Abstract
Functional lateralization is a fundamental organizational principle of the human brain, yet the neural mechanisms underlying inter-individual variability in hemispheric dominance remain poorly understood. In this study, we investigated the causal network dynamics contributing to hemispheric lateralization within the face perception system, focusing on the fusiform face area (FFA) and occipital face area (OFA). Using Dynamic Causal Modelling (DCM) combined with Parametric Empirical Bayes (PEB) in a large sample of 110 participants, we examined how individual differences in lateralization indices (LI) relate to effective connectivity in the bilateral core face network. Two complementary approaches were applied: a hypothesis-driven model comparison and additionally an exploratory model reduction. Both analyses consistently showed that lateralization in the FFA and OFA was explained by distinct network mechanisms. FFA lateralization was primarily driven by processes in the left hemisphere, reflected in a face-specific modulation of self-inhibition in the left OFA. Specifically, increased left-lateralization was associated with reduced self-inhibition (i.e., increased excitability) in the left OFA. In contrast, OFA lateralization depended on interhemispheric interactions involving both hemispheres, most prominently between the left and right FFAs. Notably, in both cases, lateralization arose from network-level interactions rather than changes within the regions themselves, highlighting the distributed nature of hemispheric specialization in face processing.