Each target opens with a current interpretation and its main uncertainty. Studies are ordered by clinical relevance and design, not simply by date. Related analyses and follow-ups are grouped with known cohort overlap explained. Historical observations and research background remain optional; source filters, searches and shared finding links still reach every record.
Choose a clinical focus and target, then select a finding in the Evidence panel. The selected study and finding type appear above the brain. Checked tracts and cortical regions provide independent anatomical context; their colors do not show treatment benefit.
Use Show labels for floating names, or hover to reveal one name and its explanation. Click a structure to keep its label and highlight, even with Show labels off; its explanation stays below the brain until you clear the selection. Small brown dots anchor anatomy labels; colored study markers have distinct roles identified in the legend. Reset camera returns to the oblique view while keeping your selections. Reset exploration restores the current target’s suggested view and clears study, finding, search and additional region selections. Targets, Evidence and Layers share one inspector beside the brain. Search atlas also finds lessons and functional background.
Green structures are background anatomy and do not respond to hover or clicks. Selected target and finding anatomy, colored region layers, tracts and study coordinates remain interactive. Blue related anatomy is shown when a finding does not specify a side; it does not imply a bilateral effect. Colors identify what is shown, not treatment efficacy.
Johnson 2026 and Ji 2026 add ten findings and two guided lessons under Mechanism & network research. Animal experiments, exploratory human associations and retrospective network validation are labeled separately. Whole-region surfaces provide orientation; they do not reproduce either study’s network, a stimulation field or a validated clinical target. Johnson’s connectivity change did not survive correction, and Ji did not test DTN-guided treatment in a randomized trial. Source review and limitations
Dixon’s integrative review adds proposed functional roles for sgACC, pgACC, medial and lateral OFC, and lateral PFC. These overlapping roles have different levels of support and do not establish treatment mechanisms. In particular, the sgACC account has limited human causal evidence and the medial OFC account is a hypothesis. The broad rostral ACC parcel does not define pgACC; dACC/aMCC and vmPFC labels depend on each study’s definition.
Tract names follow the source atlas, including the anterior and superior corticostriatal tracts and the frontoparietal segment of the cingulum. General tract lines provide reference anatomy. The separately labeled Li fibers show a released response model; no line represents an individual patient’s treatment pathway. Source tract terminology
sgACC means subgenual anterior cingulate cortex. It is gray matter, where neuronal cell bodies are concentrated. SCC DBS targets nearby white matter, which contains connecting fibers. The pink cortical region helps locate this area but does not mark the exact DBS stimulation site.
Release notes & earlier QA reports
Regional functions · v0.15.0
Dixon 2017 adds functional descriptions, sgACC–pgACC and medial–lateral OFC teaching comparisons, and terminology guidance. Clinical findings and anatomical boundaries retain their existing sources. Review and verification
NMAtlas · v0.14.2
Updated app branding and a password-protected online edition. The online edition includes the atlas and teaching views; original local papers remain on the author’s computer.
Persistent labels & teaching clarity · v0.14.1
Clicking a structure keeps its name connected to the anatomy even with Show labels off. Its explanation remains below the brain. The diagram and 3D view now share lesson-specific descriptions, including the different roles of sgACC in TMS and nearby SCC white matter in DBS. QA scope and checks
Routes and contact topography · v0.13.0
Seguin’s full text, supplement, source spreadsheets and public analysis repository support a new lesson comparing three proposed DLPFC–sgACC routes. Machado’s depression DBS study adds a left/right contact-cluster comparison in millimeters relative to the anterior commissure. Six new findings link to the reviewed papers. Explore the routes · Compare contact locations · Review and limits
Teaching edition · v0.12.0
Learn the circuits pairs five guided lessons with interactive diagrams and the existing reference anatomy. A separate contact comparison preserves nominal lead dimensions and reports different acute experiences at adjacent contacts in a published OCD case. It does not infer patient pathways, stimulation fields or a contact-number rule. Open teaching mode · Review and spatial limits · Full texts needed
Smooth hover and pinned focus · v0.11.3
Hover emphasis now fades smoothly and filters brief pointer crossings. Click a structure, tract or published point to pin its highlight. Click it again, click empty space, press Escape or use Clear focus to release it. While pinned, hovering still shows descriptions without moving the focus. Rotation, zoom and opacity controls preserve a visible pin; changing evidence or hiding that element clears it. Related targets remains available beside the pinned name. Interaction checks and release notes
Visual clarity QA · v0.11.2
Scene titles now sit outside the brain, selected findings are identified explicitly, and small text is easier to read. Reset camera preserves the evidence and layer choices. Repeated captions and outdated marker explanations have been simplified. Scope, fixes and verification
Clearer planning-network views · v0.11.1
The five planning ROI dots have been removed from the brain. Regional shading conveys the findings; the published coordinates remain in the relevant cards’ Interpretation & source location details. Treatment, control and existing connectivity-seed markers retain their roles.
Round 3 evidence expansion · v0.11.0
All five supplied article families are incorporated, with 12 new findings, two revised Fitzsimmons findings and eight published coordinate markers. OCD TMS planning effects are separated from treatment locations and weak pre-SMA associations. New DBS findings cover ALIC stimulation with GPe sensing, SCC long-term outcomes and safety, and preliminary scalp EEG changes. Widge’s mixed-diagnosis replication qualifies the cross-frequency coupling finding.
Source PDFs and supplied PDF, DOCX and XLSX supplements open locally. No original imaging volumes, individual tractograms or electric fields were supplied. Round 3 contributions, review details and remaining limits
Evidence and anatomy QA · v0.10.3
Every finding was checked against its displayed anatomy. Laterality, the Figee lateral-prefrontal approximation and missing SCC landmarks were corrected. Findings, corrections and spatial limits
Citation audit · v0.10.2
All 76 finding citations, the source bibliography, coordinate references and anatomical attributions were checked. Incorrect figure and panel references were corrected, including the SCC chapter figure. Published titles are restored. Audit findings and verification limits
Floating labels · v0.10.1
Use Show labels above the brain to hide or display floating names and their connector lines. Published coordinate markers and existing parcel anchor dots remain visible for the selected view. Hover points, parcel dots or highlighted regions to see names and explanations. Small brown parcel dots identify anatomical label anchors, not treatment coordinates. Broad regional surfaces receive no new target points. Background anatomy stays passive.
This atlas combines the reviewed evidence from your Figee collection with independently sourced population anatomy. Selecting a target displays the available anatomical context. Selecting a finding applies its region-specific laterality. Unhighlighted regions provide orientation where reported laterality is unspecified; new regional estimates retain their stated anatomical limits. A selected finding resets unrelated maps; manually selected comparison overlays are labeled explicitly.
Public pathway expansion · v0.10
Forceps minor, four additional cingulum subdivisions, frontal aslant, VTA and classical MFB models, associative and limbic STN pathways, an estimated pre-SMA–STN subset, the positive-weight Li OCD response fibers, and a local ITP volume are available as optional layers. Dataset audit and remaining gaps
Li fibers are a released response-associated model, separate from general anatomical bundles. The pre-SMA subset is an anatomical estimate. Neither is an individual treatment tractogram.
Softer hover focus · v0.9.4
Surrounding anatomy, tracts, points and labels now retain 50% of their normal opacity during hover, instead of 12%. The hovered element keeps its emphasis; moving away restores the original view.
Background cortex · v0.9.3
Parcels tagged “Background cortical anatomy” remain visible as orientation anatomy but no longer show hover callouts, isolate on hover or act as selection targets. Relevant regions, deep structures, reference bundles and published points retain their interactions. Opaque background surfaces still prevent picking hidden anatomy through them.
SCC anatomical highlight · v0.9.2
The Subcallosal cingulate view now highlights the existing AAL3 sgACC gray-matter surfaces alongside its reference fiber bundles. Nearby NAc and medial orbitofrontal anatomy remain background context. The gray-matter estimate locates the region; it is not the SCC white-matter DBS target or a study-specific tract intersection.
Hover callouts · v0.9.1
Explanations now use a smaller, translucent ivory callout with dark text and a fine connector to the hovered location. The card stays still while you explore the same element, and chooses another side near panel edges. Existing labels, descriptions and hover fading are preserved.
Hover exploration & cortex control · v0.9
Hover over a visible structure, fiber bundle or published point to bring it forward while other elements fade. Its label and short context note identify why it is included. Background parcels are explicitly described as orientation anatomy. Moving away, dragging, zooming, Escape or changing a control restores the view. Hover never changes the selected finding, checked layers or saved link.
The Cortex slider dedicates its first 90% of travel to the useful 0–40% surface-opacity range; its final 10% reaches full opacity. Shared links retain the actual opacity, so earlier views keep their appearance. Region and functional-map opacity remain separate controls.
Tract context & supplemental sources
TMS target views now suggest available structural tract families alongside cortical anatomy. Each preset links its clinical, mechanistic and anatomical rationale beside the fiber controls. Presets are editorial anatomical context, not treatment-specific tract reconstructions. Finding highlights retain their reported laterality while bilateral reference fibers remain independently filterable. Changing a fiber checkbox preserves your custom selection when choosing findings; switching targets restores that target’s suggestions.
An estimated pre-SMA–STN subset is available from released STN pathway geometry; it is not an individual or study-specific tractogram. The exact Jimenez ALIC–OFC tract, the right-insula–left-DLPFC route and complete indirect DLPFC–sgACC routes remain unavailable. Existing corticostriatal fibers are not relabeled as those pathways. Functional-map target views start with a translucent map so underlying fibers remain visible. Earlier TMS links with empty fiber defaults adopt the new suggestions; v0.8 links preserve an explicit empty custom selection.
New supplemental circuit references are listed below. They are separate from the clinical evidence sources and do not add clinical outcome cards. Seguin 2026 was initially reviewed at abstract/data-availability level. Version 0.13 adds full-text, supplement, source-data and public-code review, clinical evidence cards and a route teaching lesson. Study-specific streamline geometry remains unavailable.
Cortical region layers · v0.7
Use the checkboxes beside the fiber controls to combine translucent, differently colored regions across any atlas view. The image legend lists checked regions. The background cortex remains green. Parcel labels and published coordinate points keep their separate anatomical and study roles. Shared VLPFC/OFC parcels carry both colors; right OFC uses the same OFC color. Region opacity is independent of background cortex opacity.
DLPFC, VLPFC and OFC follow broad cortical parcels. New bounded estimates use AAL3: medial superior-frontal anatomy for dmPFC, SMA anterior to the AC plane (MNI y > 0) for pre-SMA, and the subgenual ACC labels for sgACC. Pre-SMA uses the gross anterior-SMA rule described in Ji 2021; the patient-specific superficial restriction and connectivity targets are not reproduced. The dmPFC proxy includes anterior medial cortex and is not a unique cytoarchitectonic definition.
AAL3 surfaces retain their source MNI single-subject coordinates. Small mismatches with the background template are possible; no new nonlinear registration is claimed. These surfaces are anatomical estimates, not validated treatment masks, stimulation fields or individual targets. Clinical laterality remains in the finding and its anatomy summary. Checked regions remain independent reference layers; functional maps temporarily pause their colors.
“Subglossal” is interpreted as subgenual. The uncertain “right bitemporofrontal/bifrontal” name remains unresolved; right OFC is a documented OCD target and a plausible match, not a confirmed identification. VLPFC–amygdala mechanistic citations are separate from treatment-efficacy evidence. One new clinical card covers the negative controlled Dunlop 2020 dmPFC depression trial (published abstract reviewed).
Round 2
20 selected papers contribute 41 findings after separating the two Tyagi tractography networks: 7 depression TMS, 6 OCD TMS, 4 depression DBS and 3 OCD DBS papers. Jimenez is also accessible in depression TMS. Original findings are retained. Each new study includes targeting, sample, design, and interpretation limits. Li and Tyagi share a cohort; participant totals should not be pooled across papers.
The supplied Cole and Taylor papers and supplements are linked locally. Taylor’s visual-abstract correction was checked. Fitzsimmons was initially reviewed from the abstract; its supplied full text and supplements were incorporated in v0.11.0. No uploaded originals were changed.
Controlled depression DBS trials
Holtzheimer 2017 (SCC) and Dougherty 2015 (VC/VS) add two controlled outcome cards. Neither blinded primary comparison demonstrated a significant active-versus-sham benefit. These additions were checked against published abstracts; full text was not independently reviewed. Counts of findings are not counts of independent studies.
Regional cortical parcels are context for broadly named TMS regions, not individual coil locations or simulated electric fields. Whole-structure masks are not substituted for anteromedial STN, lateral habenula, ventral ALIC, or other smaller clinical subdivisions. Those distinctions are stated in each target view.
Selected streamlines were subsampled and resampled for browser display. Colors and display density convey identity and visibility, not connection strength, excitation, inhibition, or treatment efficacy. The source papers remain linked; their full texts are not redistributed.