Importance: Hyperactivity in the subgenual cingulate cortex (SGC) is associated with major depressive disorder (MDD) and anticorrelated with activity in the dorsolateral prefrontal cortex (DLPFC). This association was found to be predictive of responsiveness to repetitive transcranial magnetic stimulation (rTMS) treatment. Such findings suggest that DLPFC-SGC connectivity is important for understanding both the therapeutic mechanism of rTMS in patients with MDD and the underlying pathophysiology of MDD. Objective: To evaluate SGC hyperactivity in patients with MDD before and after rTMS treatment. Design, Setting, and Participants: In this diagnostic study, among participants recruited from the adult and geriatric mood and anxiety services at the Centre for Addiction and Mental Health, Toronto, Ontario, Canada, who had participated in a randomized clinical trial, baseline SGC activity of patients with MDD was compared with healthy controls. In patients with MDD, SGC activity was compared before and after active or sham high-frequency rTMS treatment. Data collection started in July 2008 and concluded in March 2012. Neurophysiological data analysis started in January 2017 and ended in May 2018. Main Outcomes and Measures: Hyperactivity in the SGC before and after rTMS treatment was measured. Subgenual cingulate cortex hyperactivity activity was quantified using significant current density (SCD), and effective connectivity between the left DLPFC and SGC was computed using significant current scattering (SCS). Both measures were computed around TMS evoked potentials standard peak latencies prior to rTMS and after rTMS treatment, comparing patients with MMD treated with active and sham rTMS. Patients with MDD were assessed with the 17-item Hamilton Rating Scale for Depression. Results: Of 121 patients with MDD in the initial trial, 30 (15 [50.0%] women) were compared with 30 healthy controls (15 [50.0%] women) at rTMS treatment baseline. The mean (SD) age of the cohort with MDD was 39.1 (10.9) years, and the mean (SD) age of healthy controls was 37.0 (11.0) years. Following rTMS treatment, 26 patients with MDD who had active rTMS treatment (21.5%) were compared with 17 patients with MDD who had sham treatment (14.0%). At baseline, the SGC mean (SD) SCD and mean (SD) SCS at 200 milliseconds after TMS pulse were higher in participants with MDD compared with healthy controls (SCD: 1.04 × 10-6 [1.41 × 10-6] μA/mm2 vs 3.8 × 10-7 [7.8 × 10-7] μA/mm2; z = -2.95; P = .004; SCS: 0.87 [0.86] mm vs 0.54 [0.87] mm; z = -2.27; P = .02). Baseline source current density was able to classify MDD with 77% accuracy. Scores on the 17-item Hamilton Rating Scale for Depression were correlated with current density at the SGC (ρ = 0.41; P = .03). After rTMS treatment, SGC mean (SD) SCD and mean (SD) SCS at 200 milliseconds after rTMS pulse were attenuated to approximately the standard TMS-evoked potential latencies in the active rTMS group compared with the sham rTMS group (SCD: 1.57 × 10-7 [3.67 × 10-7] μA/mm2 vs 7.00 × 10-7 [7.51 × 10-7] μA/mm2; z = -2.91; P = .004; SCS: 0.20 [0.44] mm vs 0.74 [0.73] mm; z = -2.78; P = .006). Additionally, the SGC SCS change was correlated with symptom improvement on the 17-item Hamilton Rating Scale for Depression in the active rTMS group (ρ = 0.58; P = .047). Conclusions and Relevance: The findings of this study further implicate left DLPFC-SGC effective connectivity and SGC excitability in the pathophysiology of MDD and treatment with rTMS. These findings suggest that DLPFC-SGC connectivity may be a marker of rTMS treatment responsiveness. Trial Registration: ClinicalTrials.gov identifier: NCT01515215.