NETWORK-BASED INTEGRATION OF IMAGING MODALITIES IN DYNAMIC BRAIN NETWORKS TO ASSESS NEUROMODULATION-INDUCED STRUCTURAL CHANGES

Brain disorders are increasingly recognized as system disturbances of circuit connectivity and dynamics rather than isolated region dysfunctions, fostering a network-based view of brain organization. Graph-theoretical frameworks model brain regions as nodes and their interactions as edges, enabling single-level network analyses using structural, functional and diffusion-tensor (DTI) Magnetic Resonance Imaging (MRI). Neuromodulation targets distributed circuits, matching the network view of brain disorders. However, single-modality analyses dominate despite rich network data, lacking methods to integrate multi-level structural information to quantify the systems-level effects of neuromodulation.
To address this gap, we develop a framework that integrates volumetric and structural information into a unified network representation, enabling the assessment of neuromodulation effects on network topology using a publicly available neuromodulation dataset with a multi-timepoint protocol and multimodal neuroimaging data (T1-weighted imaging, DTI). To achieve this, we derived volumetric measures from structural MRI (e.g., grey matter volumes and cortical thickness), alongside DTI diffusion-derived anisotropy and tractography measures (e.g., FA, tract length). These single-modality features were then integrated into personalised, subject-level brain networks following established methodology (Zachariou M., 2018, DOI:10.1016/j.jprot.2018.03.009), yielding morphological and topological properties for each patient. Pre- and post-neuromodulation networks were assessed for rewiring patterns, identifying key brain regions exhibiting the largest changes across the neuromodulation protocol. This resulted in capturing patient-specific structural adaptations per region, demonstrating the method's sensitivity to localised circuit remodelling.
Overall, by enabling consistent pre-post multimodal comparisons, this network integration approach quantifies network-level effects of neuromodulation in brain networks and can be extended to additional modalities like functional MRI.