LAUSR

OBJECTIVE Functional MRI (fMRI) has been used to investigate the therapeutic mechanisms underlying deep brain stimulation (DBS) for Parkinson's disease (PD). However, the alterations in stimulation site-seeded functional connectivity induced by DBS at the internal globus pallidus (GPi) remain unclear. Furthermore, whether DBS-modulated functional connectivity is differentially affected within particular frequency bands remains unknown. The present study aimed to reveal the alterations in stimulation site-seeded functional connectivity induced by GPi-DBS and to examine whether there exists a frequency band effect in blood oxygen level-dependent (BOLD) signals related to DBS. METHODS Patients with PD receiving GPi-DBS (n = 28) were recruited for resting-state fMRI with DBS on and DBS off under a 1.5-T MR scanner. Age- and sex-matched healthy controls (n = 16) and DBS-naïve PD patients (n = 24) also received fMRI scanning. The alterations in stimulation site-seeded functional connectivity in the stimulation-on state versus stimulation-off state, as well as the relationship between alterations in connectivity and improvement in motor function induced by GPi-DBS, were examined. Furthermore, the modulatory effect of GPi-DBS on the BOLD signals within the 4 frequency subbands (slow-2 to slow-5) was investigated. Finally, the functional connectivity of the motor-related network, consisting of multiple cortical and subcortical regions, was also examined among the groups. In this study, p < 0.05 with Gaussian random field correction indicates statistical significance. RESULTS Functional connectivity seeding from the stimulation site (i.e., the volume of tissue activated [VTA]) increased in the cortical sensorimotor areas and decreased in the prefrontal regions with GPi-DBS. Alterations in connectivity between the VTA and the cortical motor areas were correlated with motor improvement by pallidal stimulation. The alterations in connectivity were dissociable between the frequency subbands in the occipital and cerebellar areas. The motor network analysis indicated decreased connectivity among most cortical and subcortical regions but increased connectivity between the motor thalamus and the cortical motor area in patients with GPi-DBS compared with those in DBS-naïve patients. The DBS-induced decrease in several cortical-subcortical connectivities within the slow-5 band correlated with motor improvement with GPi-DBS. CONCLUSIONS These findings indicate that the alterations in functional connectivity from the stimulation site to the cortical motor areas, as well as multiple connectivities among the motor-related network, were associated with the efficacy of GPi-DBS for PD. Furthermore, the changing pattern of functional connectivity within the 4 BOLD frequency subbands is partially dissociable.