Abstract

Neural adaptation is a change over time in the responsiveness of neurons to a stimulus or situation. It has been extensively studied in sensory circuits, yet its consequences to social behaviour and choice is relatively unknown. Here we demonstrate how this fundamental property of the human brain explains a phenomenon with overreaching societal impact – escalation of corrupt behaviour. We combined brain imaging with a behavioural task in which individuals can repeatedly and voluntarily act dishonestly without being required to admit so. Specifically, on each trial participants advised a second participant regarding the amount of British Pennies in a glass jar. On some trials dishonesty was advantageous for the participant, while on others honesty was. The specific task structure enabled us to estimate the amount by which a participant was lying on each and every trial. Under this controlled laboratory setting we demonstrate that the extent to which participants engaged in dishonesty grows over time. Using fMRI we show that BOLD signal reduction in the amygdala is sensitive to the history of dishonest choices. Critically, we provide evidence that such adaptation acts as a teaching signal, predicting subsequent escalations of corruption. Real world examples in which corruption escalates are plentiful; from scientific misconduct to financial fraud, plagiarism and infidelity. Our findings suggest that despite being trivial at the outset, engagement in small immoral acts can trigger a biological process that generates a “slippery slope”: what begins as small acts escalate into larger instances with harmful consequences for those at the receiving end.