This repository contains Matlab code for running the analyses described in the manuscript "Spatiotemporal visual statistics of aquatic environments in the natural habitats of zebrafish" By Cai et al.
The dataset of terrestrial and aquatic videos should be downloaded separately from:
https://doi.org/10.5281/zenodo.7502451
and the path should be manually updated in the code.
The repository is split into three parts:
This code runs the analyses of global and local contrast statistics (Figures 2, 3, and 5B/C). The script "main" sets the appropriate paths and calls a set of scripts to run each analysis, save plots, and report statistical tests.
This code runs the analyses of spatiotemporal statistics (Figures 4 and 5A). The script "main" sets the appropriate paths and calls a set of scripts to run each analysis, save plots, and report statistical tests.
This code demonstrates the preprocessing steps to go from raw output from the Insta 360 cameras to the sampled videos used in the analyses. See the readme file within this directory for more information
The dataset you download from Zenodo contains all of the terrestrial (air_g_brc) and aquatic (water_g_brc) videos (sometimes called "cubes" in the code) that were subsampled from the original field recordings, as well as mat files (air_g_brc_valid_cubes.mat, water_g_brc_valid_cubes.mat) with the names of the files that were used for the final analysis after applying exclusion criteria. There is also a zip archive called calibration_data that includes the camera calibration files that are used in the preprocessing step to generate the video samples.