Religious beliefs, particularly those concerning divine influence, are among humanity's most widespread and personally significant beliefs, yet they also exhibit remarkable diversity. This variability in religious belief and non-belief is as ubiquitous as the presence of religion across human cultures. The impact of these differences is substantial, influencing aspects from personal identity formation and social affiliations to national and international political landscapes. Understanding the neurocognitive underpinnings of individual differences in belief is therefore crucial. Existing theoretical frameworks within psychology and anthropology suggest that religious beliefs arise, at least partially, from perceptual mechanisms evolved for predictive processing of environmental information. While direct observation of the co-evolution of religious beliefs and human information processing is impossible, variations in religious and perceptual information processing among individuals offer valuable insights. Specifically, individual differences in relevant perceptual mechanisms for predictive processing of environmental stimuli may predispose individuals towards or away from religious beliefs. A widely accepted principle in human information processing frameworks is the interplay between automatic bottom-up processes (driven by sensory and perceptual input) and top-down processes (more accessible to conscious awareness). These cognitive strata are not entirely separate, and bottom-up processes likely formed the evolutionary foundation for more explicit top-down cognition. Several theories of human cognition outline how bottom-up processes guide individuals toward specific explicit beliefs. These theories propose that intuitions (experiences of knowledge not consciously learned) emerge from bottom-up processing of environmental signals, influencing subsequent conscious interpretation through explicit belief narratives that rationalize beliefs and behaviors. Intuitions frequently bias top-down judgments, making beliefs rooted in intuition particularly compelling and resistant to change. This has been observed across diverse sensory information processing modalities, such as the reliance on rapid, unconscious face processing in forming intuitions of trustworthiness. Existing theories posit that evolved neurocognitive processes contribute to default intuitions about the environment, shaping explicit beliefs. For instance, neurobiological mechanisms evolved for human cooperation are linked to intuitions about morality and fairness, which in turn may bias individuals toward beliefs in divine justice. Similarly, pathogen-avoidance mechanisms have been connected to intuitions about cleanliness and purity-related religious beliefs. The automatic bias to detect agency in the environment may contribute to over-attribution of intentional agency, supporting beliefs in supernatural agents. However, empirical support for this has been mixed, with some research suggesting under-attribution of agency, and others demonstrating a link between belief in the paranormal and illusory agency detection. Other research points towards a link between top-down, analytic thinking (critical examination of automatic responses) and religious disbelief, although replication attempts have yielded inconsistent results. A core tenet across major religions is that the sequence of events reflects an underlying order determined by divine intervention. This interventionist belief—a key focus of religious practices in various faiths—suggests that identifying individual differences that influence this aspect of belief could provide crucial insights into overall religious belief variation. Given the link between intuition and religious beliefs, and the established role of order-related perceptual information processing in shaping intuitions, particularly implicit pattern learning (IL-pat), the authors hypothesize that IL-pat could influence belief in an intervening/ordering God. This hypothesis is based on the idea that IL-pat contributes to intuitions of order, which in turn influence beliefs.

The existing literature suggests a complex interplay between bottom-up and top-down cognitive processes in shaping religious beliefs. Evolutionary psychology and cognitive science offer perspectives on how innate predispositions and environmental factors interact to form beliefs. Several studies have explored the role of intuitive thinking versus analytic thinking in belief formation, with some evidence suggesting that a more intuitive cognitive style is associated with greater religious belief, while analytic thinking might be correlated with skepticism. However, replications of some of these findings have been inconsistent. The concept of implicit learning, where patterns are acquired without conscious awareness, has also been implicated in belief formation. This research suggests that individuals' ability to implicitly learn patterns in their environment may influence their intuitions about order and purpose in the world, ultimately shaping their religious beliefs. Furthermore, the importance of cultural and social contexts in shaping religious beliefs is emphasized in various studies. The authors of this study aim to address the lack of research incorporating non-Western samples, focusing on both US and Afghan participants to examine the cross-cultural generalizability of their hypothesis.

The study employed a cross-cultural design, comparing participants from the United States and Afghanistan. The sample comprised 199 participants from the US and 148 from Afghanistan, recruited from diverse backgrounds. The US sample consisted of a majority of Christian participants and a significant number of unaffiliated individuals. Religious affiliation was not directly investigated in the Afghan sample due to potential safety concerns. A modified Serial Reaction Time Task (SRTT) was used as the primary measure of implicit pattern learning (IL-pat). Participants responded to visually presented stimuli (circles appearing at different screen positions) by pressing corresponding keys. The SRTT consisted of both pattern blocks (sequences of stimuli following a set pattern) and random blocks (sequences with no discernible pattern). IL-pat was calculated as the difference in response time between pattern and random blocks. To assess explicit awareness, participants were asked after each block whether they perceived a pattern. Measures of belief in God included the Belief in Divine Intervention Scale (BDIS), the Overlapping Circles Task (to assess self-overlap and world-overlap; indicating the perceived extent of God's influence on the individual and the world), and a lab-developed measure to assess change in belief from childhood to adulthood. Potential confounding variables were measured, including schizotypal ideation (using the Schizotypal Questionnaire) and parental religious belief. A separate measure of universal order (UO) intuitions was included, using statements about order in the universe without religious connotations. This measure was administered only to US participants and an additional online European sample. The Cognitive Reflection Test (CRT) was administered to a subsample of US participants to assess analytic thinking. The Belief in Science Scale was used in the US subsample and the European sample to assess attitudes towards science. Statistical analyses included linear regression models to investigate the relationship between IL-pat and belief measures, while controlling for potential confounding variables. Mediation analysis was employed to examine the potential mediating role of UO intuitions in the relationship between IL-pat and belief. Data analysis was conducted separately for each sample to assess cross-cultural replicability.

The study replicated its key findings across the US and Afghan samples, demonstrating cross-cultural generalizability. Implicit pattern learning (IL-pat), as measured by the SRTT, was a significant predictor of interventionist belief (IB), even when controlling for schizotypal ideation and parental religious belief. Specifically, stronger IL-pat was associated with stronger IB in both samples. Furthermore, IL-pat significantly predicted changes in belief strength from childhood to adulthood in both samples. Importantly, IL-pat was not correlated with explicit awareness of patterns in the SRTT, supporting the notion that the observed effects were driven by implicit learning. In the US sample, mediation analysis revealed that intuitions of universal order (UO) significantly mediated the relationship between IL-pat and IB, and also mediated the relationship between IL-pat and changes in belief across time. The results indicated that stronger IL-pat was associated with stronger UO, which in turn was associated with stronger IB. IL-pat was not significantly associated with a measure of belief in God's existence generally, indicating that the influence of IL-pat may be more specific to the interventionist aspect of belief. The association between IL-pat and belief was found to be independent of the performance on the Cognitive Reflection Test (CRT), a measure of analytic thinking, suggesting that the observed effects were not attributable to differences in cognitive styles. Finally, belief in science was inversely correlated with IB, but unrelated to IL-pat or UO intuitions.

The findings strongly support the hypothesis that implicit pattern learning contributes to the formation and development of beliefs about a divine intervening force in both culturally distinct samples. The study's strength lies in its successful replication across vastly different socio-religious contexts, suggesting that the relationship between IL-pat and religious belief may be a fundamental aspect of human cognition, rather than a culturally specific phenomenon. The mediation analysis provides further support for the proposed causal pathway: superior implicit learning of patterns leads to stronger intuitions of universal order, which in turn predisposes individuals towards accepting belief narratives that posit an ordering deity. The lack of correlation between IL-pat and explicit pattern recognition emphasizes the importance of unconscious processes in shaping religious beliefs. The results suggest a bottom-up influence on religious belief, as opposed to a top-down influence where religious belief would lead to more pattern seeking. The independence of this effect from analytic reasoning as assessed by the CRT further strengthens this bottom-up interpretation. While the study focused on the bottom-up pathway, the possibility of complex interplay between bottom-up and top-down processes in shaping religious beliefs is acknowledged. Future research could explore longitudinal studies to establish stronger causal links and investigate potential top-down influences. However, the results of this study provide strong initial empirical support for the theoretical framework presented.

This study provides compelling cross-cultural evidence for a link between implicit pattern learning and belief in an intervening God. Stronger implicit learning of patterns is associated with stronger belief in a divine ordering force and increased belief strength over time. The mediating role of intuitions of universal order suggests a pathway where perceptual processes shape broader intuitions that, in turn, influence religious beliefs. The findings highlight the significant role of unconscious cognitive processes in the development of religious beliefs. Future research should focus on longitudinal studies to further explore the causal relationships and investigate the broader implications of these findings for understanding religious diversity and the neurocognitive foundations of belief.

The study's cross-sectional design limits the ability to definitively establish causal relationships between implicit pattern learning and religious belief. While the findings suggest a bottom-up influence, the complex interplay of bottom-up and top-down processes warrants further investigation. The use of retrospective self-report for assessing childhood beliefs might introduce recall bias. While the sample sizes are large, they may not fully represent the diversity within the US and Afghan populations. The lack of direct measures of religiosity in the Afghan sample limits the scope of some analyses. Additionally, the mediation analysis is correlational in nature and thus does not directly demonstrate causality. Finally, the study's focus on interventionist belief does not preclude the possibility of other factors contributing to overall religious belief.