Leveraging fMRI to uncover antidepressant mechanisms of repetitive transcranial magnetic stimulation

Problem: Major depressive disorder (MDD) is a widely prevalent and disabling condition, yet mechanisms of treatment remain incompletely understood. Repetitive transcranial magnetic stimulation (rTMS) is an effective intervention for treatment-resistant depression (TRD), but substantial variability in clinical outcomes persists. rTMS is believed to exert antidepressant effects by modulating distributed brain circuits, with functional magnetic resonance imaging (fMRI), including emerging concurrent TMS-fMRI approaches, providing a powerful translational tool for probing these mechanisms. However, critical gaps remain regarding how rTMS engages circuits to produce downstream mood changes. Further, fMRI relies on the blood-oxygen-level-dependent (BOLD) signal, yet it is unknown whether rTMS alters neurovascular coupling, which could influence interpretation of BOLD-based measures. Methods: This thesis applies fMRI across three studies to evaluate antidepressant mechanisms of rTMS. Chapter 2 interrogates whether individualized functional targets (iFTs), defined using either subgenual anterior cingulate (SGC) or causal depression circuit (CDC) connectivity, could improve therapeutic outcomes. iFTs were retrospectively computed from baseline fMRI scans in TRD who received standard targeting as part of clinical trials. iFT engagement (measured through both proximity and electrical field (e-field) magnitude) was examined in relation to treatment outcomes. Chapter 3 and 4 draw on a concurrent TMS-fMRI dataset acquired in MDD patients prior to their rTMS treatment course. Chapter 3 uses electrocardiogram data to assess engagement of the central autonomic network by rTMS, reflected in altered coupling between heart-rate variability (HRV) and BOLD activity. Chapter 4 applies a blind deconvolution approach to estimate voxel-wise hemodynamic response function (HRF) parameters and evaluate whether rTMS perturbs neurovascular coupling. Conclusions: Chapter 2 found that iFT engagement did not predict antidepressant outcomes, suggesting limited clinical utility of current approaches, though future prospective validation using high-quality fMRI data and integration of e-field modeling is warranted. Chapter 3 showed rTMS causally engages the CAN, driven by altered right insula-HRV coupling, highlighting a potential mechanistic pathway. Chapter 4 revealed that rTMS induces slower and smaller HRFs, particularly within somatomotor regions, which did not confound measures of functional connectivity (FC). Together, these multidimensional approaches advance understanding of rTMS mechanisms and support the development of more precise, mechanism-informed treatments for depression.