RodentFlowRodent behavior, tracked and analyzed on iPhone

Track top-down videos, score assays by hand or import tables, and compare groups with the right statistics. Everything runs on the phone, and nothing is uploaded.

Get started

How tracking works

Source on GitHub

中文版

Video tracking on the phone

Ten assays can be scored straight from a top-down video, from the open field to fear conditioning. The tracker reads the whole recording at 10 samples per second, on the device.

Analyze a video

Calibrated arenas and mazes

Drag a few handles onto the arena corners or the maze arms. Positions are converted to centimeters, and zones, arms and quadrants follow from the calibration.

From positions to metrics

One result per animal

Every video, manual entry or imported row becomes a result for one animal and one session. Zones and thresholds can be changed afterwards without tracking the video again.

Results

Manual entry and table import

Fifteen assays, including marble burying, sucrose preference, beam walking, gait and a two-armed bandit. CSV, TSV and JSON files are accepted, with English or Chinese column names.

Enter or import data

Group statistics that fit the design

The animal is the unit of analysis. The test follows from the design, from Welch's t to mixed ANOVA with Greenhouse–Geisser correction, and each test was checked against SciPy.

Statistics

Private and open source

No account, no analytics and no network connection. Videos, results and figures stay on the iPhone until you share them. The full source code is public under the MIT License.

Source code

Language: English · 简体中文

How it works

A video becomes positions, positions become metrics for one animal, and the metrics of many animals become a group comparison. You can stop at any stage and export what you have.

Assays

Area	Assays	From video
Anxiety & affect	Open field / center, elevated plus maze, light–dark box, marble burying, sucrose preference	Open field, EPM, light–dark box
Learning & memory	Y-maze, T-maze, novel object recognition, Morris water maze, Barnes maze, fear conditioning	All six
Motor function	Beam walking, gait analysis	— (entered or imported)
Reward & decision	Conditioned place preference, two-armed bandit	CPP
Other	Any table of per-animal values	— (imported)

Open source

RodentFlow is open source under the MIT License. The tracker, every metric, the statistics engine and this manual are in one public repository. You can read exactly how each number is computed, build the app yourself, or adapt it to your lab.

github.com/seesaw-earth/RodentFlow

Where to start

Quick startScore your first open-field video in about five minutes.Recording tipsCamera position, lighting and contrast that make tracking reliable.Assay referenceEvery assay, what to set, and how each metric is defined.ValidationHow close the app gets to known ground truth, and where it does not.