Anisotropic interaction and motion states of locusts in a hopper band
Cite this dataset
Weinburd, Jasper; Simpson, Stephen; Sword, Gregory; Buhl, Camille (2024). Anisotropic interaction and motion states of locusts in a hopper band [Dataset]. Dryad. https://doi.org/10.5061/dryad.n02v6wwzz
Abstract
Swarming locusts present a quintessential example of animal collective motion. Juvenile locusts march and hop across the ground in coordinated groups called hopper bands. Composed of up to millions of insects, hopper bands exhibit coordinated motion and various collective structures. These groups are well-documented in the field, but the individual insects themselves are typically studied in much smaller groups in laboratory experiments. We present the first trajectory data that detail the movement of individual locusts within a hopper band in a natural setting. Using automated video tracking, we derive our data from footage of four distinct hopper bands of the Australian plague locust, Chortoicetes terminifera. We reconstruct nearly twenty-thousand individual trajectories composed of over 3.3 million locust positions. We classify these data into three motion states: stationary, walking, and hopping. Distributions of relative neighbor positions reveal anisotropies that depend on motion state. Stationary locusts have high-density areas distributed around them apparently at random. Walking locusts have a low-density area in front of them. Hopping locusts have low-density areas in front and behind them. Our results suggest novel interactions, namely that locusts change their motion to avoid colliding with neighbors in front of them.
README: Trajectory data for locusts in a hopper band
https://doi.org/10.5061/dryad.n02v6wwzz
This repository contains data used in creating the work:
- J Weinburd, J Landsberg, A Kravtsova, S Lam, T Sharma, SJ Simpson, GA Sword, C Buhl. Anisotropic interaction and motion state of locusts in a hopper band. Proc. R. Soc. B. **291: **20232121. https://doi.org/10.1098/rspb.2023.2121
The contents include numerical position-time data extracted from raw video footage of locust hoppers in the field. The video was recorded by Drs. Stephen J Simpson, Greg A Sword, and Camille Buhl and provided as part of an ongoing collaboration. Students who have contributed to this project include: Anna Kravtsova (Eastern Washington U), Shanni Lam (HMC), Jacob Landsberg (Haverford), and Tarush Sharma (HMC).
Give a brief summary of dataset contents, contextualized in experimental procedures and results.
Description of the data and file structure
The main data set is contained in two Matlab data files .mat which include cleaned and processed numerical trajectories and two subdirectories.
- data_recording.mat, data_recording2.mat
- contains the the full data set used in the manuscript linked above.
examples contains all data files for the two example clips analyzed in pipeline_examples.m (see the link below to code hosted on GitHub). Includes the starting preprocessed video, intermediate processed video and .xml data files, and final data data_recording_examples.mat.
manual contains the two ground-truth data sets obtained by manual tracking in TrackMate. Students Jacob Landsberg and Tarush Sharma did the bulk of the manual tracking. The original video files are also included so that the .xmls can be opened in Fiji>Plugins>Tracking>Load a TrackMate File.
The data variables themselves use Matlab's struct data format and can be somewhat opaque. Using a script in the code linked below (struct2data.m) one can convert the data variables into more intuitive 3D arrays where each row is a locust, each column is a feature, and the third dimension corresponds to time.
Code/Software
The code used to extract, clean, and process the trajectory data, to conduct the analysis, and to generate the figures and tables of the manuscript is available on the public GitHub repository:
Methods
See Methods in the associated manuscript: https://doi.org/10.1101/2021.10.29.466390
See also the README file in the associated GitHub repository: https://github.com/weinburd/locust_trajectory_data
Funding
National Science Foundation, Award: DMS–1902818
Australian Research Council, Award: LP150100479
Australian Research Council, Award: FT110100082
National Science Foundation, Award: DMS–1757952