Indian cohort study maps non-clinical eating behavior network architecture using mixed graphical models.

Eating behaviors and their associated cognitions exist along a biopsychosocial continuum, yet their structural organization remains largely unmapped in non-Western contexts. Adopting a dimensional network perspective, this study characterizes the architecture of non-clinical eating behaviors in India, a region defined by a unique interplay of cultural, structural, and psychological influences. We utilized Mixed Graphical Models (MGMs) to estimate a weighted network of 35 variables from a geographically diverse Indian cohort (N=1,508). Our analysis reveals that the Indian eating behavior landscape is a highly optimized, small-world system (S=54.64) defined by a dual-layered hierarchy of influence. We found that structural and cultural variables, notably HomeTypes and Religion, serve as the primary local anchors (highest Expected Influence), driving the state of their immediate modules. Conversely, systemic integration across the entire network is maintained by a "socio-economic and regulatory bridge" comprising Employment, Education, and Self-Esteem. These nodes exhibited the highest betweenness centrality, functioning as the critical "highways" that link disparate socio-economic, psychological, and behavioral modules. Notably, while Shape and Weight Concern were highly predictable, they functioned as local cluster nodes rather than global integrators, directly challenging the body-image-centric models dominant in Western literature. These results demonstrate that in the Global South, structural social determinants form the primary scaffold of the biopsychosocial system. Our findings provide a data-driven blueprint for systemic, culturally attuned public health interventions that prioritize structural stability alongside individual regulatory resilience.