A Multiscale Spatiotemporal Causal Mapping Algorithm for Revealing Neural Network Mechanisms of Transcutaneous Auricular Vagus Nerve Stimulation.

Transcutaneous auricular vagus nerve stimulation (taVNS) has shown promise in enhancing cognitive and emotional functions, yet its neural mechanisms remain unclear largely because existing analytical methods cannot characterize multiscale functional connectivity nor reliably infer causal interactions between brain regions in the presence of hemodynamic delays in fMRI signals. To address these limitations, we propose a Multiscale Spatiotemporal Causal Mapping (MSTCM) algorithm that integrates community-aware multiscale functional connectivity with delay-compensated causal inference. This design enables MSTCM to characterize multiscale connectivity structure and infer directed information flow with enhanced robustness. In evaluations using simulated fMRI data, MSTCM significantly outperformed seven existing causal inference algorithms across multiple evaluation metrics, including precision, sensitivity, Matthews correlation coefficient (MCC), and area under the receiver operating characteristic curve (AUC). Applied to resting-state fMRI across four predefined large-scale cortical networks before and after taVNS, MSTCM revealed that taVNS reduced functional coupling between the left lateral sensorimotor cortex (L-LSMC) and the right intraparietal sulcus (R-IPS), increased global efficiency, enhanced causal integration within the salience network (SN), weakened causal connectivity within the dorsal attention network (DAN), and strengthened information flow from DAN to SN. These findings suggest that taVNS may enhance cognitive flexibility and emotional regulation by shifting information processing from exteroceptive toward interoceptive pathways and improving large-scale network efficiency. Consequently, this study provides not only a novel methodological approach but also new neuroimaging evidence supporting the clinical potential of taVNS.