attitude as a static parameter that is implemented in the individual brain (Huettel et al., 2006, Tobler et al., 2009, Symmonds et al., 2011, Takahashi, 2012). From this ‘individual-preference’ view, a person taking risky options frequently was regarded as a risk-prone individual, whereas a person who tended to avoid risky choices was classified as a risk-averse individual. On the other hand, ethological studies reported a systematic change of risk attitude associated with the internal resource state, such as hunger (Caraco et al., 1980, Stephens and Krebs, 1986, St Onge and Floresco, 2009, Inagaki et al., 2014). In line with this ‘strategy optimization’ view, an individual needing to satisfy a resource shortage strategically takes risky options irrespective of individual preference. Although the respective ‘individual preference’ aspect and ‘strategy optimization’ aspect concerning risk attitude were investigated and confirmed separately, just how these two views are integrated remains poorly understood.
We assumed that the ‘quota severity’ concept could be of help for this issue. Imagine that you are the coach of a soccer team and you need to decide on a game strategy. When your team plays a fun match (i.e., the result is immaterial), you can apply a favorite strategy based on your individual risk preference, because your team has no quota constraint. On the other hand, when your team plays a tournament game (i.e., winning the game is essential), you need to optimally utilize an expected-value-based strategy (e.g., strengthening the defense) and risky strategy (e.g., increasing the number of attackers) depending on the quota severity (i.e., number of goals needed to win) in order to maximize the chance to win. This example implies that the human agent flexibly utilizes two types of risk attitudes (i.e., ‘individual-preference mode’ and ‘strategy-optimization mode’) depending on the quota severity in real-life situations.
Several brain structures have been implicated in risk-related decision-making. The ventro-medial prefrontal cortex (vmPFC) is known to mediate encoding of subjective value (Kable and Glimcher, 2009, Euston et al., 2012, Levy and Glimcher, 2012), and its dysfunction leads to the deficit of economic decision-making (Bechara et al., 1997, Rushworth et al., 2011). The striatum has been also known to encode reward signal and is emphasized in value-based decision-making (O'Doherty, 2004, Schultz, 2015). The dorsal anterior cingulate cortex (dACC) and anterior insula (AI), which have been referred to, by reflecting subjective arousal, as the ‘salience network’ (Ridderinkhof et al., 2004, Craig, 2009, Singer et al., 2009, Eisenberger, 2012), were also implicated in human risk-sensitive decision-making (Preuschoff et al., 2008, Christopoulos et al., 2009). These structures thus could play a crucial role during risk-related decision-making under quota. We also expected the involvement of the dorso-lateral prefrontal cortex (dlPFC), which has been implicated in cognitive flexibility (Milner, 1963, Wallis et al., 2001, Buckley et al., 2009), because further flexibility would be required to utilize multiple types of risk attitudes.
In this study, we performed an fMRI experiment with a novel gambling task that requires participants to solve a quota (required units to clear a task), to investigate the neural substrates of flexible utilization of risk attitude depending on the quota severity. Because the accompanying activation of dlPFC and other areas such as dACC has been reported in a variety of cognitive tasks including decision-making tasks (Duncan, 2010, Duncan, 2013), we focused not only on the activation level but also the functional-connectivity level and aimed to elucidate how the interplay of brain areas serves to utilize ‘individual-preference mode’ and ‘strategy-optimization mode’.
Section snippets
Participants
Twenty-nine task-naive participants (all male, aged 20–54 years, mean ± SD = 30.9 ± 10.4 years, 27 right-handed) participated in the current study. None of them had a history of neurological or psychiatric disorders according to a pre-interview based on DSM-IV-TR. We tested only male participants in order to use the same data set in a study of gambling disorder, which is highly prevalent in the male population (Shaffer and Martin, 2011). All participants provided written informed consent. This study
Computational simulation
Prior to the experiment with participants, a computational simulation was run in order to define the five ‘quota conditions’, the criteria used for behavioral and neuronal analyses. In the simulation, four computer-generated models (EV, RISKY, SAFE, RAND) played the GIG task (Fig. 1A), and performance of each model was recorded. When we calculated the average acquired units of each model, the EV model (expected-value-based choice model: always take a higher expected-value option) yielded the
Discussion
In the present study, we reported the influence of quota severity on risk attitude using a newly developed gambling task. Behavioral analysis revealed that the participants flexibly utilized individual risk preference and state-dependent optimal strategy in a single session. The fMRI analysis showed neural correlates of quota severity in the activation patterns. Notably, functional-connectivity patterns exclusively reflected the process of optimal strategy execution, demonstrating neural
Conclusion
Our results illustrated the human ability to utilize different decision-making modes under quota, as well as their neural substrates. Quota-dependent risk attitude potentially explains various phenomena concerning risk-sensitive decision-making in real-life situations. We are hopeful that our findings might contribute to a better understanding of psychiatric disorders or social conditions with altered risk-sensitive decision-making.
Acknowledgments
This study was conducted using the MRI scanner and related facilities of Kokoro Research Center, Kyoto University. We are greatly indebted to Dr Takanori Kochiyama, Research Advisor at ATR-Promotions Inc., for his technical advice regarding SPM analyses. We also thank the staffs of the Department of Psychiatry, Kyoto University, for expertise and discussion. This work was supported by grants-in-aid for scientific research A [24243061], and for Innovative Areas [23120009, 16H01504], from the
References (55)
- et al.
fMRI of thermal pain: effects of stimulus laterality and attention
NeuroImage
(2002) - et al.
An empirical demonstration of risk-sensitive foraging preferences
Anim. Behav.
(1980) AFNI: software for analysis and visualization of functional magnetic resonance neuroimages
Comput. Biomed. Res.
(1996)The multiple-demand (MD) system of the primate brain: mental programs for intelligent behaviour
Trends Cogn. Sci.
(2010)The structure of cognition: attentional episodes in mind and brain
Neuron
(2013)- et al.
The role of medial prefrontal cortex in memory and decision making
Neuron
(2012) - et al.
Psychophysiological and modulatory interactions in neuroimaging
NeuroImage
(1997) - et al.
Modeling regional and psychophysiologic interactions in fMRI: the importance of hemodynamic deconvolution
NeuroImage
(2003) - et al.
Gambling disorders
Lancet
(2011) - et al.
Neural signatures of economic preferences for risk and ambiguity
Neuron
(2006)
Independent, reciprocal neuromodulatory control of sweet and bitter taste sensitivity during starvation in Drosophila
Neuron
The neurobiology of decision: consensus and controversy
Neuron
Multiple neural mechanisms of decision making and their competition under changing risk pressure
Neuron
The root of all value: a neural common currency for choice
Curr. Opin. Neurobiol.
A generalized form of context-dependent psychophysiological interactions (gPPI): a comparison to standard approaches
NeuroImage
Flexible modulation of network connectivity related to cognition in Alzheimer's disease
NeuroImage
Risk-sensitive foraging: a review of the theory
Bull. Math. Biol.
Reward representations and reward-related learning in the human brain: insights from neuroimaging
Curr. Opin. Neurobiol.
The neural bases of cognitive processes in gambling disorder
Trends Cogn. Sci.
Frontal cortex and reward-guided learning and decision-making
Neuron
The expected value of control: an integrative theory of anterior cingulate cortex function
Neuron
A common role of insula in feelings, empathy and uncertainty
Trends Cogn. Sci.
Dynamic coding for cognitive control in prefrontal cortex
Neuron
Do humans make good decisions?
Trends Cogn. Sci.
Deconstructing risk: separable encoding of variance and skewness in the brain
NeuroImage
Monoamines and assessment of risks
Curr. Opin. Neurobiol.
Striatal connectivity changes following gambling wins and near-misses: associations with gambling severity
Neuroimage Clin.
Cited by (7)
Framing effects on financial and health problems in gambling disorder
2020, Addictive BehaviorsCitation Excerpt :Continued gambling is considered to reflect risky decision-making in the financial domain. While it has long been argued whether the individual risk preference consists of a unitary component or a multidimensional component (Slovic, 1964), in recent years a growing body of experimental evidence has indicated that it varies with contexts of choice (Vries, Holland, & Witteman, 2008; Figner & Weber, 2011; Fujimoto & Takahashi, 2016; Fujimoto et al., 2017; MacGregor, Slovic, Dreman, & Berry, 2000; Rottenstreich & Hsee, 2001). In addition, the contexts include which domain is being targeted, and it has been reported that the individual risk preference varies across domains (Figner & Weber, 2011; Hanoch, Johnson, & Wilke, 2006; Weber, Blais, & Betz, 2002).
How Goals Erase Framing Effects in Risky Decision Making
2021, Proceedings of the 43rd Annual Meeting of the Cognitive Science Society: Comparative Cognition: Animal Minds, CogSci 2021Prospect Theory and Optimal Risky Choices with Goals
2020, Proceedings for the 42nd Annual Meeting of the Cognitive Science Society: Developing a Mind: Learning in Humans, Animals, and Machines, CogSci 2020Salience-driven value construction for adaptive choice under risk
2019, Journal of NeuroscienceDecision-making for multi-criteria optimization of process planning
2019, Mechanics and IndustryTrait and state-dependent risk attitude of monkeys measured in a single-option response task
2019, Frontiers in Neuroscience