We hypothesized that religiosity, a set of traits variably expressed in the population, is modulated by neuroanatomical variability. We tested this idea by determining whether aspects of religiosity were predicted by variability in regional cortical volume. We performed structural magnetic resonance imaging of the brain in 40 healthy adult participants who reported different degrees and patterns of religiosity on a survey. We identified four Principal Components of religiosity by Factor Analysis of the survey items and associated them with regional cortical volumes measured by voxel-based morphometry. Experiencing an intimate relationship with God and engaging in religious behavior was associated with increased volume of R middle temporal cortex, BA 21. Experiencing fear of God was associated with decreased volume of L precuneus and L orbitofrontal cortex BA 11. A cluster of traits related with pragmatism and doubting God's existence was associated with increased volume of the R precuneus. Variability in religiosity of upbringing was not associated with variability in cortical volume of any region. Therefore, key aspects of religiosity are associated with cortical volume differences. This conclusion complements our prior functional neuroimaging findings in elucidating the proximate causes of religion in the brain.
Obviously these sorts of studies need to be viewed skeptically, the intersection of religion & fMRI seems very biased toward a high level of "sexiness." Nevertheless, there are dozens of books on the psychology of religion, and, we know that religiosity is moderately heritable, so at some point the cognitive neuroscientists need to get in on the game of normal human variation in religious orientation (as opposed to studies of mystical brain states which seem focused on outliers). Here's the conclusion, which takes some sides in long festering arguments about the evolutionary origins of religion:
This study is correlational; therefore, it does not imply causality. Moreover, it was performed in adults. Subjects may have been predisposed to follow specific patterns of religious behavior by their individual brain development or their religious behavior may have contributed to volume changes of certain brain areas. Regardless, the fact that there is no brain area correlating with religiosity of upbringing (PC2) argues against religious nurture independently accounting for regional brain variability. Therefore, religiosity in adult life may reflect innate "susceptibilities", perhaps genetic or early developmental, which are non-modifiable during upbringing, or any initial effect of religious upbringing may be dissipated by experiences in later life. The brain areas identified in this and the parallel fMRI studies are not unique to processing religion, but play major roles in social cognition. This implies that religious beliefs and behavior emerged not as sui generis evolutionary adaptations, but as an extension (some would say "by product") of social cognition and behavior. Furthermore, the current study suggests that evolution of certain areas that advanced understanding and empathy towards our fellow human beings (such as BA 7, 11 and 21) may, at the same time, have allowed for a relationship with a perceived supernatural agent (God) based on intimacy rather than fear. The idea that how you relate to "thy God" parallels how you behave to "thy neighbor" is a long-cherished claim of many religions (and is backed by some empirical evidence...In this study, we see that this link is elaborated in the cognitive and neural foundations of religion.
Citation: Kapogiannis D, Barbey AK, Su M, Krueger F, Grafman J, 2009 Neuroanatomical Variability of Religiosity. PLoS ONE 4(9): e7180. doi:10.1371/journal.pone.0007180