Skip to content

STRIDE

STRIDE (Stride-level Tracking and Integrated Decision Extraction) turns one video of a mouse running a T-maze into two phenotypes: what the animal decided and how it walked. A single camera below the transparent maze floor records the rewarded-alternation task; SLEAP tracks a 15-keypoint pose and segments the maze into seven regions; STRIDE scores arm choices, junction exploration and snout probes, and detects every stride to compute the stride-level gait measures.

cognition arm choice, correctness, latencies, junction exploration, snout probes
gait stride and step measures, step width, duration, duty factor, symmetry, body sway and its phase

This is the documentation of STRIDE 1.0.0. Code: talmolab/STRIDE. The study STRIDE was built for, and its results, are on the project page.

Which STRIDE do I need?

For Goes in Comes out Command
STRIDE T-maze T-maze recordings, with your own trained pose and maze-ROI models videos, two model paths per-trial and per-stride CSV tables, a summary, run logs stride tmaze
STRIDE Gait any recording with pose tracking, no maze needed videos + pose files (or a pose model), and the scale (px/cm) per-stride, per-step and per-video gait tables stride gait

Both take the scale (px/cm) from a value you give or from a ChArUco board in view (Scale calibration, stride calibrate).

The authors' own intake workflow

The repository also holds lab/, the workflow the authors use for their own recordings (sorting and renaming camera clips, a per-day scale check, then STRIDE T-maze). It is not needed to use STRIDE; it is described in Authors' lab workflow.

flowchart LR
  V[Videos] --> S[Scale px/cm]
  S --> P[Pose inference]
  P --> R[ROI inference + consensus ROI]
  R --> D[Decisions + decision rules]
  R --> G[Gait: strides, filters]
  P --> G
  D -. optional .-> M[Mixed models, classifier]
  G -. optional .-> M

How STRIDE is designed

  • One command, or one file. Each pipeline runs from the command line alone or from one commented YAML file (stride init-config); a flag on the command line overrides the file (Configuration reference).
  • Every step can be switched on or off, and a second run does not repeat finished work.
  • No extra hardware. Gait comes from the cognitive task's own video.
  • Your files, your names. Any video file names work; metadata is optional.
  • Honest exits. Anything that stops a run exits non-zero with one line saying why; videos that had to be left out are listed with the reason. Warnings never change a value (Logs, warnings and exit codes).
  • Reproducible. Every stage writes a run manifest recording the code version, models, settings, inputs and outputs.
  • Documented as computed. The column glossary describes every number exactly as the code computes it.

Citation

Li Y, Ling L, Chou R, Ketterer-Sykes EB, Andrews NA, Han C, Liao C, Lee CT, Lee KF*, Pereira TD*. Dual cognitive and motor phenotyping for early detection of Alzheimer's Disease progression in mouse models. In preparation. * corresponding authors.

Software: talmolab/STRIDE 1.0.0, MIT license. A software DOI will be added when the release is archived. BibTeX: Citation.