Personal core values modulate risky choice evaluation and subsequent risk taking behavior: an fMRI study

Abstract

Evaluated the associations between two disconnected topics ,i.e., personal values and risk-taking behaviors, by both behavioral and BOLD fMRI responses. ROI approach, functional connectivity, and graph theory analysis were implemented. Security and hedonism were found to serve as an accelerator and a brake, respectively, with respect to risky behaviors.

Full text

dACC
References
1. Schwartz, S. H. (2012). An overview of the Schwartz theory of basic values. Online Readings in Psychology and Culture, 2(1), 1-20.
2. Brosch, T., & Sander, D. (2013). Neurocognitive mechanisms underlying value-based decision-making: from core values to economic
value. Frontiers in human neuroscience,7, 398.
3. Goh, J. O., Su, Y.-S., Tang, Y.-J., McCarrey, A. C., Tereshchenko, A., Elkins, W., & Resnick, S. M. (2016). Frontal, Striatal, and Medial
Temporal Sensitivity to Value Distinguishes Risk-Taking from Risk-Aversive Older Adults during Decision Making. Journal of Neuro-
science, 36(49), 12498-12509.
Acknowledgements
This work was supported by Taiwan Ministry Science and Technology grants 103-2410-H-002-082-MY2, 105-2420-H-002-002-MY2, and
105-2410-H-002-055-MY3.
Correspondence
Yun-Shiuan Chuang: yunshiuan.chuang@gmail.com
Yu-Shiang Su: yushiangsu@gmail.com
Joshua Oon Soo Goh: joshuagoh@ntu.edu.tw
Brain and Mind Laboratory
http://gibms.mc.ntu.edu.tw/bmlab/
Graduate Institute of Brain and Mind Sciences
National Taiwan University College of Medicine
Section 1, No. 1 Ren-ai Rd, 15th oor, Rm 1554,
Zhongzheng District, Taipei 10051, Taiwan
Personal core values modulate risky choice evaluation and
subsequent risk-taking behavior: an fMRI study
Yun-Shiuan Chuang1,2,3, Yu-Shiang Su2, Joshua Oon Soo Goh1,2,3
1. Department of Psychology, National Taiwan University.
2. Graduate Institute of Brain and Mind Sciences, College of Medicine, National Taiwan University.
3. Neurobiological and Cognitive Science Center, National Taiwan University.
Introduction
1. Personal core values motivate many of our decision be ha vi-
ors1,2.
2. Using fMRI, we investigated neural correlates of individual
diffe rences in hedonism and security that accelerate and
brake risk-taking behavior, respectively.
3. We hypothesized that the effect of individual core values
should be modulated by contextual factor, being more
dominant when outcomes are less important.
4. Also of interest were individual differences in information
exchanges between motivational, affective, and regulatory
neural circuits in reecting these two core value dimensions
during decisions, explored via a graph thoery measure of
functional connectivity.
Methods and Materials
Participants
• N = 40 (16 male and 24 female); Age = 23.1 ± 2.0 years.
Lottery Choice Task3
Trial Conditions
Schwartz Value Survey
Value Denition Sample Items
Hedonism Pleasure and sensuous
gratication for oneself
Pleasure, Enjoying life,
Self-indulgent
Security
Safety, harmony and
stability of society, of rela-
tionships, and of self
National security,
Reciprocation of favors,
Family security, Social order
Behavioral Data: GLMM with MCMC stimulation4
1
Overview of core values and
risk-taking behaviors
1(Cor.)Mean (SD)
1.Hedonism ─ -.01 (1.39)
2.Security -.07 -.08 (0.67)
Overall Acceptance Rate: Mean(SD)
H0.03 (0.08) 0.05 (0.13) 0.54 (0.25) 0.97 (0.07) 0.96 (0.08)
M0.02 (0.04) 0.05 (0.08) 0.42 (0.23) 0.97 (0.08) 0.97 (0.07)
L0.11 (0.19) 0.15 (0.23) 0.63 (0.22) 0.96 (0.10) 0.92 (0.18)
LL ML MM MH HH
Winning Probability
Magnitude
Results
fMRI acquisition and pre-processing
• Five runs of task echo-planar imaging: 218 volumes; TR=2s; TE=24ms;
38 axial slices; 64x64 matrix, 3.4275 x 3.4375 x 4mm.
Preprocessing for GLM analysis
• Time-slicing, realignment, co-registration, normalization, spatial smooth-
ness*
Extra Preprocessing Steps for Degree Centrality (DC)
• De-trend, bandpass, global signal removal, deep WM/ CSF signal re-
moval, and stimulus-induced signals removal.
*Note: no spatial smoothness for DC preprocessing
GLM analysis
Winning Probability (%)
Magnitude
(points)
LL ML MM MH HH
4-15 24-35 44-55 64-75 84-95
High
(101-115) LL_H ML_H MM_H MH_H HH_H
Medium
(51-65) LL_M ML_M MM_M MH_M HH_M
Low
(1-15) LL_L ML_L MM_L MH_L HH_L
Degree Centrality (DC): fcMRI measure5
GLM: Y = β0 + Choice(β10+β11Prob + β12Mag + β13(Prob*Mag))
+ Accept(β20+β21Win+β22Loss)
+ Reject (β30+β31Miss+β32Dodge)+Cov.+ ε
GLM: β13(first) = β0 + β1Security + β2Hedonism + Gender + ε
Responseit ~ Bern (pit )
Logit(pit) =
γ0i + β0 + β1 Securityi + β2 Hedonismi + γ3i Probit + γ4i Magit+ γ34i (Probit * Magit)+
β13 (Securityi * Probit) + β14 (Securityi * Magit)+β134 (Securityi * Probit * Magit )+
β23 (Hedonismi * Probit) + β24 (Hedonismi * Magit)+β234 (Hedonismi * Probit * Magit)
+
β4 Genderi + β5 IQi + εit
*Note: = 50; Burn-in =100,000; Iteration = 400,000
5Degree Centrality (DC): Core values effect on degree of information exchanges
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Probability of acceptance
Low Magnitude Trials
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Medium Magnitude Trails
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Security
High
Low
High Magnitude Trials
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Probability of acceptance
Low Magnitude Trials
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Medium Magnitude Trails
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 25 50 75
Probability(%)
Hedonism
High
Low
High Magnitude Trials
*Interaction of Hehonism x Prob. x Mag. , pMCMC < .05
Hedonism enhances risk-taking
with weaker magnitude
2
GLM: Low-Magnitude enhanced core values effect on probability sensitivity
4
Frontoparietal network Striatum
Anterior Cingulate Cortex
Security accelerates rejection and impedes risk-taking
3
1000
1500
2000
2500
0 25 50 75
Probability(%)
Response Time (ms)
Acceptance
1000
1500
2000
2500
0 25 50 75
Probability(%)
Security
High
Low
Rejection
Response Time Acceptance Rate
*Main effect of Security , pMCMC < .05 *Main effect of Security , pMCMC < .03
*Interaction of Security x Prob. , pMCMC < .05 *Interaction of Security x Resp., pMCMC < .01
*Note: β13 is the interested magnitude-modulated probability sensitivity.
*Note: β1 and β2 are interested core value effect.
L IPL L DLPFC
L VLPFC
x =
-
41 x =
-
5
R Caudate
y = 18z = 11
R aINS
Hedonism
(Positive)
Security
(Positive)
SMA
x = 7
y = 1z = 4
Lateral habeluna (LHb)
x =
-
8y =
-
9
z = 14
L Put & L Pal
y =
-
1z =
-
4
dACC
vACC
x = 4
Striatum Anterior Cingulate Cortex
L SPG
x =
-
23
L DLPFC
x =
-
44
Both core values show magnitude-modulated effect in right DLPFC regions
a
Hedonism specic magnitude-modulated effect
b
Both core values link to increased DC of SMA and insula
a
Hedonism specically
links to decreased DC of LHb
b
Security specically
links to increased DC of striatum and ACC
c
Security specically
links to decreased DC of PFC and parietal region
d
y = 35
Conclusion
• Derive functional connectivity MRI (fcMRI) time-series from task-
based runs, which was then used to generate the DC map.
• DC is a graph theory measure. It captures the extent of the intrin-
sic connectivity (or, intrinsic information ows) between a voxel
and the whole brain voxels.
• DC map generation:
The fcMRI time-series of whole brain voxels were used to gene-
rate Pearson’s R correlation matrix per each voxel. For each
voxel, strong edges (r >.25) are summed and z-transformed,
ending up with a whole brain degree centrality map.
1. Common neural loci of core value inuence during decision-making include regulatory and salience regions.
a. Higher hedonism and security enhanced R DLPFC probability sensitivity during low magnitude trials, reecting
higher affective regulation during affordable trials, particularly for individuals with stronger risk-relevant values.
b. Higher hedonism and security enhanced DC of insula and SMA during the LCT task, reecting higher salience
modulation for individuals with stronger risk-relevant values.
2. Different core values differentially dissociate affective, motivation, and regulatory circuit decision functions.
a. Hedonism enhanced low magnitude frontoparietal, ACC, and striatal probability sensitivity and decreased LHb
DC, consistent with increased subjective value and lower aversion in these individuals.
b. Security reduced frontoparietal but increased ACC and striatal DC, consistent with overall attentional disengage-
ment and increased behavioral gating associated with increased rejection rates.
3. Our ndings support the role of personal core values in risk-taking evaluation and behaviors, having implications for
understanding the neural bases for individual differences in human value-based decisions.
x = 42
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Low Magnitude
Medium Magnitude
High Magnitude
−3 −2 −1 0 1 2 −3 −2 −1 0 1 2 −3 −2 −1 0 1 2
−0.02
0.00
0.02
0.04
0.06
Hedonism
Parameter Estimate of Prob. Sensitivity (a.u.)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
Low Magnitude
Medium Magnitude
High Magnitude
−1 0 1 2 −1 0 1 2 −1 0 1 2
−0.04
−0.02
0.00
0.02
0.04
Security
Parameter Estimate of Prob. Sensitivity (a.u.)
Parameter Estimate of
Conditional
Prob. Sensitivity (a.u.)
y = 34
Hedonism
(Low-Mag. enhaced) Security
(Low-Mag. enhaced)
• Statistical thresoholds are set as pFDR < .05, derived from random field
theory using MCMC, which correspond to punc. < .005 and k > 20.
First level
Second level
4. Jaeger, T. F. (2008). Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed
models. Journal of memory and language, 59(4), 434-446.
5. Buckner, R. L., Sepulcre, J., Talukdar, T., Krienen, F. M., Liu, H., Hedden, T., et al. (2009). Cortical hubs revealed by
intrinsic functional connectivity: mapping, assessment of stability, and relation to Alzheimer’s disease. Journal of Neuro-
science, 29(6), 1860-1873.
R DLPFC (54, 32, 23)
p < .01 n.s.
p < .08 n.s. n.s.
p < .08
Parameter Estimate of
Conditional
Prob. Sensitivity (a.u.)
Personal
Core values
− Potential
Loss
＋Potential
Gain
Perceived
Loss
Perceived
Gain
Acceptance
or
Rejection
Objective
Expected Value
Subjective
Economic Value
Decision
Behavior
Contextual
Factor
Magnitude
Gambled
ITI ITIChoice Outcome
4s 1 / 3 / 5s 2s
ITI
48 points
Win 71% Loss 21%
Accept Reject
Accept
+ 48 points
148 points
−48 points
52 points
＋
Win
Lose
Reject
(+ 48 points)
100 points
(−48 points)
100 points
Miss
Dodge
Low Magnitude Trials Medium Magnitude Trials High Magnitude Trials
Low Magnitude Trials Medium Magnitude Trials High Magnitude Trials
Acceptance Rejection
Probability(%) Probability(%) Probability(%)
Probability(%) Probability(%) Probability(%)Probability(%)Probability(%)