Journal Club · Functional Neurosurgery

DBS for Treatment-Resistant Depression

When two negative randomized trials did not settle the question — and why the field is busier than ever.

A trainee reading of the pivotal sham-controlled trials of deep brain stimulation for treatment-resistant depression — BROADEN and RECLAIM — framed by the open-label promise that preceded them and the durability and connectomic data that followed. The real lesson is about why rigorous trials can fail to confirm a real effect.

Mayberg Neuron 2005 BROADEN Lancet Psych 2017 RECLAIM Biol Psychiatry 2015 Crowell Am J Psych 2019

Why This Topic

DBS for treatment-resistant depression (TRD) is the cautionary tale — and the unfinished story — of psychiatric neurosurgery. Open-label studies were dramatic; the two industry-sponsored randomized sham-controlled trials both stopped for futility; and yet long-term data and a new generation of individualized, circuit-guided approaches have kept the field very much alive. For a functional neurosurgery audience the value is less about a single result than about why rigorous trials can fail to confirm a real effect — and what that teaches about trial design in neuromodulation.

One-Line Takeaway

Two sham-controlled trials of DBS for depression failed at short blinded endpoints — but durable open-label benefit, the superiority of individualized connectomic targeting, and biomarker-guided closed-loop stimulation argue the trials measured the wrong thing at the wrong time in an un-individualized way, not that the premise is wrong.

Part I

The Arc Of The Evidence

1.The Promise — Mayberg et al. (Neuron 2005)

The modern era began with an open-label proof-of-concept (n=6) stimulating the subcallosal cingulate (SCC) white matter / Brodmann area 25, a region shown on functional imaging to be metabolically overactive in depression. Four of six patients achieved a striking, sustained response, with corresponding reductions in area-25 activity. This reframed depression as a treatable circuit disorder and launched DBS for TRD. (Seminal open-label; details from the published report.)

2.BROADEN — The Subcallosal Cingulate RCT (Holtzheimer, Lancet Psychiatry 2017)

A multisite, prospective, randomized, double-blind, sham-controlled trial. Patients with bilateral SCC leads were randomized to 6 months of active vs sham stimulation, then 6 months open-label. The primary outcome was response frequency (≥40% reduction in depression severity) averaged over months 4–6 of the blinded phase, with a pre-specified futility analysis at roughly half the planned sample.

Before the futility analysis, 90 participants were randomized to active (n=60) or sham (n=30). Both groups improved, but there was no significant difference in response during the blinded phase — 12/60 (20%) active vs 5/30 (17%) sham. The trial confirmed safety and feasibility but not efficacy at 6 months; the sponsor halted it. The authors explicitly named clinical features and electrode placement as factors future studies should address.

3.RECLAIM — The Ventral Capsule/Ventral Striatum RCT (Dougherty, Biol Psychiatry 2015)

The first completed RCT of DBS for TRD: 30 patients randomized to active vs sham at the VC/VS target, blinded for 16 weeks, then open-label. Primary outcome: response = ≥50% MADRS improvement.

There was no significant difference — active 3/15 (20%) vs sham 2/14 (14.3%), and no difference in continuous MADRS change at 16 weeks. The trial was terminated early for futility (30 of a planned ~200 enrolled). Open-label response was 20–26.7% at 12/18/24 months.

The One-Two Punch Two sham-controlled trials, two different targets, two different sponsors — both stopped for futility at fixed blinded endpoints within about two years of each other. This is what people mean when they say “the DBS depression trials failed,” and it largely ended industry investment for several years.

4.The Data That Kept The Field Alive

Durability (Crowell et al., Am J Psychiatry 2019). An open-label, long-term follow-up of 28 SCC-DBS patients (20 major depression, 7 bipolar II) over 4–8 years found response (≥50%) and remission (≥30%) rates maintained through years 2–8; three-quarters met response criteria for more than half their time in the study, and 21% showed continuous response from year 1. The implication is pointed: a blunt 6-month blinded endpoint can score as a failure a therapy whose benefit is real but slow and durable.

A different time course (Schlaepfer et al., Biol Psychiatry 2013). Stimulation of the superolateral medial forebrain bundle in 7 patients produced a rapid antidepressant effect — mean MADRS fell >50% by day 7 — with 6 of 7 responders at last observation, at a low mean current (2.86 mA). Target and circuit choice shape both magnitude and latency of response.

Part II

Why The Trials Probably Failed

5.Design And Biology, Not A Refuted Premise

  • The endpoint window fought the biology. Efficacy was judged at 4–6 months (BROADEN) or 16 weeks (RECLAIM), but response often emerges and consolidates over many months to years.
  • Targeting was anatomical, not individualized. The SCC is a white-matter convergence zone; millimeter differences change which tracts are engaged. BROADEN’s authors named electrode placement as a likely culprit.
  • “TRD” is biologically heterogeneous. A single fixed target and program across a mixed population dilutes any true effect.
  • Underpowering and futility stops. RECLAIM enrolled 30 of ~200 planned; BROADEN ~90 of its target. Low responder counts in both arms on small samples gave little chance to detect a moderate effect.
  • Sham and discontinuation designs are hard in implanted patients. Expectancy effects, imperfect blinding, and ethical limits on sham duration all confound interpretation.
  • No bedside readout for programming. Unlike tremor (visible suppression on the table), depression offers no immediate signal to optimize the contact; optimization was slow and non-standardized across sites.
  • The stops were also business decisions. Sponsor-driven futility analyses ended the regulatory pathway and chilled investment — “failure” partly reflects commercial risk tolerance.
Part III

Why Interest Remains

6.The Case For Continuing

Durable open-label response

Sustained benefit out to 8 years (Crowell 2019) is difficult to dismiss as placebo.

Connectomic targeting

Prospective tractography-guided SCC targeting (Riva-Posse/Mayberg 2018) reported ~73% response at 6 months and ~82% at 1 year — far above BROADEN’s blinded result.

Closed-loop, biomarker-driven stimulation

Scangos et al. (Nat Med 2021) mapped a patient-specific depression biomarker and delivered state-dependent stimulation with rapid benefit (n=1 proof-of-concept).

The OCD precedent

The 2009 FDA Humanitarian Device Exemption for VC/VS DBS in severe OCD — the same region as RECLAIM — proves psychiatric neuromodulation can clear a regulatory bar.

A Sobering Counterpoint Even with an FDA pathway, DBS-for-OCD adoption remained low — psychiatric DBS faces implementation and acceptance hurdles, not only evidentiary ones. And the historical lesional record (anterior capsulotomy for OCD and depression) is a reminder that limbic-circuit modulation has decades of precedent but also a cautionary legacy that demands careful patient selection and ethics oversight.

Questions For The Table

  • If response is slow and durable, what is the right blinded endpoint — and is a fixed-duration sham phase even the correct design?
  • How much of the failure is targeting (fixable by tractography) versus patient heterogeneity (fixable by selection/biomarkers)?
  • Does the connectomic responder rate (~73–82%) justify a new pivotal trial, and how would you power and blind it?
  • What ethical guardrails should govern DBS for depression given the lesional history and the vulnerability of the population?
  • Is closed-loop, biomarker-driven stimulation the paradigm shift, or another promising n=1?
  • A negative randomized trial is not the same as an ineffective therapy — BROADEN and RECLAIM failed at short blinded endpoints with un-individualized targeting.
  • Open-label benefit can be durable to 8 years (Crowell), and connectomic targeting markedly outperforms anatomical targeting.
  • Target and circuit choice shape both the size and the speed of response (e.g., rapid medial-forebrain-bundle effects).
  • The OCD HDE shows psychiatric DBS can clear a regulatory bar; the depression failures look like problems of selection, targeting, and design.
  • The field’s question has shifted from “does stimulating area 25 work?” to “how do we target the right circuit in the right patient and read out the right biomarker?”

Selected References

  1. Mayberg HS, Lozano AM, Voon V, et al. Deep brain stimulation for treatment-resistant depression. Neuron. 2005;45(5):651–660. Seminal open-label, n=6.
  2. Holtzheimer PE, Husain MM, Lisanby SH, et al. Subcallosal cingulate deep brain stimulation for treatment-resistant depression: a multisite, randomised, sham-controlled trial. Lancet Psychiatry. 2017;4(11):839–849.
  3. Dougherty DD, Rezai AR, Carpenter LL, et al. A randomized sham-controlled trial of DBS of the ventral capsule/ventral striatum for chronic treatment-resistant depression. Biol Psychiatry. 2015;78(4):240–248.
  4. Crowell AL, Riva-Posse P, Holtzheimer PE, et al. Long-term outcomes of subcallosal cingulate deep brain stimulation for treatment-resistant depression. Am J Psychiatry. 2019;176(11):949–956.
  5. Schlaepfer TE, Bewernick BH, Kayser S, et al. Rapid effects of deep brain stimulation for treatment-resistant major depression. Biol Psychiatry. 2013;73(12):1204–1212.
  6. Riva-Posse P, Choi KS, Holtzheimer PE, et al. A connectomic approach for subcallosal cingulate DBS surgery: prospective targeting in treatment-resistant depression. Mol Psychiatry. 2018;23(4):843–849. Figures abstract-level; confirm against full text.
  7. Scangos KW, Khambhati AN, Daly PM, et al. Closed-loop neuromodulation in an individual with treatment-resistant depression. Nat Med. 2021;27(10):1696–1700. n=1 proof-of-concept; abstract-level.