Neural bases of ingroup altruistic motivation in soccer fans

Humans have a strong need to belong to social groups and a natural inclination to benefit ingroup members. Although the psychological mechanisms behind human prosociality have extensively been studied, the specific neural systems bridging group belongingness and altruistic motivation remain to be identified. Here, we used soccer fandom as an ecological framing of group membership to investigate the neural mechanisms underlying ingroup altruistic behaviour in male fans using event-related functional magnetic resonance. We designed an effort measure based on handgrip strength to assess the motivation to earn money (i) for oneself, (ii) for anonymous ingroup fans, or (iii) for a neutral group of anonymous non-fans. While overlapping valuation signals in the medial orbitofrontal cortex (mOFC) were observed for the three conditions, the subgenual cingulate cortex (SCC) exhibited increased functional connectivity with the mOFC as well as stronger hemodynamic responses for ingroup versus outgroup decisions. These findings indicate a key role for the SCC, a region previously implicated in altruistic decisions and group affiliation, in dovetailing altruistic motivations with neural valuation systems in real-life ingroup behaviour.


Detailed description of psychometric measures
Group identification has long been studied as one of the chief constructs of Social Identity Theory (1,2), therefore having several available measures, differently from each other (3)(4)(5). We choose a validated measure of identification with soccer fans (6), the Football Supporter Team Identification Scale, which is composed of seven items that must be answered on a 10-point Likert scale in which 1 = strongly disagree and 10 = strongly agree.
Psychological kinship measures the extent to which an individual perceives other group members as family. It has three items and has been previously correlated to identity fusion and extreme pro-group behaviors (7). This construct has a direct relationship with evolutionary theories, being hypothesized as one of the proximal mechanisms to explain in-group cooperation with non-relatives (8).
The construct of entitativity (9), refers to the perception of a collection of persons as being bonded in a coherent unit. We used the measure created by Rüsch (2009), inspired on the study by Lickel and collaborators (2001) and based on Campbell (1958) and Rothbart and Park (2004) suggested components of entitativity: similarity, common fate and goals, and being a distinguishable and recognizable group in society.
The measure has four items, one related to the perceived entitativity of a focal group and the other three related to each of the aforementioned components. Figure S1. fMRI task design with an example of each condition. A cue indicating trial type was presented after a fixation cross jitter period. After cue presentation, participants pressed the handgrip dynamometer (effort period for Reward and Effortonly conditions) or passively observed the cue in Cue-only conditions. Real-time feedback of the amount earned during the effort period was provided, followed by a green outline indicating the end of the trial and the amount earned. Each condition was pseudo-randomly presented and the Effort-only condition was presented after 1/3 of each Reward condition. The cue for the Effort-only condition was always the same, indicating that participants should employ the same effort (as indicated by the number at the centre) as in the preceding Reward trial. The letters on the silhouettes stands for: VOC, "você" (Self); FLA, "Flamengo" (one of the soccer teams); STI, "Sem Time"

Detailed depiction of the fMRI task
(Non-fans).

Effects of self-concerned and altruistic reward conditions versus Effort-only
One-sample t-tests were performed to compare BOLD responses for each condition comparing Reward vs. Effort-only tasks. This high-level contrast enabled a tight control for effects of physical effort, while retaining the specificity of the motivation to obtain rewards for self and others. Whole-brain cluster FWE correction of each contrast showed robust BOLD responses in a priori ROIs as well in additional regions ( Figures S1-S3 and Tables S1-S3). However, the interaction contrasts only indicated significant effects in the mOFC and the VS for (Self -Effort-only) > (Non-fan -Effort-only) (see Table 1 in the main text).

Beta parameters for the left VS a priori ROI for each Reward vs. Effort-only contrasts
To further explore the null result in the VS for the interaction contrast [Fans -Effortonly] > [Non-fans -Effort-only], we extracted the parameter estimates from the a priori left VS ROI (-12, 10, −6) 13 for all the three Reward vs. Effort-only contrasts (Fig. S5).
Although the mean beta values were in line with our hypotheses, the effect in the VS for the interaction contrast Fans > Non-Fans did not reach significance due to the high standard errors.