Self-allocation bias in performance-based cooperative decisions is driven by self-interest rather than distorted performance encoding.

Human cooperation often involves performing joint tasks, where success relies on how collective rewards are allocated among cooperating parties based on their individual performance and contribution to task outcomes. However, it remains unclear whether and how individual performance and contribution give rise to self-related biases in such allocation decisions. Here, we developed a novel performance-based social allocation task that manipulated how individual performance contributed to joint outcomes. Across two experiments, participants exhibited a robust self-allocation bias: they allocated more rewards to themselves and disproportionately disregarded their own performance, particularly when their performance did not causally contribute to the joint outcome. This self-allocation bias was amplified in individuals with stronger individualistic social preferences, as measured by social value orientation. At the neural level, self-relevant (versus self-irrelevant) allocation decisions were associated with increased activity in the medial prefrontal cortex extending into the anterior cingulate cortex, insula, and temporoparietal junction (TPJ). Moreover, the dorsomedial prefrontal cortex, lateral orbitofrontal cortex, and TPJ tracked trial-by-trial variations in relative performance as a function of contribution structure, independent of self-relevance. Together, these findings suggest that self-allocation bias in performance-based decisions is unlikely to arise from distorted neural encodings of performance. Instead, self-interest may shape how contribution-structured performance information is used in social-allocation choices, providing a more precise account of how self-serving behavior emerges in cooperative contexts.