## **Testing the Effects of Stress on Discounting, Self-Efficacy, and Executive Control** ## As part of the larger project goal to understand the mechanism through which stress effects adherence behaviors described above, the current study will test whether the three targets -- temporal discounting (preferences), self-efficacy (beliefs), and executive control/memory (constraints) -- are indeed causally affected by stress. We will therefore induce stress in the lab via four different stress induction paradigms and measure its effect on measures of the three targets identified in Study 1. **Stress induction** We will induce stress in four different ways (across subjects): the Trier Social Stress Test (TSST) induces social stress (Haushofer et al., 2013; Kirschbaum et al., 1993); the Cold Pressor Task (CPT) induces physical stress (Hines & Brown, 1932); the Centipede Game induces financial stress (Rosenthal, 1981); and hydrocortisone administration induces the neurobiological consequences of stress (high cortisol levels; e.g. Cornelisse et al., 2013). In each of the four stress studies, participants will be randomized to either the active treatment group (for example, receive hydrocortisone pill) or an appropriate control group instead of being exposed to the stressor (for example, receive placebo pill). Assessing the causal mechanisms of stress through these four different manipulations builds upon our previous work, showing that discounting is increased by hydrocortisone administration, but not the Trier Social Stress Test (Haushofer et al., 2013). **Acute and chronic stress** As an important innovation, we will capture the within-subjects effects of acute vs. chronic stress exposure by exposing subjects to the stress protocols described above once on day 1 of the study (treated as acute) and repeatedly over the course of 7 days (treated as chronic), following the protocol developed by Kandasamy et al. (2014). The stress literature has long distinguished between these two types of stress (McEwen, 2004), and it has recently been shown that risk aversion is affected by chronic but not acute stress (Kandasamy et al., 2014); however, few other studies of the effects of stress on behaviors related to adherence study chronic stress (which likely is the more relevant kind in generating low regimen adherence). **Outcome measures** We will assess the effect of stress on our targets through both psychological (questionnaires) and behavioral (incentivized computer tasks) outcome measures. These measures will be chosen through an initial round of measure development under Study 1. Each subject will be tested on the outcome measures for all of the targets, with the order of tasks counterbalanced across subjects (this will also allow us to assess the evolution of stress effects over time (Henckens et al., 2010, 2011; Joëls et al., 2011). As a manipulation check for stress induction, we will also measure salivary cortisol, positive/negative affect (PANAS) (Watson et al., 1988), and subjective stress and pain (visual analog scales).