Risky decision-making strategies mediate the relationship between amygdala activity and real-world financial savings among individuals from lower income households: A pilot study

Abstract

Lower financial savings among individuals experiencing adverse social determinants of health (SDoH) increases vulnerabilities during times of crisis. SDoH including low socioeconomic status (low-SES) influence cognitive abilities as well as health and life outcomes that may perpetuate poverty and disparities. Despite evidence suggesting a role for financial growth in minimizing SDoH-related disparities and vulnerabilities, neurobiological mechanisms linked with financial behavior remain to be elucidated. As such, we examined the relationships between brain activity during decision-making (DM), laboratory-based task performance, and money savings behavior. Participants (N = 24, 14 females) from low-SES households (income<$20,000/year) underwent fMRI scanning while performing the Balloon Analogue Risk Task (BART), a DM paradigm probing risky- and strategic-DM processes. Participants also completed self-report instruments characterizing relevant personality characteristics and then engaged in a community outreach financial program where amount of money saved was tracked over a 6-month period. Regarding BART-related brain activity, we observed expected activity in regions implicated in reward and emotional processing including the amygdala. Regarding brain-behavior relationships, we found that laboratory-based BART performance mediated the impact of amygdala activity on real-world behavior. That is, elevated amygdala activity was linked with BART strategic-DM which, in turn, was linked with more money saved after 6 months. In exploratory analyses, this mediation was moderated by emotion-related personality characteristics such that, only individuals reporting lower alexithymia demonstrated a relationship between amygdala activity and savings. These outcomes suggest that DM-related amygdala activity and/or emotion-related personality characteristics may provide utility as an endophenotypic marker of individual's financial savings behavior.

Keywords: Alexithymia; Amygdala; Balloon Analogue Risk Task; Decision-making; Financial savings; Social determinants of health; Socioeconomic status; fMRI.

Fig. 1.

Modified BART. On each trial, participants were instructed to reflect on the task rules during the “think phase” (average duration: 5 s, range: 2–6 s). This phase was followed by a “pump phase” (duration determined by participants’ behavior) when participants were instructed to inflate a virtual balloon using a MRI compatible scroll device. A “wait phase” (average duration: 3 s, range: 1–5 s) then preceded an “inflate phase” (average duration: 4 s range: 2–6 s) during which participants were presented with an animation of the balloon inflating. Participants saw either a positive/win outcome or a negative/pop outcome during the “outcome phase” (average duration: 3 s, range:1–5 s). For the positive outcome, participants saw a visual “$” symbol and an audio of coins clanking while for a negative outcome, participants saw an animation of the balloon popping and an audio of a small explosion. If the balloon did not pop, the pump amount (one cent/pump) was added to an accumulating total which was visible in the lower right part of the screen (Green box). If the balloon popped, money was not won or lost, and the total accumulated amount of money remained unchanged. The “at stake” money amount (i.e., denoting the number of pumps currently registered on the response device) was visible on the blue box at the right side of the screen. The current trial number out of the 20 trials in a run remained visible in the pink box. Information regarding the pop value from the previous trial was presented on a box on the left side of the screen presented. A “rest phase” (average duration: 4 s, range:1–7 s) separated each trial.

Fig. 2.

BART-related brain activity. (A) Risky decision-making related brain activity (pump phase, p_corrected<0.05) was observed in multiple regions including the bilateral amygdala, bilateral striatum, bilateral cingulate gyrus, bilateral middle temporal gyrus, left precuneus, left angular gyrus. (B) Brain activity related to the positive outcome (wins, p_corrected<0.05) was observed in the bilateral caudate, bilateral cingulate, bilateral middle frontal gyrus, bilateral angular gyrus, left fusiform gyrus, left supramarginal gyrus, right lentiform nucleus, right thalamus. (C) Brain activity related to the negative outcome (pops, p_corrected<0.05) was observed in the bilateral inferior frontal gyrus, bilateral lingual gyrus, bilateral precentral gyrus, bilateral inferior frontal gyrus, left angular gyrus, left precuneus, left medial frontal gyrus, and right inferior parietal lobe (see Table S1 for brain activity coordinates).

Fig. 3.

Brain-behavioral relations. DM-related activity (pump vs. baseline) in the right amygdala was positively corelated with an objective measure of strategic-DM in the BART (r[22]=0.4, p_{Bonferronni-corrected[n=2]}=0.04), such that increased amygdala activity was associated with higher strategic-DM. However, risk-related activity in the right amygdala was not correlated with either personality metrics (r[22]=0.05, p = 0.83), or real-world behavior (r[22]=0.17, p = 0.43).

Fig. 4.

An objective measure of strategic-DM mediated right amygdala’s influence on money saving behaviors. (A, B) Given that strategic-DM (i.e., effective pump bet, EPB) correlated with both right amygdala activity and money saved, we conducted a formal mediation analysis to characterize the effect of right amygdala activity (X) on real-world financial behavior (Y) through BART performance (M). (C) BART performance fully mediated the influence of right amygdala activity on money saved. That is, when including BART performance as a mediator, the right amygdala’s direct effect on money saved failed to reach significance (c’ path: β = −85.17 p = 0.623), whereas the indirect effect was significant (ab path: β = 240.3: 95%CI=59.39, 528.27). Furthermore, self-reported alexithymia appeared to moderate this mediation. (D) When alexithymia was low, the conditional indirect effect of amygdala-performance-savings was significant (β = 625.14, 95%CI= 126.98, 1258.79). However, when alexithymia was high, the effect failed to reach significance (β = 131.48, 95%CI =−203.87, 360.32). Solid regression lines represent significant effect and dotted lines do not.