DBS Targeting

1.A Short History of How We Got Here

Stereotactic neurosurgery started with Horsley and Clarke in 1908, who built a device for placing electrodes into the cerebellum of monkeys using external skull landmarks as references. The system worked for non-human primates because their skulls are predictably related to their brains. In humans, it does not — variability in skull shape, sinus pneumatization, and brain position relative to the cranium makes external landmarks an unreliable proxy for deep targets.

The breakthrough came in 1947, when Spiegel and Wycis published the first human stereotactic frame and the first human stereotactic atlas, using ventriculography — air or contrast injected into the lateral ventricles — to identify intracranial reference points directly. The pineal gland, the foramen of Monro, and the third ventricle outline could now be seen on plain films, and structures could be referenced to them rather than to skull landmarks.

Talairach refined this through the 1950s into the system we still use. He recognized that two small white-matter commissures — the anterior commissure (AC) and the posterior commissure (PC) — could be reliably identified on ventriculography (and later on MRI), and that the deep gray matter structures of the basal ganglia and thalamus hung from the AC-PC line in a remarkably consistent geometry. The AC-PC system, codified in his 1957 Atlas d'Anatomie Stéréotaxique, became the universal coordinate language for stereotactic neurosurgery.

Schaltenbrand and Bailey published a printed brain atlas in 1959; Schaltenbrand and Wahren expanded it in 1977 into the version most of you have seen — a series of axial, coronal, and sagittal plates spaced 1–2 mm apart, each showing the deep gray matter referenced to AC-PC. For four decades, indirect targeting using the Schaltenbrand-Wahren atlas was the only way to plan a DBS case.

The shift to MRI-based stereotaxis in the 1990s — first T2 for direct visualization of the STN, then SWI/QSM, then proton density and FGATIR for thalamic and pallidal subnuclei — added a layer rather than replacing the substrate. Direct visualization tells you where this patient's STN actually is. AC-PC is still how you describe that location, register it to the workstation, and verify it on post-op imaging.

2.The AC-PC Line and Why It Persists

The anterior commissure is a thin, oval white-matter tract that crosses the midline at the rostral end of the third ventricle, just above the optic chiasm. On a sagittal midline image it appears as a small white spot superior to the optic recess and inferior to the floor of the lateral ventricle. The posterior commissure crosses at the rostral end of the cerebral aqueduct, where the third ventricle narrows into the aqueduct. The line connecting their geometric centers — the intercommissural line — defines the y-axis of every stereotactic atlas in current use.

The deep brain hangs from this line. The thalamus sits above it, the substantia nigra and STN below it, the GPi anterolateral and superior to it. These structures' positions relative to AC-PC vary by only a few millimeters across the population, while their positions relative to skull landmarks or cortical sulci vary by a centimeter or more. This is why AC-PC has outlasted ventriculography, frame-based stereotaxis without imaging, and the predictions that direct visualization would render coordinate-based planning obsolete.

The convention used at UPMC, and at most modern DBS centers, is to set the mid-commissural point (MCP) as the origin: x = 0 at midline, y = 0 halfway between AC and PC, z = 0 in the AC-PC plane. Some older literature (and some European centers) uses the anterior commissure as origin. Be careful when reading older papers — a coordinate of "y = 12" means very different things in MCP-origin and AC-origin systems.

Standard sign conventions:

• X — lateral distance from midline (positive = right or left depending on local convention; many systems use unsigned X with a separate L/R designation)
• Y — anterior–posterior, positive = anterior to MCP
• Z — superior–inferior, positive = superior to AC-PC plane

So the UPMC STN starting recipe of 12 / -3 / -4 means: 12 mm lateral from midline, 3 mm posterior to MCP, 4 mm inferior to the AC-PC plane. Memorize this notation. You will be asked for coordinates in the OR, and "twelve, minus three, minus four" should leave your mouth before you have to think about it.

3.Identifying AC and PC Reliably

This is the single most error-prone step in indirect targeting, and the one most likely to be delegated to the workstation's auto-detect feature without verification. Don't.

Sequence. A high-resolution sagittal T1 (or a 3D MPRAGE reformatted to sagittal) is the workhorse. Proton-density and FGATIR are alternatives that show the commissures with excellent contrast.

Where to look. True midline sagittal slice — confirm by checking that the cerebral aqueduct, fornix, and septum pellucidum all align in a single plane. If they don't, your image is off-midline and your AC and PC marks will be off. AC appears as a small bright oval just superior to the optic recess of the third ventricle, anterior to the columns of the fornix. PC is a small bright structure at the junction of the third ventricle and the cerebral aqueduct, posterior to the habenular commissure.

Pitfalls.

• Habenular commissure mistaken for PC. The habenular commissure sits 1–3 mm anterior to PC and is a common error in beginners. PC is the commissure at the aqueduct, not the one above it.
• Lamina terminalis mistaken for AC. AC is a discrete oval; the lamina terminalis is a thin sheet anterior to AC and inferior to it.
• Tilted head. If the patient is not centered in the scanner, the apparent midline sagittal is not the true midline. Always reformat to the AC-PC plane after marking.
• Hydrocephalus or large ventricles. Distorts the geometry and makes both commissures harder to find. In these patients, FGATIR or proton-density is more forgiving than T1.

A useful sanity check: after you've placed AC and PC, the workstation will report the AC-PC distance. Normal is 24–28 mm, mean ~26 mm. Anything outside that range — a 22 mm AC-PC or a 32 mm AC-PC — should make you re-check your marks before you proceed.

4.Indirect Targeting: The Atlas Approach

Indirect targeting is the application of standard, atlas-derived coordinates relative to AC-PC, without reference to whether the structure is actually visible on this patient's MRI. The Schaltenbrand-Wahren atlas is the canonical source. The starting coordinates we'll discuss in Part III are all indirect.

The strength of indirect targeting is robustness. It works in patients with poor MRI quality, in dystonia where direct visualization is unreliable, in centers without high-field magnets. It is also the language of comparison: when you read a paper reporting "active contact at 11.2 / -1.9 / -3.4," you can interpret that coordinate against your own practice without seeing the imaging.

The weakness is that it ignores individual anatomical variation. STN volumes vary by a factor of two across individuals. Brains with significant atrophy, large ventricles, or asymmetry between hemispheres will have STNs that sit slightly differently than the population mean. A pure indirect approach will be off in these patients.

5.Direct Targeting: The MRI Approach

Direct targeting is identification of the actual structure on the actual patient's MRI, with coordinates derived from the visualized anatomy rather than from an atlas. The technique was pioneered by Bejjani and Dormont (2000) at the Pitié-Salpêtrière, who showed that the STN was directly visible as a hypointense ovoid on multiplanar T2 sequences and that targeting the directly visualized structure improved accuracy over pure atlas coordinates. Sequences and contrast mechanisms have evolved since, but the principle has not: you target the structure you can see, and you use the indirect coordinate as a cross-check.

The UPMC workhorse: FGATIR. We obtain Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR) on every DBS case. FGATIR uses an inversion-recovery preparation tuned to null white-matter signal, producing strong contrast between gray-matter structures and the surrounding white-matter tracts. For the STN, FGATIR shows the nucleus with usable intensity differentiation from the adjacent internal capsule and substantia nigra, complementing what SWI shows by iron-based contrast. For the thalamus, FGATIR shows the thalamus–internal capsule boundary clearly and helps confirm the lateral coordinate for Vim, even though the Vim/Vop/Voa subnuclei themselves remain atlas constructs. For the GPi, FGATIR shows the medial medullary lamina (separating GPi from GPe) and the inferior border with the optic tract better than any other sequence we use routinely.

STN-specific. SWI/QSM remains the highest-contrast sequence for the STN itself — the iron content of the nucleus produces a hypointense almond-shaped signal on axial slices through the upper midbrain that is unmistakable. T2 works as well but with less sharp borders. We use SWI alongside FGATIR; the two sequences are complementary, and disagreement between them is informative.

GPi-specific. Proton density is a useful adjunct to FGATIR for the medial medullary lamina, particularly in older studies and on lower-field magnets. T2 is acceptable but lower contrast.

Vim. The hard one. Standard MRI does not show Vim/Vop/Voa subnuclei directly. FGATIR clarifies the lateral and dorsal boundaries against the internal capsule. DRTT tractography on diffusion imaging can guide Vim placement, but the technique is center- and dataset-dependent. For most centers, including ours, Vim remains predominantly an indirect target with FGATIR-confirmed boundaries.

6.The Synthesis: Starting Coordinate as Sanity Check

The right framework for a modern DBS planner is neither pure indirect nor pure direct, but a layered approach:

1. Calculate the indirect coordinate using the standard recipe (Part III). This is your starting point and your sanity check.
2. Visualize the structure directly on the appropriate sequence. Adjust the coordinate based on what you actually see.
3. Compare. If the direct and indirect coordinates agree within 1–2 mm, you're confident. If they disagree by 3 mm or more, stop and find the source of the discrepancy before you trust either one.

The most common cause of a large indirect/direct mismatch is not anatomical variation — it is a registration or AC-PC marking error. If your direct coordinate puts the STN where the indirect coordinate doesn't, you may have misidentified PC, mis-set the AC-PC plane, or have an image distortion problem. Ventricular enlargement is a real cause of mismatch in elderly patients but tends to be modest (1–2 mm) rather than dramatic.

The starting coordinate is your sanity check, not your endpoint. The directly visualized anatomy is your endpoint, not your starting point. Skipping either step is how avoidable misses happen.

7.Subthalamic Nucleus (STN)

The STN is an ovoid lens-shaped structure at the junction of the diencephalon and midbrain, bordered by the internal capsule laterally, the zona incerta and thalamus superiorly, the red nucleus posteromedially, the substantia nigra ventrally, and oculomotor fibers anteromedially. It is functionally tripartite, with the dorsolateral two-thirds being sensorimotor (the DBS target), the ventromedial portion being associative/limbic. The entire nucleus is small — typically 6 × 4 × 5 mm — which is why a 1 mm error matters here in a way it does not in larger structures.

Direct visualization. We use FGATIR and SWI together for every STN case. On axial SWI through the upper midbrain, you will see two paired hypointense structures: the larger, more medial pair are the red nuclei; the more lateral, almond-shaped pair are the STNs. The substantia nigra sits just inferior and slightly more lateral to the STN, and on adjacent slices the STN and SN can appear continuous — the SWI signal alone does not always tell you where the STN ends and the SN begins. FGATIR at the same level resolves this: the STN appears with usable intensity differentiation from the adjacent internal capsule (lateral border) and from the substantia nigra (inferior border), confirming that the SWI hypointensity is true STN. Use the axial slice at maximum red nucleus diameter — typically 3–4 mm below MCP — for your reference. This is also the slice where Bejjani's line is drawn.

The Bejjani line. Bejjani et al. (2000) introduced an individualized anatomical landmark that every trainee should know and most don't use enough. The technique:

1. On axial T2 or SWI at the slice showing the red nucleus at maximum diameter (≈ 4 mm below MCP), draw a line through the anterior border of the red nucleus, perpendicular to the AC-PC line.
2. The intersection of this line with the medial border of the STN (or the lateral edge of the RN) defines the y-coordinate of the target.
3. Move 3 mm lateral from the lateral border of the RN to find the typical x-coordinate.

Why this matters: the RN is a much larger and more reliably visualized structure than the STN itself. In patients where the STN is poorly seen on direct imaging — older 1.5 T scanners, motion artifact, atypical anatomy — the RN gives you a robust patient-specific anatomical reference for the y-coordinate that the indirect recipe alone cannot. The published validation studies (including the Karger group's cross-center evaluation and recent Lead-DBS atlas work) consistently show that Bejjani-line-derived coordinates agree closely with the modern probabilistic sweet spot, and that the dorsolateral STN sweet spot lies approximately 2 mm below the superior margin of the RN at 3 mm lateral to its lateral edge.

A practical workflow: calculate 12 / -3 / -4. Pull up the axial SWI at the RN's maximum diameter. Draw the Bejjani line. Confirm that 12 / -3 / -4 lands within the directly-visualized STN, just lateral to the lateral RN border, in the dorsolateral portion of the nucleus. If it doesn't, ask why — and believe the anatomy over the recipe once you've ruled out a registration error.

Other anatomical landmarks worth knowing. The mammillothalamic tract (MTT) can be seen on axial proton-density or T2 just medial to the STN — when visible, it provides another patient-specific medial reference. The interpeduncular cistern and the contour of the cerebral peduncle anchor the anterior border. The Sukeroku sign (the appearance of the STN as an oval hypointensity with the SN below, resembling the silhouette of the kabuki character) is a Japanese-school descriptive landmark useful as a recognition gestalt. Trainees who only know "twelve, minus three, minus four" and not these landmarks are vulnerable when the recipe and the anatomy disagree.

Common misses and what they mean.

• Too lateral → internal capsule → stim-induced contralateral muscle contractions, reduced therapeutic window
• Too medial → limbic/associative STN → mood lability, hypomania, cognitive side effects
• Too anterior → associative STN → cognitive and behavioral side effects without good motor benefit
• Too posterior → sensory thalamus or medial lemniscus → paresthesias as the dominant side effect
• Too superior → zona incerta / thalamic fasciculus (sometimes useful for tremor, but not the PD target)
• Too inferior → substantia nigra → mood depression on stimulation

8.Ventral Intermediate Nucleus (Vim)

Vim is the most demanding indirect target in functional neurosurgery because, unlike STN and GPi, the subnuclei themselves are not reliably visible on standard MRI. FGATIR — which we obtain on every DBS case — shows the thalamus as a distinct gray-matter structure and clarifies the lateral boundary against the internal capsule, narrowing the lateral coordinate; but the Vim/Vop/Voa boundaries within the thalamus remain atlas constructs that you reach by recipe.

The 25% rule. This is a Schaltenbrand-derived heuristic: the Vim sits at approximately 25% of the AC-PC distance anterior to PC, in the AC-PC plane (z = 0), at 14–15 mm from midline. Working through the math: if AC-PC is 26 mm in this patient, then 25% of 26 = 6.5 mm anterior to PC, which is 13 - 6.5 = 6.5 mm posterior to MCP, so y ≈ -6 to -7. The "25% from PC" formulation is preferable to a fixed millimeter offset because it scales with AC-PC length, which varies across patients.

X coordinate — a contested point. Two conventions exist:

• Absolute distance from midline (the UPMC convention): 14–15 mm lateral, regardless of third ventricle width.
• Relative to the third ventricle wall: measure the distance from midline to the lateral border of the third ventricle, then add 11 mm. This convention adapts to wide third ventricles, which push the Vim laterally; it is preferred at some European centers.

The two converge in patients with normal-sized ventricles. They diverge in patients with significant ventricular enlargement, where the absolute-from-midline approach risks placing the lead too medial. We use the absolute coordinate but check the third-ventricle width in every Vim case as a sanity check.

The Vop/Voa distinction. Vim is the receiving nucleus for the cerebellum (via DRTT) and the standard target for tremor (essential tremor, parkinsonian tremor, MS tremor). Vop sits 2 mm anterior to Vim and is the receiving nucleus for pallidum; some centers use Vop for parkinsonian tremor with rigidity, or as a combined Vim/Vop target. Voa sits another 2 mm anterior and receives input from GPi; it is occasionally targeted for dystonic tremor or as a combined Voa/Vop target for Holmes tremor and post-traumatic tremor syndromes. The starting recipe — Vim, Vop +2, Voa +4 — captures the geometry; the indication drives which one you use.

Anatomical landmarks for confirmation. Vim sits just superior to the AC-PC plane (z = 0 to z = +2 in some practices), with the internal capsule immediately lateral. The thalamus–internal capsule border is visible on FGATIR. A trajectory passing through the lateral thalamus toward Vim that ends up in the IC means you are too lateral; clinical clues during awake testing include contralateral motor contractions or paresthesias.

The cZi alternative. The caudal zona incerta / posterior subthalamic area, ~15 mm lateral, 6–7 mm posterior to MCP, 4 mm below AC-PC, is an emerging target for tremor with comparable or superior efficacy to Vim in several series (Plaha and Blomstedt, 2009 onward). It is not the UPMC primary tremor target, but trainees should know that "Vim is the only tremor target" is no longer accurate.

9.Globus Pallidus Internus (GPi)

GPi is the largest of the three workhorse targets and the most forgiving. It is bordered by the internal capsule medially (the anterior limb anteriorly, the genu and posterior limb posteriorly), GPe laterally (across the medial medullary lamina), the optic tract inferiorly, and the basal nucleus of Meynert anteriorly.

Direct visualization. Proton density and FGATIR show the medial medullary lamina (separating GPi from GPe) and the inferior border with the optic tract. SWI is less helpful here than for STN because the iron content of GPi is more uniform. T2 is acceptable but lower contrast.

Anatomical landmarks.

• The optic tract sits ~3 mm below the typical GPi target. The z = -4 starting coordinate places you safely above it. Closer than 2 mm and the patient will have phosphenes on stimulation.
• The internal capsule sits medial. As you move posteriorly along the long axis of the pallidum, the IC bows laterally — a common reason for anteriorly-aimed trajectories ending up inadvertently medial at depth. This is one reason posteroventrolateral targeting is forgiving anteriorly but unforgiving posteriorly.
• The basal nucleus of Meynert lies just anterior to the GPi target. Trajectories aimed too anterior will pass through it; the cognitive consequences are debated but worth being aware of.

Functional organization and the trajectory. GPi is also functionally tripartite, with the sensorimotor portion in the posteroventrolateral region — y = +2 (slightly anterior to MCP), x = 22 (well lateral), z = -4 (just above the optic tract). A long-axis trajectory through the GPi from a frontal entry point will sample associative dorsally and sensorimotor ventrally, with the goal of placing the ventral contacts in the sensorimotor zone.

Common misses.

• Too medial → internal capsule → stim-induced contractions, dysarthria
• Too lateral → GPe (which is not the therapeutic target, though it's not dangerous)
• Too inferior → optic tract → phosphenes (resolves immediately on turning stim off, but the lead position is wrong)
• Too posterior → posterior limb of internal capsule
• Too dorsal → associative GPi → mediocre motor benefit, possible cognitive effects

Why GPi is more forgiving than STN. The structure is several times larger, the side-effect tolerances are more graceful (a 2 mm lateral STN miss puts you in the IC; a 2 mm lateral GPi miss puts you in GPe, which is functionally inert), and the dyskinesia-suppressing effect of GPi stimulation is robust across a wide range of contact positions. This is one reason GPi is often recommended for surgeons earlier on the learning curve, and one reason CSP 468 found equivalent motor outcomes despite the practical difficulty difference.

10.The Surface Decision

Entry-point planning gets attention from every trainee — usually the right kind of attention. The standard considerations:

• Distance from midline: typically 2–3 cm lateral to the sagittal midline, depending on target.
• Distance from coronal suture: anterior to the suture for STN and GPi long-axis trajectories; the precoronal frontal entry is the workhorse.
• Avoid superficial bridging veins draining to the superior sagittal sinus. These are the highest-risk surface vessels. T1 with gadolinium and dedicated MR venography are essential — the contrast study is non-negotiable. A tractography-quality T2 SPACE without contrast will not show bridging veins reliably.
• Avoid sulcal entries. A burr hole that lands over a sulcus risks injury to pial vessels in the sulcal floor and increases CSF egress with subsequent pneumocephalus and brain shift. Aim for a gyral crown.
• Plan around the motor strip. Especially for GPi long-axis trajectories that come in more posteriorly, avoid the precentral gyrus.

11.The Underrated Part: The Deep Trajectory

This is where most trainee planning falls short and where most senior surgeons earn their reputation. The trajectory between the cortex and the target — particularly the portion above 15 mm from target, where the microelectrode/macroelectrode is deployed and the trajectory is committed — gets less attention than it deserves.

The key conceptual point: at our institution and most others, the cannula stops about 15 mm above target, and the microelectrode (or directly the DBS lead) extends from there to the target. The deep portion of the trajectory — from cortex to ~15 mm above target — is committed once the cannula is in place. You cannot adjust at that depth without backing out and re-planning. So that portion of the trajectory must be vessel-free before the case starts.

Vessels at risk along the deep trajectory:

• Lateral lenticulostriate vessels lining the lateral border of the caudate, supplying the striatum. A trajectory cutting through the body of the caudate or transgressing the caudate-IC border passes through this territory.
• Subependymal veins lining the lateral wall of the lateral ventricle. Transventricular trajectories cross these.
• Internal cerebral veins and septal veins medially, near the foramen of Monro. A medial trajectory aiming for a centrally-placed target is particularly vulnerable.
• Choroidal vessels in the choroid plexus and the choroidal fissure — relevant for trajectories that pass through or near the ventricle.
• Vessels in sulcal walls deep to the entry point. A trajectory aimed slightly off-axis from the gyral long axis will cross a sulcus at depth even if the entry is on a gyral crown.

Workflow for deep-trajectory review:

1. After choosing entry and target, generate a probe's-eye view along the trajectory.
2. Scroll the T1 + gadolinium sequence along the trajectory from entry to ~15 mm above target. Look for any vessel within 2–3 mm of the planned path.
3. Re-scroll on SWI/SWAN for the same depth range. SWI will show small veins that T1+Gd misses.
4. Pay particular attention to the first 4 cm of the trajectory below cortex — this is where most peri-procedural hemorrhages originate, and it is the portion many trainees do not review systematically.
5. If the deep trajectory has a vessel within 2 mm, adjust the entry point, not the target. Move the entry 5–10 mm in any direction that opens the deep path; the target stays fixed.

Avoid the ventricle when possible, but not at all costs. A transventricular trajectory roughly doubles the risk of clinically significant hemorrhage in most series. It also increases CSF egress and brain shift. But forcing an extraventricular trajectory at the cost of crossing a major vessel or transgressing the caudate body is a worse trade. The decision is patient-specific and depends on the geometry of the individual brain.

Avoid the caudate when possible. Transcaudate trajectories cross the lateral lenticulostriate territory and are associated with a measurable increase in hemorrhage rate in published series. A trajectory that grazes the caudate-IC border is acceptable; one that transits the body of the caudate is worth re-planning.

12.Trajectory Verification

Before signing out the plan:

1. Coronal and sagittal reformats along the trajectory — both with and without contrast. The vessel you missed will be on the slice you didn't reformat.
2. Probe's-eye view scrolled from entry to target.
3. Final coordinate confirmation against the directly-visualized target. The plan should make sense at every depth.
4. Document the rationale in the plan — entry point, target, why this trajectory rather than an alternative. Future-you (debriefing a complication) will thank present-you.

A practical step-by-step for the trainee sitting at the workstation for the first time. This is the workflow we use; centers vary in details but the steps are common.

1. Verify imaging. T1+Gad, T2, SWI, and FGATIR are present and free of motion artifact. Re-scan if there is motion that affects the target region.
2. Co-register CT to MRI. The workstation does this automatically; verify by checking the alignment of bony landmarks and ventricular margins. Misregistration here propagates through the entire plan.
3. Mark AC and PC on midline sagittal T1. Verify true midline by checking aqueduct, fornix, septum alignment. Confirm AC-PC distance is in the 24–28 mm range.
4. Set MCP as the origin. The workstation now reports coordinates relative to MCP.
5. Apply indirect coordinate: 12 / -3 / -4 for STN, contralateral to the more-affected side first (or per surgeon preference).
6. Cross-check on axial SWI at maximum RN diameter. The point should land in the lateral STN; the dorsolateral sweet spot is approximately 2 mm dorsal and just lateral to the upper border of the RN.
7. Cross-check on FGATIR at the same level to confirm the STN borders against the internal capsule and substantia nigra.
8. Draw the Bejjani line through the anterior RN border. Confirm the y-coordinate visually.
9. Adjust if direct and indirect disagree by >2 mm. Investigate the source first (registration? AC-PC marks? anatomical variant?). When you trust the discrepancy, prefer the directly visualized target over the recipe.
10. Plan trajectory. Choose entry point on a gyral crown, anterior to coronal suture, ~3 cm lateral to midline. Avoid bridging veins on T1+Gad.
11. Probe's-eye review of the entire trajectory from entry to target on T1+Gad and SWI. Look for vessels within 2 mm of the path. Particular attention from cortex to 4 cm depth and along the deployment portion (15 mm above target down to target).
12. Adjust entry point if any deep vessel is too close. Target stays fixed; entry moves.
13. Coronal and sagittal reformat verification. Document the plan and rationale.
14. Repeat for contralateral side. Plans should be symmetric for x; y and z usually identical, sometimes adjusted 1 mm for asymmetric anatomy.
15. Sign out, present at functional conference, revise based on team input. The senior surgeon's marks on your plan are the most efficient learning tool you have. Save them.

What comes after planning. Preoperative planning ends when the plan is signed out at functional conference. Case-day workflow then adds two layers of intraoperative confirmation that the plan does not capture. Microelectrode recording (MER) maps the trajectory through the target by sampling single-unit activity along the descent — the dorsal border of the STN is identified by a characteristic shift to high-frequency, irregular, movement-related firing; the ventral exit into substantia nigra is recognized by a distinct firing pattern; the GPi is identified by high-frequency tonic discharge with characteristic border cells separating it from GPe. Macrostimulation then tests therapeutic benefit (tremor suppression, rigidity reduction) and side-effect thresholds (capsular contractions, paresthesias, phosphenes for GPi, mood/autonomic effects for STN) at planned and adjacent contact positions. We use MER on every case where it's feasible. Awake versus asleep DBS, MER versus image-guided-only approaches, and the role of evoked resonant neural activity (ERNA) in modern intraoperative confirmation are subjects worth their own readings; here it suffices to note that the plan is the starting point of case day, not the endpoint.