Exploring occupational neuroplasticity using a novel DAG-based effective connectivity model with fMRI.

Abstract

Occupational neuroplasticity shaped by occupational experiences offers valuable insights into neuropsychological health, cognitive interventions, and occupational selection. However, the underlying neural mechanisms are still not fully understood. Currently, investigating these mechanisms by estimating effective connectivity (EC) networks from fMRI data represents a promising approach. Nevertheless, existing models face challenges including low temporal resolution, high dimensionality, and limited interpretability. To address these challenges, the paper proposes a novel DAG estimation model, GroupDAGs, which uses an M-matrix to construct acyclic constraints and incorporates modularity and group similarity. Compared to existing methods, GroupDAGs enhance the accuracy and interpretability of brain EC estimation. Its performance was extensively validated through simulations with various noise types and graph structures. Furthermore, using seafarers as an example, it was applied to collected pre- and post-voyage fMRI data to explore the neuro-causal mechanisms of occupational neuroplasticity. The results showed that seafarers' brain occupational neuroplasticity is reflected in enhanced brain network modularity, increased cerebellar specialization, and improved task-related attentional control and motor coordination abilities. Key brain regions in seafarers linked to changes in emotional regulation and social cognition were also identified. Together, this study not only introduces a novel method for calculating brain EC networks but also provides new evidence for occupation-related neural neuroplasticity.