Helping of conspecifics under threat has been observed across species. In humans, the dominant view proposes that empathy is the key proximal mechanism driving helping motivation in a threatening context, but little is known about how one s own defensive responses to the threat may guide helping decisions. In this pre-registered study, we manipulated threat imminence to activate the entire defensive brain circuitry, and assess the impact of different defensive responses on risky helping behaviour. Forty-nine participants underwent fMRI scanning while making trial-by-trial decisions about whether or not to help a co-participant avoid aversive shocks at the risk of receiving a shock themselves. Helping decisions were prompted under imminent and distal threat, based on the spatiotemporal distance to the administration of the shock to the co-participant. We found that greater engagement of reactive fear circuits (insula, ACC, PAG) during the threat presentation led to helping decisions, whereas engagement of cognitive fear circuits (hippocampus and vmPFC) preceded decisions not to help. Relying on representational similarity analysis, we identified how the defensive circuitry uniquely represented the threat to oneself, and the distress of the co-participant during the task. Importantly, we found that the strength with which the amygdala represented the threat to oneself, and not the other s distress, predicted decisions to help. Our results demonstrate that defensive neural circuits coordinating fast escape from immediate danger may also facilitate decisions to help others, potentially by engaging neurocognitive systems implicated in caregiving across mammals. Taken together, our findings provide novel insights into the proximal basis of altruistic responding, suggesting that defensive responses may play a more important role in helping than previously understood.