The image is from a paper (ungated) by UCSD (political) scientists Darren Schreiber, Alan Simmons, Christopher Dawes, Taru Flagan, James Fowler, and Martin Paulus. Below is the abstract:
We matched public voter records to 54 subjects who performed a risk-taking task during functional imaging. We find that Democrats and Republicans had significantly different patterns of brain activation during processing of risky decisions. Amygdala activations, associated with externally directed reactions to risk, are stronger in Republicans, while insula activations, associated with internally directed reactions to affective perceptions, are stronger in Democrats. These results suggest an internal vs. external difference in evaluative process that illuminates and resolves a discrepancy in the existing literature. This process-based approach to political partisanship is distinct from the policy-based approach that has dominated research for at least the past half century. In fact, a two parameter model of partisanship based on amygdala and insula activations achieves better accuracy in predicting whether someone is a Democrat or a Republican than a well established model in political science based on parental socialization of party identification.
We reported earlier on a different paper that suggests Republicans have a thing with their Amygdala although that study saw an increased size of the anterior cingulate cortex among Democrats. Here is Andy Gelman’s take.