Individuals' risk attitudes are known to guide choices about uncertain options. However, in the presence of others' decisions, these choices can be swayed and manifest as riskier or safer behavior than one would express alone. To test the mechanisms underlying effective social 'nudges' in human decision-making, we used functional neuroimaging and a task in which participants made choices about gambles alone and after observing others' selections. Against three alternative explanations, we found that observing others' choices of gambles increased the subjective value (utility) of those gambles for the observer. This 'other-conferred utility' was encoded in ventromedial prefrontal cortex, and these neural signals predicted conformity. We further identified a parametric interaction with individual risk preferences in anterior cingulate cortex and insula. These data provide a neuromechanistic account of how information from others is integrated with individual preferences that may explain preference-congruent susceptibility to social signals of safety and risk.
This is a preview of subscription content, access via your institution
Open Access articles citing this article.
Scientific Reports Open Access 23 August 2022
Face masks increase compliance with physical distancing recommendations during the COVID-19 pandemic
Journal of the Economic Science Association Open Access 13 November 2021
Scientific Reports Open Access 12 April 2018
Subscribe to Journal
Get full journal access for 1 year
only $8.25 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Get time limited or full article access on ReadCube.
All prices are NET prices.
van de Waal, E., Borgeaud, C. & Whiten, A. Potent social learning and conformity shape a wild primate's foraging decisions. Science 340, 483–485 (2013).
Lee, I.H. Market crashes and informational avalanches. Rev. Econ. Stud. 65, 741–759 (1998).
Myers, D.G. & Lamm, H. The group polarization phenomenon. Psychol. Bull. 83, 602 (1976).
Biele, G., Rieskamp, J., Krugel, L.K. & Heekeren, H.R. The neural basis of following advice. PLoS Biol. 9, e1001089 (2011).
Izuma, K. & Adolphs, R. Social manipulation of preference in the human brain. Neuron 78, 563–573 (2013).
Klucharev, V., Hytönen, K., Rijpkema, M., Smidts, A. & Fernández, G. Reinforcement learning signal predicts social conformity. Neuron 61, 140–151 (2009).
Lohrenz, T., Bhatt, M., Apple, N. & Montague, P.R. Keeping up with the Joneses: interpersonal prediction errors and the correlation of behavior in a tandem sequential choice task. PLoS Comput. Biol. 9, e1003275 (2013).
Huettel, S.A., Stowe, C.J., Gordon, E.M., Warner, B.T. & Platt, M.L. Neural signatures of economic preferences for risk and ambiguity. Neuron 49, 765–775 (2006).
Preuschoff, K., Bossaerts, P. & Quartz, S.R. Neural differentiation of expected reward and risk in human subcortical structures. Neuron 51, 381–390 (2006).
Holt, C.A. & Laury, S.K. Risk aversion and incentive effects. Am. Econ. Rev. 92, 1644–1655 (2002).
Rangel, A., Camerer, C. & Montague, P.R. A framework for studying the neurobiology of value-based decision making. Nat. Rev. Neurosci. 9, 545–556 (2008).
Bernoulli, D. Exposition of a new theory on the measurement of risk. Econometrica 22, 23–36 (1954)[transl].
Hare, T.A., Camerer, C.F. & Rangel, A. Self-control in decision-making involves modulation of the vmPFC valuation system. Science 324, 646–648 (2009).
Chib, V.S., Rangel, A., Shimojo, S. & O′Doherty, J.P. Evidence for a common representation of decision values for dissimilar goods in human ventromedial prefrontal cortex. J. Neurosci. 29, 12315–12320 (2009).
Fan, J., Hof, P.R., Guise, K.G., Fossella, J.A. & Posner, M.I. The functional integration of the anterior cingulate cortex during conflict processing. Cereb. Cortex 18, 796–805 (2008).
Ruff, C.C. & Fehr, E. The neurobiology of rewards and values in social decision making. Nat. Rev. Neurosci. 15, 549–562 (2014).
Sip, K.E., Smith, D.V., Porcelli, A.J., Kar, K. & Delgado, M.R. Social closeness and feedback modulate susceptibility to the framing effect. Soc. Neurosci. 10, 35–45 (2014).
Somerville, L.H., Kelley, W.M. & Heatherton, T.F. Self-esteem modulates medial prefrontal cortical responses to evaluative social feedback. Cereb. Cortex 20, 3005–3013 (2010).
Zaki, J., Schirmer, J. & Mitchell, J.P. Social influence modulates the neural computation of value. Psychol. Sci. 22, 894–900 (2011).
Behrens, T.E., Hunt, L.T., Woolrich, M.W. & Rushworth, M.F. Associative learning of social value. Nature 456, 245–249 (2008).
Burke, C.J., Tobler, P.N., Baddeley, M. & Schultz, W. Neural mechanisms of observational learning. Proc. Natl. Acad. Sci. USA 107, 14431–14436 (2010).
Christopoulos, G.I. & King-Casas, B. With you or against you: Social orientation dependent learning signals guide actions made for others. Neuroimage 104, 326–335 (2015).
Christopoulos, G.I., Tobler, P.N., Bossaerts, P., Dolan, R.J. & Schultz, W. Neural correlates of value, risk, and risk aversion contributing to decision making under risk. J. Neurosci. 29, 12574–12583 (2009).
Ridderinkhof, K.R., Ullsperger, M., Crone, E.A. & Nieuwenhuis, S. The role of the medial frontal cortex in cognitive control. Science 306, 443–447 (2004).
Campbell-Meiklejohn, D.K., Bach, D.R., Roepstorff, A., Dolan, R.J. & Frith, C.D. How the opinion of others affects our valuation of objects. Curr. Biol. 20, 1165–1170 (2010).
Eliaz, K., Ray, D. & Razin, R. Choice shifts in groups: A decision-theoretic basis. Am. Econ. Rev. 96, 1321–1332 (2006).
Attanasio, O., Barr, A., Cardenas, J.C., Genicot, G. & Meghir, C. Risk pooling, risk preferences, and social networks. Am. Econ. J. Appl. Econ. 4, 134–167 (2012).
Chein, J., Albert, D., O′Brien, L., Uckert, K. & Steinberg, L. Peers increase adolescent risk taking by enhancing activity in the brain′s reward circuitry. Dev. Sci. 14, F1–F10 (2011).
King-Casas, B. & Chiu, P.H. Understanding interpersonal function in psychiatric illness through multiplayer economic games. Biol. Psychiatry 72, 119–125 (2012).
Raafat, R.M., Chater, N. & Frith, C. Herding in humans. Trends Cogn. Sci. 13, 420–428 (2009).
Burke, C.J., Tobler, P.N., Schultz, W. & Baddeley, M. Striatal BOLD response reflects the impact of herd information on financial decisions. Front. Hum. Neurosci. 4, 48 (2010).
Miller, N., Garnier, S., Hartnett, A.T. & Couzin, I.D. Both information and social cohesion determine collective decisions in animal groups. Proc. Natl. Acad. Sci. USA 110, 5263–5268 (2013).
Pfeifer, J.H. et al. Entering adolescence: resistance to peer influence, risky behavior, and neural changes in emotion reactivity. Neuron 69, 1029–1036 (2011).
Edelson, M.G., Dudai, Y., Dolan, R.J. & Sharot, T. Brain substrates of recovery from misleading influence. J. Neurosci. 34, 7744–7753 (2014).
Medic, N. et al. Dopamine modulates the neural representation of subjective value of food in hungry subjects. J. Neurosci. 34, 16856–16864 (2014).
Hsu, M., Bhatt, M., Adolphs, R., Tranel, D. & Camerer, C.F. Neural systems responding to degrees of uncertainty in human decision-making. Science 310, 1680–1683 (2005).
Kahneman, D. & Tversky, A. Prospect theory: An analysis of decision under risk. Econometrica 47, 263–291 (1979).
Sokol-Hessner, P. et al. Thinking like a trader selectively reduces individuals′ loss aversion. Proc. Natl. Acad. Sci. USA 106, 5035–5040 (2009).
Allen, J.P., Chango, J., Szwedo, D., Schad, M. & Marston, E. Predictors of susceptibility to peer influence regarding substance use in adolescence. Child Dev. 83, 337–350 (2012).
Ouimette, P.C., Finney, J.W. & Moos, R.H. Twelve-step and cognitive-behavioral treatment for substance abuse: A comparison of treatment effectiveness. J. Consult. Clin. Psychol. 65, 230 (1997).
Friston, K.J. et al. Statistical parametric maps in functional imaging: a general linear approach. Hum. Brain Mapp. 2, 189–210 (1994).
Clithero, J.A. & Rangel, A. Informatic parcellation of the network involved in the computation of subjective value. Soc. Cogn. Affect. Neurosci. 9, 1289–1302 (2014).
We thank R. Montague, T. Lohrenz and S. LaConte, and gratefully acknowledge the technical assistance of J. Lu, J. Shin, and members of the Chiu and King-Casas laboratories. This work was supported in part by the US National Institutes of Health (MH091872 and MH087692 to P.H.C. DA036017 to B.K.-C.).
The authors declare no competing financial interests.
Integrated supplementary information
Each trial started with a ‘New Round’ screen. Participants made a series of choices between two gambles, one of which was objectively riskier than the other. The decisions were made (a) alone ('Solo' trials) and also (b) with information provided about other players' selections ('Info' trials). At the start of each trial, a new pair of gambles was presented. For each pair, the safer ('safe') gamble had lower payoff variance, and the riskier ('risky') gamble had greater payoff variance (see Online Methods for full gamble details). On Info trials, two others' decisions were displayed prior to the participant’s cue to enter a choice. Info and Solo trials were intermixed with a new random order for each participant and drawn without replacement from a uniform distribution.
Eight unique paired lottery menus were used (adapted from Holt & Laury10). For each menu, high and low payoffs for the safe and risky gambles were held constant, but the probabilities varied. (a) The high and low payoffs for each menu of gambles are shown here. To avoid potential biases for or against any particular menu, for each participant, gambles from four of the eight lottery menus were randomly selected to be presented, and trial order was randomized for menu x probability x trial type with a unique order per participant. (b) For scanning, we used the 6 pairs symmetric around the mean indifference point (for each menu, gambles 40%, 50%, 60%, 70%, 80% and 90% probability of high payoff).
Supplementary Figure 3 Behavioral data from pilot and scanned participants indicating risk-aversion in a majority of participants.
(a) The gambles were piloted in N = 54 behavioral participants. These pilot data verified that the chosen payoff amounts and format were appropriate for revealing individual differences in risk preferences. As in Holt & Laury10, participants were risk averse, with the mean indifference ('switch over') point between the high payoff probabilities of 60 and 70%, to the right of risk neutral (dotted line, between 40 and 50%). (b) Individual risk preferences were estimated from Solo trial choices from the main study and indicated risk aversion (ρ < 1) for a majority of participants. (c ) Mix Info trials are those in which the others’ displayed decisions were mixed (one safe, one risky gamble displayed). Regardless of the order of the mixed information (‘Player 1 safe, Player 2 risky’, orange; or ‘Player 1 risky, Player 2 safe’, green), participants’ choices were comparable with that of Solo trials (black). Eight participants whose choices yielded no unique solution for the Solo risk preference estimates were excluded from these plots. Error bars show s.e.m.
To assess whether the observed influence of others was a social or more general information effect, we implemented a separate behavioral experiment instructing participants that ‘Info’ trials were computer-generated choices. The visual aspects and trial structure of the original game were maintained, and as in the original task, participants (N=30, no overlap with the N=70 scanned participants) chose between two gambles. Participants were instructed that on some trials (previously the ‘Info’ trials), prior to the participant’s decision, two computers would randomly pick among the options, and these two options would be presented (‘Computer Info’ trials). As in the original experiment, ‘Solo’ trials were interspersed with the Computer Info trials. No influence of computer-selected options on participants’ choices was observed (repeated measures ANOVA, F(3, 87) = 0.71, P = 0.55; paired t-tests: Safe vs Solo, t(29) = 0.61, P = 0.55; Mix vs Solo, t(29) = 0.90, P = 0.37; Risky vs Solo, t(29) = -0.52, P = 0.37). Error bars show s.e.m.
Supplementary Figure 5 Other-conferred utility (OCU) predicts asymmetric, risk preference-congruent, conformity: risk-seeking example.
The proposed other-conferred utility model hypothesizes that Uwith OCU = Usolo + OCU, where OCU is a constant increase in utility of the gamble chosen by others. Depending on the concavity (ρ) of an individual’s utility power function, Uwith OCU will exceed a limited range of the Usolo predicted gambles; those gambles for which Uwith OCU > Usolo have an increased likelihood of being chosen. The model and predictions are schematically depicted here for ρ > 1, risk seeking. The figure shows Usolo for a typical risk seeking individual (plotted here using basic power utility function U(x) = xρ and our top quintile average ρ = 1.48, black line; see Online Methods for complete model estimation details). For clarity, we limit our illustration to a single menu of uniformly distributed gambles in which the high and low payoffs for the safe and risky gambles are held constant, but the probabilities vary (per Holt & Laury10). The x-axis represents each gamble’s Usolo value (i.e., V = Usolo–1(EU), where EU = expected utility = [pgamble (Vhigh payoff)ρ + (1 – pgamble)(Vlow payoff)ρ]. The reference point (V0) is the value associated with the mean expected utility of safe gambles (V0 = Usolo–1(EUsafe)) which is flanked to the left and right by values associated with risky gambles. When others choose the safe gamble (blue arrow at V0), OCU is conferred to that option, and the resultant Uwith OCU(V0) exceeds Usolo for originally preferred gambles between V0 and V1 (indicated by the blue bar). When others choose the risky gamble (red arrow at V2, for comparison mirroring V1 and symmetric around V0), the resulting Uwith OCU(V2) exceeds Usolo for originally preferred gambles between V2 and V3 (indicated by the red bar). Notice that in all cases, OCU from risky others will have influence over a greater range of gambles (red bar) than will OCU from safe others (blue bar; simply: V3 – V2 > V1 – V0), demonstrating that for risk seeking, the influence of risky others will always exceed that of safe others. As detailed throughout, this asymmetry predicts preference-biased conformity when making decisions about uncertain options. The complementary example for risk-averse individuals is shown in the main text, Fig. 2a.
Supplementary Figure 6 Risk preference predicts conformity bias; association between distance and conformity is independent from payoff probability.
(a) Conformity bias was defined as the conformity difference between 'safe, safe' and 'risky, risky' trials. Individual risk preferences estimated from Solo trials were significantly, negatively correlated with the conformity bias (Pearson’s correlation, r = –0.46, P = 1.51e–04); risk seeking individuals conformed more when others chose the risky gamble (risky influence), and risk averse individuals conformed more when others chose the safe gamble (safe influence). Each point is an individual participant, and the red line is the regression line between risk preference and conformity bias. (b) Because choice behavior in the current task was modeled using a softmax function, distance (difference between others' choices and one's own decision probability), is correlated with probability of earning high payoff. (c) However, conformity is not correlated with probability of earning high payoff. These results demonstrate that the association between distance and conformity is independent from payoff probability. Error bars show s.e.m.
Supplementary Figure 7 Increased vmPFC response for OCU-modified variables is consistent across multiple additional variables and greatest for Uwith OCU chosen gamble - Uwith OCU unchosen gamble (pink).
It is possible that other variables than ‘Uwith OCU chosen gamble – Uwith OCU unchosen gamble (Fig. 3ai)’ show an enhanced vmPFC signal when modified by OCU, as Uwith OCU is posited to be a closer approximation of participants’ subjective value (on Info trials) than is Usolo. To address this possibility, we ran several additional analyses with and without OCU modifying several different variables (U_safe, U_risky, U_safe – U_risky, U_chosen, and U_total) to examine whether the OCU-modified utility signals show stronger neural correlates than the unmodified variables. We focused on vmPFC activity in the INFO trials because OCU modification only exists on these trials (for Solo trials, OCU = 0 as there is no information from others). Using these additional regressors, we observed that in every case, vmPFC activity for utility differences between the chosen and unchosen gambles (pink, Uwith OCU chosen gamble – Uwith OCU unchosen gamble) was greater than neural responses associated with each of the other additional regressors (gray bars; using the Fig. 3ai cluster as the ROI; paired t-tests: Urisky, t(55) = –2.87, P = 0.0059; Usafe, t(55) = –3.36, P = 0.0014 ; Urisky-Usafe, t(55) = –2.79, P = 0.0072; Utotal, t(55) = –2.98, P = 0.0043; Uchosen, t(55) = –2.08, P = 0.042). Equally important, for 4 of these 5 additional regressors, the OCU-modified variable showed at least a trend toward better performance than the un-modified variable (gray vs black bars; paired t-tests: Urisky, t(55) = –2.49, P = 0.016; Usafe, t(55) = 1.28, P = 0.21; Urisky–Usafe, t(55) = –2.03, P = 0.047; Utotal (Usafe+Urisky), t(55) = –2.22, P = 0.031; Uchosen, t(55) = –1.32, P = 0.19). The better performance for the OCU-modified variables is notable in its consistency across variables, and provides additional support for our hypothesis that others choices confer subjective value to those options. That is, under this hypothesis, we expected to see stronger vmPFC encoding of OCU-modified variables, given that Uwith OCU is posited to be a closer approximation of participants’ subjective value (on Info trials) than is Usolo alone. In addition, to test the differences between beta for Uchosen OCU – Uunchosen OCU and each of the listed other OCU-modified variables, we ran a series of paired t-tests. The vmPFC beta obtained from Uchosen OCU – Uunchosen OCU was larger than each of the other 5 OCU-modified variables at least at the trend level (each variable tested vs beta for Uchosen OCU – Uunchosen OCU, paired t-tests: Urisky OCU, t(55) = –1.68, P = 0.099; Usafe OCU, t(55) = –3.88, P = 0.00028; Urisky OCU – Usafe OCU, t(55) = –1.79, P = 0.079; Utotal OCU, t(55) = –2.22, P = 0.030; Uchosen OCU, t(55) = –1.52, P = 0.13). We note that beta for Uchosen OCU – Uunchosen OCU is not necessarily expected to be significantly better than each of the other OCU-modified variables. Specifically, other OCU-modified variables (including and beyond those tested here), may contribute with varying weight to participants’ choices. We focused here on beta for Uchosen OCU – Uunchosen OCU as the broadest OCU-modified variable likely to be encoded. It is likely (and suggested by these analyses) that individual differences exist for the types of information to which others choices confer value and that are compiled into a participant’s decisions. Unmodified variables are shown in gray, OCU-modified variable are shown in black. Error bars show s.e.m.; * P < 0.05, n.s., not significant.
Supplementary Figure 8 Neural other-conferred utility signal and preference-biased responses to direction of others’ choices are robust in the smallest subset of included participants.
To check the robustness of the results, we duplicated all analyses leaving out all participants who were excluded in any behavioral model fitting for OCU or Solo risk preferences (due to lacking a unique solution, leaving N = 49); each analysis showed comparable patterns and survived comparable statistical thresholds as for the larger sample reported in the main manuscript. (a, b) All neural findings reported the main text Fig. 3a and 3c were duplicated leaving out all participants who were excluded in any individual estimations or for scanning artifacts (leaving N=49 as the smallest subset of included participants). All displayed maps are displayed at P <.001 uncorrected, k > 10 contiguous voxels; the vmPFC and dACC clusters are significant at P < 0.02 FWE, SVC; the insula cluster is significant at P < 0.02 FWE.
Supplementary Figure 9 Risk preference-dependent conformity and neural signatures of other-conferred utility are robust to participants’ gender.
(a) Male and female participants showed no between-group differences in either risk preference (independent sample t-test, t(60) = 0.87, P = 0.39, two-tailed), conformity (independent sample t-test, t(60) = 1.20, P = 0.24, two-tailed), or conformity bias (independent sample t-test, t(60) = –0.56, P = 0.58, two-tailed). We also included gender as a covariate in the behavioral analyses to assess potential gender differences in the preference-dependent conformity predicted by the OCU model (per main text Fig. 2b). Both males and females show preference-dependent conformity biases, and gender was not a predictor of this pattern (multiple linear regression, t = –1.15, P = 0.25). (b, c) Nonetheless, to evaluate and control for potential gender effects, we also performed additional analyses adding gender as an extra nuisance regressor to the neural general linear model (GLM) analyses presented in the main text. All results with the gender covariate added to the GLM are comparable with those reported in the main text Fig. 3a and 3c. All displayed maps are displayed at P <.001 uncorrected, k > 10 contiguous voxels; the vmPFC and dACC clusters are significant at P < 0.02 FWE, SVC; the insula cluster is significant at P < 0.02 FWE; error bars show s.e.m.
About this article
Cite this article
Chung, D., Christopoulos, G., King-Casas, B. et al. Social signals of safety and risk confer utility and have asymmetric effects on observers' choices. Nat Neurosci 18, 912–916 (2015). https://doi.org/10.1038/nn.4022
This article is cited by
Scientific Reports (2022)
Face masks increase compliance with physical distancing recommendations during the COVID-19 pandemic
Journal of the Economic Science Association (2021)
Nature Neuroscience (2019)
Scientific Reports (2018)
npj Science of Learning (2017)