Tiago Soares Bortolini¹,², Patrícia Bado¹,², Sebastian Hoefle², Annerose Engel²,³, Jean-Claude Dreher4 and Jorge Moll² 
¹ Graduate Program in Morphological Sciences, Federal University of Rio de Janeiro, Brazil 
² Cognitive and Behavioral Neuroscience Unit, D’Or Institute for Research and Education, Rio de Janeiro State, Brazil
³ Day Care Clinic for Cognitive Neurology University Hospital Leipzig, Max Planck Institute for Human Cognitive and Brain Sciences,Leipzig, Germany
4 Neuroeconomics Laboratory: Reward and Decision Making, Cognitive Neuroscience Center, Unité e Mixte de Recherche, Bron, France

Decades of behavioral studies have contributed with key knowledge on the phenomena of in-group bias, showing that people tend to favor members of own groups even in anonymous contexts. Neuroimaging studies have focused mainly on the hedonic aspects of in-group bias, demonstrating activity of the brain reward system by the act of cooperating with in-group members, largely based on the minimal group paradigm. In this study, we aim to investigate the neural bases of in-group cooperation, focusing on neural regions previously related to in-group perception and affiliative behavior (i.e. subgenual cortex and septo-hypothalamic area). To this aim, we developed a task in which participants could collaborate with other individuals from a meaningful social group: soccer fans of same soccer club. In a new event-related fMRI design, participants (N=25) could earn money for themselves (‘self-reward condition’), for anonymous participants that supported the same soccer team (‘team’ condition) and for anonymous participants that did not support any soccer team (‘non-team’ condition) by pressing a dynamometer while inside the scanner (24 trials in each condition). There was also an ‘effort only’ control condition (one third of the trials) in which participants were asked to perform an effort without monetary incentive for self or others. The amount of effort made on each trial determined how much money was accumulated in each condition. Participants benefited themselves directly only in the self-reward condition. Moreover, to reinforce in-group feelings, participants watched 15 clips of soccer fans of their own team in a stadium, as well as 15 clips of a ‘neutral’ group of fans in a similar context (interspersed across the experiment). Behavioral analysis showed that participants made more effort to earn money for themselves than for supporters of their own team or to non-fans. At the functional level, main effects of all effort (rewarded) conditions revealed robust BOLD activity in the medial orbitofrontal cortex (mOFC) and  posterior cingulate cortex (p < 0.001 uncorrected). Contrasts between each reward condition and their respective effort only conditions revealed common and distinct activated brain areas: frontopolar cortex (FPC) survived only in the non fan – effort only contrast, septal area survived only in the  fan – effort only contrast and the subgenual cortex survived only in the self – effort only contrast (a priori ROIs). Finally, among all clusters in the areas mentioned above, only the parameters estimates of the subgenual cluster had a positive association with the difference score on entitativity - how much participants perceived fans or non fans as an unit (r = 0.44, p = 0.027). These results indicate both a common brain network for in-group cooperation, ‘altruistic’ donation and self-reward, which included the mOFC and PCC, and different brain areas related to each condition. FPC activation during altruistic decisions may represent moral evaluations, above and beyond self reward and in-group preferences; while the subgenual cortex was more associated with effortful in-group cooperation, the septal area and adjacent ventral OFC may be more involved with positive emotional valence or reward.