movement

Measuring behaviour as movement

Defining behaviour is tricky, but many have tried.

The total movements made by the intact animal (Tinbergen 1951).

We can quantify movements through various tools:

🎥 Video cameras
📱 Inertial measurement units (IMUs)
🛰️ GPS-based biologgers

From videos to motion

“CRISPR Ants Lose Ability to Smell” (2017)

Markerless pose estimation

Animal behaviour 🤝 computer vision

What happens after tracking?

Lack of standardised data formats and tools
Lots of fragile ‘in-house’ scripts
Piles of un-analysed data

movement: overview

The movement dataset

Dataset types

movement input/output

Cleaning data

Quick plots

Quantifying motion

Scripting with movement

from movement.io import load_dataset
from movement.filtering import rolling_filter
from movement.kinematics import compute_speed

ds = load_dataset(
  "path/to/my_data.h5", source_software="DeepLabCut", fps=30
)

ds = ds.sel(time=slice(600, 2400))

ds["position_smooth"] = rolling_filter(
  ds["position"], window=5, statistic="median"
)

ds["speed"] = compute_speed(ds["position_smooth"])

ds.to_netcdf("my_data_processed.nc")

movement GUI

Drawing regions of interest

Using regions of interest

Load into as polygons or lines

Compute distances and angles relative to regions

Compute region occupancy

Keeping it light

Package available on PyPI and conda-forge
| | Works across OSes
No dedicated GPU required

Long-term vision

movement as the scikit-image for animal motion data.

If you prefer the R ecosystem, check out the animovement toolbox by Mikkel Roald-Arbøl.

Behaviour segmentation?

Aligning data streams

movement community

~90k downloads | 42 code contributors | ~370 merged pull requests

Join the movement

movement.neuroinformatics.dev

github.com/neuroinformatics-unit/movement

neuroinformatics.zulipchat.com > movement

neuroinformatics.dev | Neuroinformatics Unit

nikosirmpilatze.com | @niksirbi@neuromatch.social

Animals in Motion

References

“CRISPR Ants Lose Ability to Smell.” 2017. Nature 548 (7667): 263–63. https://doi.org/10.1038/d41586-017-02337-4.

Mathis, Alexander, Pranav Mamidanna, Kevin M. Cury, et al. 2018. “DeepLabCut: Markerless Pose Estimation of User-Defined Body Parts with Deep Learning.” Nature Neuroscience 21 (9): 1281–89. https://doi.org/10.1038/s41593-018-0209-y.

Pereira, Talmo D., Nathaniel Tabris, Arie Matsliah, et al. 2022. “SLEAP: A Deep Learning System for Multi-Animal Pose Tracking.” Nature Methods 19 (4): 486–95. https://doi.org/10.1038/s41592-022-01426-1.

Tinbergen, Niko. 1951. The Study of Instinct. Clarendon Press.