Brain-outcome associations for risk taking depend on the measures used to capture individual differences

Abstract

Biological markers of risk taking are prominent targets for clinical, developmental, and longitudinal research. With respect to brain function, several regions are considered central for risky choice, yet insights into the neural basis of risk taking stem primarily from studies using single measures. Considering that recent studies suggested different risk-taking measures cannot be used interchangeably, it is currently unclear whether core regions of the brain involved in risk show a measure-dependent functional dissociation. Reporting results from the imaging subsample (N = 116 young adults) of the Basel–Berlin Risk Study, we examine (1) the conjunction of average neural representations of experience-based risky choice in the Balloon Analogue Risk Task and description-based risky choice in monetary gambles, (2) the preservation of individual activation differences across the two measures, and (3) the explanatory power of the neural correlates of risky choice for behavior. Our results suggest common group-level activation increases in nucleus accumbens, inconsistent individual differences in regional activation across measures, and limited explanatory power of neural indices for behavior, within and across measures. Our findings help clarify commonalities and differences between the neural representation of experienced and described risk, and thus should inform research designs targeting individual differences in risk taking.

Convergence of Age Differences in Risk Preference, Impulsivity, and Self-Control: A Multiverse Analysis

“…Concerning the latter, our anterior insula findings are consistent with AIM predictions for age e↵ects on neural correlates of (risky) decision making [27]; relative to younger adults, older adults showed decreased activation in brain regions associated with loss anticipation, specifically the anterior insula. Confirming the crucial role of behavioral measures in brain-behavior associations [85], and by extension age e↵ects, we obtained mainly insignificant specifications for neural markers extracted from delay discounting. We did obtain a small set (n=3) of significant specifications indicating positive age e↵ects on nucleus accumbens activation di↵erences for smaller-sooner versus larger-later choices, but these were highly covariate-dependent and as such less credible than the other age e↵ects obtained from our analyses.…”

“…In the current study we aimed to address these shortcomings and performed a multiverse analysis [84] of life span trajectories of risk preference and related constructs using multiple traditional and novel elicitation methods. The current study presents a follow-up and extension of our previous work [25,31,43,46,85], combining comprehensive within-subject testing of di↵erent constructs (i.e., risk preference, impulsivity, and low self-control) and assessment methods (i.e., self report, informant report, behavioral paradigms, and biological measures) in an adult life span sample spanning from early to late adulthood. The two main questions of interest were thus: (1) Do we find age e↵ects on risk preference, impulsivity and low self-control, and do these converge given these constructs conceptual relatedness?, and (2) To what extent do age e↵ects for risk preference, impulsivity, and low self-control converge as a function of assessment method?…”

Representative Design in Psychological Assessment: A Case Study Using the Balloon Analogue Risk Task (BART)

“…Such improvements are much needed in various areas of behavioral research (cf. Frey et al, 2017;Lauriola et al, 2014;Lönnqvist et al, 2015;Millroth, Juslin, Winman, Nilsson, & Lindskog, 2020)-for instance, when investigating the functional neural architecture of risk taking in neuroimaging studies (e.g., Schonberg et al, 2012Schonberg et al, , 2011Tisdall et al, 2018).…”

“…), yet previous research has highlighted substantial psychometric limitations therein (e.g., Beauchamp, Cesarini, & Johannesson, 2017;Berg, Dickhaut, & Mc-Cabe, 2005;Eisenberg et al, 2019;Frey, Pedroni, Mata, Rieskamp, & Hertwig, 2017;Lönnqvist, Verkasalo, Walkowitz, & Wichardt, 2015;Slovic, 1962). Although valid alternatives do exist in the form of self-report measures (e.g., Steiner, Seitz, & Frey, 2020), behavioral tasks may be indispensable for applications such as studying the functional neural architecture of risk preference, which requires the simulation of risk-taking behaviors in the fMRI scanner (e.g., Helfinstein et al, 2014;Li et al, 2019;Rao, Korczykowski, Pluta, Hoang, & Detre, 2008;Schonberg, Fox, & Poldrack, 2011;Tisdall et al, 2018).…”

“…The BART (Lejuez et al, 2002) is one of the most prominent behavioral measures used to gauge individuals' risk-taking behavior, often employed in behavioral decision research (e.g., Lauriola, Panno, Levin, & Lejuez, 2014;Lejuez et al, 2002;Wallsten, Pleskac, & Lejuez, 2005), in clinical settings (e.g., Bornovalova, Daughters, Hernandez, Richards, & Lejuez, 2005;Hopko et al, 2006;Hunt, Hopko, Bare, Lejuez, & Robinson, 2005), as well as in applied contexts (e.g., Aklin, Lejuez, Zvolensky, Kahler, & Gwadz, 2005;Lejuez, Aklin, Zvolensky, & Pedulla, 2003). It has been shown to be predictive of whether or not a person is a smoker, MDMA-user, drinker, and Marijuana user (e.g., Campbell, Samartgis, & Crowe, 2013;Hanson, Thayer, & Tapert, 2014;Hopko et al, 2006;Lejuez, Aklin, Jones, et al, 2003) and has repeatedly been used to study the neural architecture of risk-taking behavior (e.g., Helfinstein et al, 2014;Li et al, 2019;Rao et al, 2008;Tisdall et al, 2018), as well as its potential ge-netic underpinnings (Mata, Hau, Papassotiropoulos, & Hertwig, 2012).…”