Social inference and the evolution of the human brain

The evolutionary forces that led to the unprecedented expansion of the human brain and the extreme cognitive prowess possessed by humans have always attracted a great deal of attention from the scientific community. Presented here is a novel theoretical perspective, where the driving force on human brain evolution was the need for enhanced ability to infer social values of conspecifics in the face of degradation and loss of chemosensory signalling mechanisms necessary for social communication present in most mammals.

The lack of chemosensory communication of biologically relevant information between humans in the face of the need to make adaptive and accurate social evaluations, led to an exaption of mammalian chemosensory brain regions for the more complex task of inferring social values from behavioural cues that are variable, ambiguous, or otherwise difficult to detect and interpret. This change in social processing from perceptual evaluation to inferential computation placed a premium on cognitive capacity, thus selecting for larger more powerful brains. These selective processes would have left an indelible mark on the human brain, where the human homologues of regions involved in mammalian conspecific chemical communication, in particular the target regions of this study the amygdala and ventromedial prefrontal cortex (VMPC), should be involved in the processing of biologically relevant information and social inference.

Several experiments were conducted to examine the role of these brain regions in social inferential processing using the lesion deficit method. First, given that conspecific chemical communication is particularly relevant for biologically imperative evaluation for the purposes of reproduction, VMPC and amygdala damage may result in abnormal mate-related decisions. Second, normal social attributions exhibit the correspondence bias, however damage to the target regions may result in an abnormal lack correspondence bias. Third, the current hypothesis is contrasted with another leading hypothesis, the Social Brain Hypothesis whose proponents predict a relationship between group-size and social cognition. Finally, if the target brain regions are truly integral in inferring social information, then damage to these regions will interfere with the ability to utilize transitive inference in social situations, and potentially in using transitive inference in general.

Damage to the target areas produces limited effects on mate-related decisions and preferences. However, the current hypothesis may suggest that the target brain regions are only involved when the problem is inferential in nature rather than simpler perception of social information. In support of this notion, damage to the target regions results in a lack of the correspondence bias when making economic decisions. This alteration in social attributions actually leads to more `rational' decision-making in this context. In contrast to the predictions of the Social Brain Hypothesis, damage to the target regions produces no observed reduction in social group size, nor is there any observed relationship between perspective-taking ability and group size. Finally, damage to the VMPC produces deficits in using transitive inference in a non-social context perhaps hinting at the underlying computations of this region in inferring social information.

In conclusion, it appears that the notion that the human brain regions that have been exapted from their duties in chemosensation and communication in mammalian brains has at least some validity. Moreover, these brain regions have been shifted by evolution to a more computationally complex process of social inference possibly providing the push toward larger and more powerful human brains.