Foundations & Circuit Targets

Evidence status. In the United States, DBS for severe treatment-resistant OCD is available under a Humanitarian Device Exemption; other psychiatric DBS and focused-ultrasound lesion procedures remain investigational or off-label. Network models are useful but not deterministic.

Burckhardt to Moniz: the birth of psychosurgery

The first deliberate operations on the brain to treat mental illness are usually traced to Gottlieb Burckhardt, a Swiss psychiatrist who between 1888 and 1889 excised areas of cortex, mainly temporal and sometimes frontal, in six agitated psychotic patients, reporting the series in 1891. His hypothesis was specific: that agitation arose from hyperactivity in a sensory center that could be disconnected. One patient died within days of surgery and another died by suicide after discharge, and the hostile reception left the idea dormant for half a century. It was revived in 1935 by the Portuguese neurologist Egas Moniz, working with the neurosurgeon Almeida Lima, who introduced prefrontal leucotomy, severing white-matter connections between the prefrontal cortex and deeper structures. Moniz reported improvement in agitation and affective symptoms, and in 1949 received the Nobel Prize in Physiology or Medicine (shared with W.R. Hess), the only one ever awarded for a psychosurgical procedure. The conceptual premise, that disconnecting frontal-subcortical pathways could relieve psychiatric suffering, was not entirely wrong; the execution and the safeguards were catastrophically immature.

The lobotomy and its collapse

In the United States, Walter Freeman, a neurologist and neuropsychiatrist who was not a trained surgeon, popularized the operation with the neurosurgeon James Watts. Freeman then radically simplified it on his own, adopting Fiamberti's transorbital route. Watts objected that the procedure was too dangerous to perform outside an operating room, grew critical of how patients were being selected, and ended the collaboration. Freeman's transorbital lobotomy (an orbitoclast driven through the orbital roof and swept to disconnect frontal white matter, performed without an operating room and often with electroconvulsive shock as the anesthetic) could be done in minutes and was applied on an enormous scale, often with minimal diagnostic rigor and no meaningful outcome measurement. Freeman personally performed several thousand lobotomies (commonly cited as about 3,400, roughly 2,500 of them transorbital) and traveled the country teaching the technique; estimates of the total number of lobotomies of all types performed in the United States between 1936 and the late 1950s are on the order of 40,000 to 50,000. The procedure produced frequent, severe, and irreversible harm: apathy, disinhibition, blunted personality, seizures, and death. Several forces ended the era nearly at once: the arrival of chlorpromazine and the first effective antipsychotics in the 1950s offered a non-surgical alternative; outcome data and personal accounts exposed the human cost; and a broad ethical and public backlash followed. By the 1970s, psychosurgery in the United States was the subject of intense scrutiny, culminating in review by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, whose 1977 report, importantly, did not ban the practice outright but concluded that carefully selected, narrowly indicated procedures could be appropriate under rigorous oversight. That conclusion, that the answer to harm was governance and precision rather than prohibition, set the template for everything that followed.

From freehand disconnection to targeted lesions

The decisive technical advance was human stereotactic surgery, introduced by Spiegel and Wycis in 1947, which allowed a defined intracranial point to be reached through a small opening using a coordinate frame and imaging. Applied to psychiatric disease, stereotaxis replaced the broad frontal sweep with small, reproducible lesions placed in specific white-matter tracts and limbic relay points. Four operations defined this era and, remarkably, all four survive in some form today:

• Anterior capsulotomy (developed by Talairach and refined by Leksell): a lesion in the anterior limb of the internal capsule interrupting prefrontal-thalamic fibers; the dominant ablative procedure for OCD.
• Anterior cingulotomy (Ballantine and colleagues, Boston): a lesion in the dorsal anterior cingulate; used for OCD and chronic pain.
• Subcaudate tractotomy (Knight, London): a lesion in the substantia innominata below the head of the caudate, principally for depression.
• Limbic leucotomy (Kelly): a combination of cingulotomy and subcaudate tractotomy.

These procedures were smaller, anatomically defined, and far safer than the lobotomy, and in carefully selected refractory OCD and depression they produced meaningful response rates that remain broadly consistent with contemporary lesioning data. They re-established that the goal was a specific node in a specific circuit, not a region of cortex. Their detailed technique and outcomes are covered on the core indications page.

Borrowing a tool from movement disorders

By the late 1990s, deep brain stimulation had matured as a treatment with adjustable stimulation for Parkinson disease and tremor. The conceptual leap was to apply the same hardware to the same anatomical targets that ablative psychiatric surgery had already explored, but with an electrode instead of a lesion, trading a planned destructive lesion for adjustable stimulation and ongoing hardware care. In 1999, Nuttin and colleagues reported the first DBS for OCD, placing electrodes in the anterior limb of the internal capsule, the same target as capsulotomy. In 2005, Mayberg and colleagues reported DBS of the subcallosal cingulate (Brodmann area 25) for treatment-resistant depression, an experiment driven explicitly by a functional-imaging model of a depression circuit rather than by lesion tradition. These two reports, more than any others, launched modern psychiatric neuromodulation.

Regulatory recognition and the humility that followed

In 2009 the FDA granted a Humanitarian Device Exemption (HDE) for Reclaim DBS Therapy for OCD, indicated for bilateral stimulation of the anterior limb of the internal capsule as an adjunct to medications and as an alternative to anterior capsulotomy in adults with chronic, severe, treatment-resistant OCD who have failed at least three SSRIs. The HDE pathway applies to humanitarian-use devices and requires probable benefit rather than the premarket-approval standard of demonstrated effectiveness, with local IRB or appropriate committee oversight for clinical use. OCD DBS thus became, and remains, the only FDA-recognized psychiatric DBS indication a program can offer clinically under HDE oversight. The same period delivered a sobering lesson: the industry-sponsored randomized trials intended to confirm DBS for depression, BROADEN (subcallosal cingulate) and RECLAIM (ventral capsule/ventral striatum), failed their prespecified endpoints, and both were stopped after futility analyses. Long-term open-label follow-up of the same cohorts nonetheless showed durable benefit in many patients, exposing a deep methodological problem: blinding, sham design, endpoint timing, and patient heterogeneity in psychiatric DBS are genuinely hard, and open-label promise does not survive naively into a randomized design. The field's response was refinement: biomarker-guided and connectomic targeting, individualized programming, and adaptive/closed-loop approaches, explored on the core indications page.

The return of incisionless lesioning

The adjustability of DBS did not make lesioning obsolete; implantation itself still carries potentially permanent risks. Two incisionless modalities revived the targeted-lesion concept with modern precision: stereotactic radiosurgery (Gamma Knife anterior/ventral capsulotomy), delivering a focused necrotizing dose without an incision, and magnetic resonance-guided focused ultrasound (MRgFUS) capsulotomy, which creates a thermal lesion under real-time MR thermometry with no hardware and no implanted infection risk. For a patient who cannot or will not accept implanted hardware and lifelong device management, a one-time lesion remains a rational choice; these modalities are detailed on the core indications page.

Governance as the price of legitimacy

The defining feature of modern psychiatric neurosurgery is that it is governed. The 2014 multi-society consensus on guidelines for stereotactic neurosurgery for psychiatric disorders, developed jointly by the world and regional stereotactic and functional neurosurgery societies and the World Psychiatric Association, codified the expectations: an independent multidisciplinary committee, documented treatment refractoriness, capacity-based informed consent, validated pre- and post-operative outcome measurement, prospective data collection or registry participation, and conduct within an appropriate regulatory framework. These are not bureaucratic accretions; they are the operational form of the lesson learned from the lobotomy. The companion selection and program page translates this framework into practice.

The arc, then, is coherent: a history of serious harm that underscores why anatomical precision, credible evidence, patient autonomy, and independent oversight must advance together. A trainee who internalizes that arc will read every subsequent indication and target on these pages with the right combination of optimism and discipline.

Loops, not regions

The cortico-striato-thalamo-cortical model organizes frontal-subcortical function into parallel loops that originate in cortex, project to striatum, relay through the pallidum and substantia nigra pars reticulata to thalamus, and return to the originating cortex. The loops are not interchangeable: the limbic loop passes through the ventral striatopallidal complex and the mediodorsal thalamus, in parallel with the dorsal motor loop's route through the internal pallidum and the ventrolateral thalamus, which is the reason psychiatric targets cluster ventrally. Three are central to psychiatric disease: a limbic/affective loop (anterior cingulate and medial orbitofrontal cortex to ventral striatum), an orbitofrontal loop (lateral OFC to ventromedial caudate), and a dorsolateral prefrontal loop (cognitive control). Within each loop, a direct pathway facilitates and an indirect pathway suppresses thalamocortical output, while the hyperdirect pathway gives frontal cortex a fast cortico-subthalamic route to brake behavior. OCD is classically modeled as relative overactivity of the direct relative to the indirect pathway in orbitofrontal-caudate circuitry, producing an inability to terminate behavioral and cognitive sequences.

Beyond the segregated loops

The strictly parallel, segregated-loop picture has been enriched. As Milad and Rauch summarized, contemporary imaging implicates not only cortico-striatal circuitry but the lateral and medial orbitofrontal cortices, the dorsal anterior cingulate, and amygdalo-cortical circuitry, and links OCD to impaired fear extinction in addition to deficient behavioral inhibition. The practical consequence for neuromodulation is that effective targets are not confined to one node: interrupting or modulating the fiber traffic that binds OFC, dACC, ventral striatum, thalamus, and the subthalamic region can all influence the same clinical syndrome, which is exactly what the target diversity in OCD DBS demonstrates.

One piece of anatomy, many target names

The anterior limb of the internal capsule (ALIC) is a mixed corridor. It carries the anterior thalamic radiation, a reciprocal system linking the mediodorsal and anterior thalamic nuclei with prefrontal and cingulate cortex, together with descending prefrontal corticopontine fibers and fibers connecting caudate and putamen. Its ventral portion additionally carries prefronto-subthalamic hyperdirect fibers, and the corridor is topographically ordered: orbitofrontal, ventromedial and anterior cingulate fibers run ventromedially while dorsolateral prefrontal fibers run dorsolaterally, which is why moving a contact a few millimeters changes which prefrontal territory is engaged. Its ventral extent, where the capsule meets the ventral striatum, is the ventral capsule/ventral striatum (VC/VS); embedded nearby are the nucleus accumbens (NAc) and the bed nucleus of the stria terminalis (BNST). The historical confusion of target names (capsulotomy, ALIC-DBS, VC/VS-DBS, NAc-DBS, BNST-DBS) largely reflects slightly different points along a single fiber corridor and the gray matter immediately abutting it. The nucleus accumbens lies inferomedially, beneath the head of the caudate; the bed nucleus of the stria terminalis lies superomedially, at about the AC-PC plane, wrapped around the anterior commissure and running posteriorly along the stria terminalis. This is why an ablative capsulotomy and a ventral-capsule electrode can produce comparable antiobsessional effects: they engage overlapping prefrontal-thalamic and prefrontal-brainstem fibers.

The connectomic reframing

Two lines of work converted this intuition into quantitative targeting. Haber and colleagues, comparing the four principal OCD DBS targets (anterior limb of internal capsule, ventral striatum, subthalamic nucleus, and a midbrain target), showed that different stereotactic locations can converge on overlapping orbitofrontal and anterior cingulate fibers traveling through the internal capsule, while each site also engages its own location-specific connections. Li and colleagues pooled ALIC, NAc, and STN cohorts and identified a candidate fiber pathway, centered on connections between the dorsal anterior cingulate and ventrolateral prefrontal cortices and the anteromedial subthalamic nucleus and consistent with a hyperdirect projection, whose estimated engagement was associated with response across cohorts. This is retrospective, model-dependent evidence derived from a normative connectome, not proof of a universal tract or a prospectively validated target. When Widge and colleagues tested tract-activation models using patient-specific imaging in a small VC/VS cohort, activation correlated with outcome but no model predicted OCD response above chance under cross-validation. The useful lesson is that a therapeutic volume may be better described by network engagement than by a point alone.

The subcallosal cingulate hub

The subcallosal cingulate (SCC, Brodmann area 25) sits at a confluence of mood-regulating networks and was chosen as the first imaging-derived depression DBS target by Mayberg. Its therapeutic effect appears to depend less on the gray-matter node than on a specific bundle of converging white-matter pathways. Riva-Posse and colleagues, comparing responders and nonresponders, found that benefit was most likely when the volume of activated tissue captured a combination of four tracts (the forceps minor, the uncinate fasciculus, the cingulum bundle, and fronto-striatal fibers), a finding that launched tractography-guided SCC targeting. A small prospective open-label cohort then used this template for surgical targeting. This is a promising targeting hypothesis, but the small open-label cohorts do not establish a universally necessary or sufficient bundle.

Reward, capsule, and thalamic depression targets

Depression DBS has been delivered to several nodes, each with a circuit rationale: the VC/VS (shared with OCD, engaging ventral striatal reward circuitry); the superolateral medial forebrain bundle (slMFB); the inferior thalamic peduncle; and the lateral habenula, a key node in negative-reward signaling. The slMFB is a tractography-defined pathway ascending from the ventral tegmental area through the ventral capsule to prefrontal cortex, whose stimulation has produced rapid antidepressant effects in small early studies (Schlaepfer; Coenen). It is best understood as a targeting construct rather than a classically dissected tract: the midbrain target region contains lateral VTA subnuclei, dopaminergic elements, descending fibers, and oculomotor nerve fascicles (the anatomic basis for the transient diplopia seen with slMFB stimulation), and its separability from the adjacent anterior thalamic radiation remains disputed. Comparative ranking of these targets is addressed on the core indications page; the anatomical point here is that they sample different entry points into an interconnected mood-and-reward network.

Targets at a glance

Other indications recruit their own nodes, most importantly the centromedian-parafascicular (CM-Pf) and CM-Voi thalamus and the globus pallidus internus (GPi) for Tourette syndrome, where the relevant loop is sensorimotor/limbic basal ganglia rather than the OCD orbitofrontal loop. The table summarizes the principal targets, their network membership, and their main indications; the clinical pages develop each in depth.