Start

Five short pages, about 30 minutes. By the end you can load, inspect, analyze, and save a dataset end-to-end.

This is the practical on-ramp. Each page does one real task and shows the result, then explains — briefly — why the API is shaped the way it is.

#	Page	What you’ll do
01	Read an image	Load a NIfTI file and get your data, with its spatial frame attached.
02	Spatial frames	Check that two images line up before you combine them.
03	ROI extraction	Pull a region’s time-by-voxel matrix and keep track of where it came from.
04	Derived maps	Compute a map (e.g. temporal SNR) that remembers how it was made.
05	Round-trip	Save it and read it back — provenance still intact, no sidecar files.

Each page ends with a short, skippable What it refuses note: the few inputs neuroim rejects on purpose so a mistake fails loudly instead of producing wrong numbers. Read it if you want the edges; skip it on a first pass.

Prefer vocabulary first? The Concepts pages are short and indexed.