Background: Stimulation adjustment is required to optimize outcomes of deep brain stimulation (DBS) for treatment-resistant depression (TRD), but controlled data on ideal stimulation parameters are limited. Using a double-blind randomized cross-over design, we measured the efficacy and safety of short (SPW) and long (LPW) pulse width subcallosal cingulate (SCC) DBS. Methods: Twenty-two patients with TRD underwent bilateral DBS implantation into SCC white matter using diffusion tensor imaging tractography in an academic hospital. Patients were randomized using computer-generated numbers to receive either SPW or LPW stimulation for 6 months. Clinician assessing outcomes was blinded to stimulation group. Keeping frequency constant (130 Hz), either pulse width or voltage was increased monthly, based on response using 17-item Hamilton Depression Rating Scale (HDRS). Non-responders (< 48% reduction in HDRS from the baseline) at 6 months were crossed-over to opposite stimulation for another 6 months. Primary outcome was change in HDRS at 6 and 12 months. Findings: One responder from SPW group suicided at 3 months, therefore data from 21 patients were analyzed, as per protocol. Response rates were 48% at both 6 and 12 months for the entire sample. Both stimulation groups showed significant reduction in HDRS (p <0.001), with no difference between groups at 6 months (p=0.61), plus within (p=0.21) and between (p=0.52) cross-over groups at 12 months. Area under the curve analysis revealed a trend for stronger clinical response in the LPW group over 6-12 months (p=0.051). LPW patients had better remission rates (30-36%) than SPW (10-20%) over 6-12 months. Adverse effects and battery drainage were comparable between groups. Add-on individualized cognitive behavioral therapy did not change HDRS scores. Interpretation: While both LPW and SPW stimulation were equally safe and effective, clinical response was stronger with LPW over time, indicating a role for LPW in stimulus optimization of SCC-DBS for TRD. Trial Registration: ClinicalTrails.gov: NCT01983904 Funding Statement: Alberta Innovates and Health Solutions Declaration of Interests: Dr. Ramasubbu has received honorarium for serving in the advisory committee of Astra Zeneca, Lundbeck, Janssen, and Otsuka. He also received investigatorinitiated grant from Astra Zeneca and Pfizer. Other authors declare no potential conflict of interest. Dr. Kiss reports that Alberta health services receives value adds for research and education from Medtronic Canada, the manufacturer of the devices used in the study. These funds are only partially available for research to the authors for research and education through the Neuromodulation program. Medtronic had no involvement in any of this study. Ethics Approval Statement: This was a double-blind randomized controlled parallel and cross-over trial approved by the Conjoint Health Research Ethics board at the University of Calgary. Written informed consent was obtained.