Repository navigation
At a Coordinate
The reverse lookup: you have a point, you want to know what is there. Left panel, At a coordinate.
Four fields:
| Field | Notes |
|---|---|
| AP (mm) | Anterior is positive |
| relative to | bregma · lambda · interaural · occipital crest |
| ML (mm) | Right is positive; matched to whichever hemisphere is closer |
| DV (mm) | Dorsal is positive, so depths are negative |
The panel then lists the nearest printed labels, ranked by 3-D distance, each with how far it sits from your point. Click one to select that structure as usual.
The line under the fields tells you which plate the AP resolves to, and its bregma — click it to go there.

Bregma −5.20, ML 2.60, DV −4.30 resolves to plate 38, where the crosshair is drawn solid. The nearest printed labels are listed with their distances — the closest, ECIC, 0.34 mm away.
Pick on the plate arms the plate; your next click on it fills the fields in — ML and DV from the point you clicked, AP from the plate it was on. Faster than typing when you are aiming at something you can see.
It reads the click, not the hover, so moving the pointer across the image on its way to the button does not disturb anything.
Your point is drawn on the plate as a crosshair, so you can see what you asked for rather than imagining it. It is drawn solid on the plate the AP resolves to and faint on any other — so you can still read your ML and DV against a neighbouring level, without the app pretending the AP matches.
- AP resolves only to the nearest plate — 350 µm. There is nothing between plates to resolve to.
- The results are printed labels, not a segmentation. "Nearest label" is not the same as "inside that structure". A point can sit squarely in a region whose abbreviation happens to be printed further away than a neighbour's.
With a working frame on, the three fields are read in your frame and inverted back to atlas space, and the landmark dropdown steps aside — lambda, interaural and the occipital crest have no ML or DV recorded anywhere, so their position in a tilted frame cannot be worked out honestly.
Cite the atlas: Radtke-Schuller S, Schuller G, Angenstein F, Grosser OS, Goldschmidt J, Budinger E (2016). Brain Struct Funct 221(Suppl 1):1–272. doi:10.1007/s00429-016-1259-0 · Report an issue · Built by Caras Lab, University of Maryland
Start here
Doing things
Reference