The study addresses the evolving debate on effective pedagogical approaches that foster not just factual knowledge and procedural proficiency, but also the ability to self-regulate learning. A key aspect of this self-directed learning is error monitoring – the capacity to detect and evaluate unexpected outcomes and adapt accordingly. Error monitoring is linked to self-regulatory behaviors and executive functions, improving across development with adult-like responses emerging by mid-adolescence. The anterior cingulate cortex (ACC) and its functional connectivity are key neural components in this process. The researchers hypothesized that educational experiences shape error monitoring and its neural correlates. Traditional pedagogy often provides delayed feedback on errors, whereas Montessori education emphasizes self-discovery and peer-learning from mistakes. Montessori students consistently demonstrate higher academic and socio-emotional achievements, and this study aimed to explore whether this advantage relates to differences in error monitoring. The researchers predicted that both groups would exhibit cingulate gyrus activity in response to errors, but that the patterns of brain activation and connectivity would differ, with Montessori students showing stronger activity and connectivity in error-monitoring regions (ACC, medial frontal cortex) and traditionally-schooled students showing effects in memory-related regions (hippocampus).

The introduction and discussion sections extensively review relevant literature. Studies on adults consistently implicate the cingulate regions, particularly the ACC, in error monitoring. Neurodevelopmental research highlights age-related changes in cingulate gyrus activity and connectivity related to error processing, especially between 8-12 years. However, few studies examine error monitoring's neural correlates in children, and none had previously investigated its susceptibility to different pedagogical approaches. Prior work by the authors indicated stronger EEG responses to error detection in traditionally schooled students compared to Montessori students, suggesting differences in self-regulated error monitoring. This aligns with research showing that incorrect responses are typically negatively valenced in traditionally schooled adults. The study thus aimed to fill this gap in research by comparing error monitoring in children from different educational backgrounds using fMRI.

Thirty-seven children (ages 8-12.3 years) participated, with 32 (16 from each school type) included in the final analyses after exclusion criteria were applied (motion artifacts, medical conditions). Participants completed a math proofreading task during fMRI scanning, judging the correctness of solutions to simple math problems. Behavioral data (response times, accuracy, missed trials) were collected. fMRI data were preprocessed using SPM12: motion correction, slice timing correction, co-registration, normalization to MNI template, and spatial smoothing. Statistical analyses employed a two-way mixed-design ANOVA with response type (correct/incorrect) and pedagogy (Montessori/traditional) as factors. A psycho-physiological interaction (PPI) analysis examined functional connectivity, using seed regions identified from the activation analysis. Control variables (non-verbal intelligence, anxiety, working memory, math skills, socioeconomic status, parental math affect, home environment) were collected to control for potential confounding factors and statistically compared between the groups using t-tests or chi-square tests.

Behavioral results showed no group differences in overall accuracy. However, Montessori students responded faster (MRT_M= 1719 ms, SD=405 vs MRT_T=2060, SD=351), had a higher rate of incorrect responses, and a significantly lower rate of missed trials compared to traditionally schooled students. fMRI analyses revealed that both groups exhibited higher brain activity for correct compared to incorrect responses in regions including the posterior cingulate cortex, precuneus, and cerebellum. Importantly, pedagogical group differences emerged in functional connectivity patterns. Traditionally-schooled students displayed stronger connectivity between the ACC, medial prefrontal cortex, and cuneus with the hippocampus following correct responses. In contrast, Montessori students showed stronger connectivity between the ACC and frontal regions following incorrect responses. These differences were not observed for correct responses in Montessori students or incorrect responses in traditionally-schooled students.

The findings suggest that children may process correct and incorrect responses differently than adults, with a greater focus on correct responses in this age group. The observed differences in functional connectivity patterns between the two groups indicate the significant impact of pedagogical experience on error monitoring and its neural correlates. The stronger connectivity between the ACC and hippocampus in traditionally schooled students may reflect a greater emphasis on memory and recall, potentially through instrumental learning or reinforcement learning mechanisms. In contrast, the stronger ACC-frontal connectivity in Montessori students following errors may reflect a more exploratory and self-corrective approach to learning, driven by self-directed learning and multisensory integration. These differing connectivity patterns are potentially linked to different learning strategies emphasized in each pedagogical approach.

This study demonstrates that pedagogical experience significantly influences error monitoring behavior and neural activity in children aged 8–12 years. Montessori education seems to foster a more process-oriented approach, potentially through greater self-correction and multisensory engagement. Future research should investigate whether these differences persist into adulthood and explore the generalizability of findings to other learning domains. Longitudinal studies with larger sample sizes are crucial to confirm the long-term impact of pedagogical approaches on the development of error monitoring and its neural substrates.

The study's cross-sectional design and non-random assignment to school types limit the ability to definitively establish causal links between pedagogy and observed outcomes. The moderate sample size might limit the power to detect subtle effects. Future research should use longitudinal designs with larger, randomly assigned samples to strengthen causal inferences. The lack of adult participants prevents determining whether observed differences are developmental or persistent across ages. Future studies should incorporate adult participants with tracked pedagogical histories for a more comprehensive understanding of the developmental trajectory of error monitoring.