Multiple timescales of temporal context in risky choice: Behavioral identification and relationships to physiological arousal

Abstract

Context-dependence is fundamental to risky monetary decision-making. A growing body of evidence suggests that temporal context, or recent events, alters risk-taking at a minimum of three timescales: immediate (e.g. trial-by-trial), neighborhood (e.g. a group of consecutive trials), and global (e.g. task-level). To examine context effects, we created a novel monetary choice set with intentional temporal structure in which option values shifted between multiple levels of value magnitude (“contexts”) several times over the course of the task. This structure allowed us to examine whether effects of each timescale were simultaneously present in risky choice behavior and the potential mechanistic role of arousal, an established correlate of risk-taking, in context-dependency. We found that risk-taking was sensitive to immediate, neighborhood, and global timescales, increasing following small (vs. large) outcome amounts, large positive (but not negative) shifts in context, and when cumulative earnings exceeded expectations. We quantified arousal with skin conductance responses, which were specifically related to the global timescale, increasing with cumulative earnings, suggesting that physiological arousal captures a task-level assessment of performance. We complimented this correlational analysis with a secondary reanalysis of risky monetary choices following the double-blind administration of propranolol and a placebo during a temporally unstructured choice task. We replicated our behavioral finding that risk-taking is context-sensitive at three timescales but found no change in temporal context-effects following propranolol administration. Our results demonstrate that risky decision-making is consistently dynamic at multiple timescales and that arousal is likely the consequence, rather than the cause, of temporal context in risky monetary decision-making.

Cognitive strategy use selectively changes temporal context effects in risky monetary decision-making

“…This research indicates that the current option values and probabilities as well as previous outcomes are important to risky choice behavior. In a recent set of studies, we found that risk-taking decreased following large positive previous outcomes 6 and replicated this effect in a subsequent novel dataset 21 .That risk-taking decreased following large previous outcomes is consistent with some previous studies 5,11 but at odds with others showing increased risk-taking 9,14,16,17 following large gains. Such inconsistencies in the literature may be the result of differences in how outcomes are realized across tasks 10 and integrated with potential future outcomes 17 .…”

Multiple timescales of temporal context in risky choice: Behavioral identification and relationships to physiological arousal

“…Below, we report how we determined the sample size, and all data exclusions, manipulations, and study measures. Data from this paper can be accessed at: https://osf.io/a7nvx/ [48]. For all data cleaning and analysis scripts: https://github.com/sokolhessnerlab/vic [49].…”

“…Data from this paper can be accessed at: https://osf.io/a7nvx/ [48]. For all data cleaning and analysis scripts: https://github.com/sokolhessnerlab/vic [49]. Data were analyzed using the package "lme4" [50] in R, version 4.0.0.…”

“…First, we estimated the extent to which current trial variables (e.g. risky gain t , in dollars, where t represents the current trial) accounted for risk-taking behavior (choice t ) by fitting a generalized linear binomial mixed effects model to the binary risky choice data (coded as 0 = safe choice, 1 = risky choice) using the package "lme4" [50] in R. For model 1, we included risky gain t , safe t, and magnitude t (which was simply the EV level of that run) because choice options on each trial were derived by the EV level plus or minus up to $2 (for full model specifications and results, see S2 Table and the study Github repository [49]; in pseudo R code: model 1 = glmer(choice(t) ~0 + risky gain amount(t) + safe amount(t) + magnitude(t) + (0 + risky gain amount(t) + safe amount(t) | Subject ID), family = "binomial")). Regression estimates revealed the expected pattern of sensitivity to current trial variables.…”

The Relationship between Physiological Arousal and Multiple Levels of Temporal Context in Risky Choice

“…Data from this paper can be accessed at: https://osf.io/a7nvx/ (Brooks & Sokol-Hessner, 2021a). For all data cleaning and analysis scripts: https://github.com/sokolhessnerlab/vic (Brooks & Sokol-Hessner, 2021b). Data were analyzed using the package "lme4" (Bates, Maechler, Bolker, & Walker, 2015) in R, version 4.0.0.…”