Microstructural changes in prefrontal–cingulate networks following dorsomedial prefrontal cortex-targeted intermittent theta burst stimulation in treatment-resistant depression

Intermittent theta-burst stimulation (iTBS) is a well-established transcranial magnetic stimulation protocol for the treatment of treatment-resistant depression; however, the neurobiological mechanisms underlying its therapeutic effects remain incompletely understood. Although the dorsolateral prefrontal cortex has traditionally been the primary stimulation target, accumulating evidence suggests that the dorsomedial prefrontal cortex constitutes a central hub within large-scale emotion regulation networks. Given its extensive structural and functional connectivity with cingulate and limbic regions, modulation of this area may induce network-level changes associated with clinical improvement. In this study, we aimed to investigate white matter microstructural alterations following bilateral dorsomedial prefrontal cortex–targeted iTBS delivered with a double-cone coil and to examine their relationship with treatment response.

Twenty-one patients with treatment-resistant depression and thirty-two demographically matched healthy controls were included. Diffusion tensor imaging (DTI) data were acquired in all participants at baseline, and repeated in the patient group following completion of a full treatment course. iTBS was administered bilaterally to the dorsomedial prefrontal cortex using a double-cone coil. Clinical severity was assessed longitudinally using standardized depression rating scales. White matter tracts connecting prefrontal, cingulate, and limbic regions were reconstructed using probabilistic tractography, and tract-specific diffusion metrics, including fractional anisotropy, mean diffusivity, axial diffusivity, and radial diffusivity, were extracted. Analyses focused on longitudinal microstructural changes, group differences relative to healthy controls, and associations between diffusion parameters and clinical outcomes. Statistical models incorporated appropriate corrections for multiple comparisons.

iTBS treatment was associated with a significant reduction in depressive symptom severity. At the microstructural level, changes were not uniformly distributed but instead demonstrated tract-specific patterns within prefrontal–cingulate networks. Notably, a significant post-treatment increase in fractional anisotropy was observed in the left anteromedial prefrontal–rostral anterior cingulate pathway, accompanied by a shift toward values observed in healthy individuals, suggesting a normalization process. In parallel, distinct microstructural alterations within dorsomedial prefrontal–cingulate pathways were significantly associated with clinical improvement, indicating that these changes may reflect functionally meaningful remodeling rather than nonspecific diffusion effects.

Baseline microstructural properties also demonstrated predictive relevance. Specifically, higher pretreatment fractional anisotropy within orbitofrontal–subgenual cingulate connections was associated with greater clinical improvement, highlighting this pathway as a potential structural biomarker of treatment response. Furthermore, comparisons with healthy controls revealed widespread baseline microstructural differences, predominantly localized in the right hemisphere. In contrast, treatment-related changes were more pronounced in the left hemisphere, suggesting a dissociation between disease-related alterations and treatment-sensitive neuroplastic processes.

Taken together, these findings support a network-level model in which dorsomedial prefrontal cortex–targeted iTBS induces tract-specific microstructural reorganization within prefrontal–cingulate–limbic circuits. Rather than reflecting a uniform effect, these alterations appear to represent coordinated remodeling across functionally distinct pathways involved in emotion regulation. The identification of both treatment-related changes and baseline predictive markers underscores the potential role of DTI-based measures in informing personalized neuromodulation strategies. Further studies with larger samples and multimodal designs are warranted to confirm these findings and to clarify the underlying mechanisms.

References

No references