Dataset DOI: 10.5061/dryad.7m0cfxq3x

Description of the data and file structure

Talitrid amphipods are an extensively studied system for navigation due to their robust ability to navigate back to the optimal burrowing zone after foraging and could be a model system in which to study the impacts of collective behavior on short-distance navigation and orientation. We investigated whether talitrid amphipods (Megalorchestia pugettensis) differ in their orientation abilities when released individually versus in a group. When released individually, the amphipods took longer to start moving (p<0.001), traveled longer paths (p=0.003), moved faster (p=0.016), had a different initial bearing (p=0.003), and exhibited more spread in their initial bearing (p=0.009) than when released in groups. There was no difference between individuals and groups in terms of their trial time, nor in the direction or spread of their final orientation. This study introduces a tractable, invertebrate species in which to study the impacts of collective movement and reveals previously unexamined differences in orientation abilities for talitrid amphipods released independently vs. in a group that have implications for experimental design in this system.

On Zenodo: The Markdown document "Main Code.rmd" along with the associated knit "Main-Code.html" contains all the code needed to recreate the analyses in the main text and supplemental of the associated paper. In addition to the code to make the figures and tables, it also contains the following:

• Loading of all of the trial data that was created in the file "Code to Load and Wrangle Data.rmd" (described below), along with detailed information about each column
• Loading of the data from the non-parametric permutation test (i.e. "bootstrapped data"). This is done to reduce the time associated with re-running the bootstrap. These files are generated in the markdown document "Nonparametric Bootstrap Analysis.rmd". The files in question are entitled "Boostrapped Underlying Dist for GS5.csv", "Boostrapped Underlying Dist for GS10.csv" and "Averaged GS1 Values.csv"
• Sourcing of the file "Circular Plots Function.R" which contains a function to create the circular plots in figure 2A
• Sourcing of the files "Violin Plots with Bootstrap pvalue.R" and "Circular Violin Plots with Bootstrap pvalue.R" which contains functions to create the violin plots (for linear and circular variables respectively) in figure 3
• Sourcing of the files "Bootstrap Histogram Plot Function.R" and "Circular Bootstrap Histogram Plot Function.R" which contains functions to create the histogram plots (for linear and circular variables) in figure S6 and S7
• Sourcing of the file "Mixed Effect Model Function.R" which contains a function to run the mixed effect models associated with table S2 and pull out the relevant variables of interest
• Sourcing of the file "Violin Plots of Weather.R" which contains functions to create the violin plots in figure S5

The original trial videos were digitized using the MATLAB app DLTdv8a. The raw output is included for each group size in the zipped folders (including both the .csv output and the .mat files) entitled "GS#_Raw DLTdv8a output.zip" where the # symbol corresponds to each of the associated GS's. The original trial videos used for each .mat are included in the zipped folders "GS# Videos.zip". The videos for GS 1 and GS10 are included, but the videos for GS5 could not be included due to exceeding the data storage limit, and due to the fact that they were not used in the final publication. These videos are available upon request to the lead author.

These raw csv files output from DLTdv8a were then compiled and tidied to use for further analysis using the code in the markdown document entitled "Code to Load and Wrangle Data.rmd". This markdown document also sources the following scripts which contain the functions needed to tidy the outputs from DLTdv8a and calculate the variables of interest: "Functions for GS1_Cleaned for Pub_V2.R", "Functions for GS5_Cleaned for Pub_V2.R", "Functions for GS10_Cleaned for Pub_V2.R". The output from this code is the two files: GSall.csv which contains the metadata and track summary statistics for each animal, and GSalltracks.csv which contains the raw path information for each animal.  More details about each of the columns of these files can be found in the beginning of the markdown document "Main Code.rmd".

Code/Software

All statistical analyses were performed using R (v. 4.2.3., R Core Team 2023)

Files and variables

File: GS1_Videos.zip

Description: Zip folders containing the raw experimental videos for the Group-Size 1 trials. File names are as follows: MMDDYY-GS1-T# for the Month, Day, and Year. The GS stands for group size and the number after designates how many individuals were in the trial. The T stands for trial, and the number after designates the trial number. Any additional parenthetical numbers included in the title are because the videos were broken into parts, and designates the order in which they should appear.

File: GS10_Videos.zip

Description: Zip folders containing the raw experimental videos for the Group-Size 10 trials. File names are as follows: MMDDYY-GS10-T# for the Month, Day, and Year. The GS stands for group size and the number after designates how many individuals were in the trial. The T stands for trial, and the number after designates the trial number. Any additional parenthetical numbers included in the title are because the videos were broken into parts, and designates the order in which they should appear.

File: GS1_Raw_DLTdv8a_output.zip

Description: Zip folders containing both the .csv and .mat output from the digitization of the trial videos for Group-Size 1. For both file types, the number immediately following "GS" in the file name indicates the number of individuals used in the trial, and the number immediately following "T" indicates the trial number. Other letters, or numbers separated by a - or _ have no significant meaning. The .mat files are generated by the MATLAB app DLTdv8a, and can only be opened and read by that software. They contain the digitization information (i.e. X and Y location for each frame, for each tracked individual), and more can be learned about this software and file formatting from https://biomech.web.unc.edu/dltdv/.

There is one trial with two .mats: Trial 1, which is associated with both GS1_T1_dvProject.mat (which was used to generate the .csv used in the analysis and is consistent with the rest of the files in the project), and 073021_GS1_T1_1_dvProject.mat, which was the first file ever digitized and the entire video was digitized, instead of only the first few minutes. This file was not used, but is included for completeness.

Each .csv contains the following, where each row corresponds to one frame of the video file:

• Latency: This column contains the word "Cup" in the row where the cup was lifted, and the word "Start" in the row where the animal began moving
• Prewall_X: The X position of the animal before it hit the wall of the arena
• Prewall_Y: The Y position of the animal before it hit the wall of the arena
• Postwall_X: The X position of the animal after it hit the wall of the arena
• Postwall_Y: The Y position of the animal after it hit the wall of the arena
• Beach_X and Beach_Y: The X and Y location of the beach. The first row is the center of the arena, the second row is the point along the edge of the arena in the direction of the beach
• North_X and North_Y: These variables are not used, but indicate the direction of True North in a manner similar to Beach
• Center_X and Center_Y: The X and Y location of the center of the arena
• Date: The Date (first row only)
• GS: The number of individuals in the trial, or group size (first row only)
• Trial: The trial number (first row only)
• Weather: A character string description of the weather (first row only)
• Temp: The temperature (first row only)
• TOD: The time of day that the trial began (first row only)
• Dig: The person who digitized the trial (first row only)

File: GS5_Raw_DLTdv8a_output.zip

Description: Zip folders containing both the .csv and .mat output from the digitization of the trial videos for Group-Size 5. These files are identical to those in GS1_Raw_DLTdv8a_output.zip except that the csv's contain the following additional columns:

• Cup: This column has the word "Cup" in the row where the cup was lifted
• Start1 - Start5: These columns contain the words "Start" in the row where the animal began moving. They also contain the word "Stop" in the rows where the animal hit the wall of the arena
• Prewall1_X, Prewall1_Y - Prewall5_X, Prewall5_Y: The X and Y positions for all animals before hitting the wall. The # following "Prewall" indicates the animal's numerical ID

File: GS10_Raw_DLTdv8a_output.zip

Description: Zip folders containing both the .csv and .mat output from the digitization of the trial videos for Group-Size 10. These files are identical to those in GS1_Raw_DLTdv8a_output.zip except that the csv's contain the following additional columns:

• Prewall1_X, Prewall1_Y - Prewall10_X, Prewall10_Y: The X and Y positions for all animals before hitting the wall. The # following "Prewall" indicates the animal's numerical ID
• Moves: This column indicates the time at which the animal started moving, where the row number corresponds to the individual in question (i.e. row 1 is the move time for Animal 1)
• WallHit: The time at which the animal hit the wall of the arena, where the row number corresponds to the individual in question (i.e. row 1 is the wall hit for Animal 1)
• Cup: The time the cup was lifted

File: GSall.csv

Description: The raw csv files output from DLTdv8a were compiled and tidied to use for further analysis using the code in the markdown document entitled "Code to Load and Wrangle Data.rmd". This markdown document also sources the following scripts which contain the functions needed to tidy the outputs from DLTdv8a and calculate the variables of interest: "Functions for GS1_Cleaned for Pub_V2.R", "Functions for GS5_Cleaned for Pub_V2.R", "Functions for GS10_Cleaned for Pub_V2.R". The output from this code is the two files: GSall.csv which contains the metadata and track summary statistics for each animal, and GSalltracks.csv which contains the raw path information for each animal.  More details about each of the columns of these files can be found in the beginning of the markdown document "Main Code.rmd".

Variables

• GS: The group size
• Trial: The trial number (the NA for GS 1 is Trial 14)
• Center_X: The x coordinates of the center of the pan
• Center_Y: The y coordinates of the center of the pan
• Beach_X: The x coordinates of the spot on the edge of the pan closest to the beach/wrackline
• Beach_Y: The y coordinates of the spot on the edge of the pan closest to the beach/wrackline
• AngleBeach: The angle between the center of the pan and the beach
• Date: Date of the trial in Year-Month-Day.
• Weather: A qualitative descriptor of the weather
• Temp: Temperature at the start of the trial in Fahrenheit
• hour: The hour of the day when the trial began in 24hr
• minutes: The minute of the day when the trial began in 24hr
• Dig: The name of the person who digitized the trial
• Dist_Actual: The sum of the distance traveled between each step (cm)
• Dist_StrtLine: The beeline distance between their starting and ending coordinates (cm)
• Sinuosity: Dist_StrtLine/Dist/Actual. Not an accurate measurement of path wiggliness b/c of non-standard steplengths
• Time: The total number of seconds between when the animal started moving to when they finished the trial
• FinalPosition: The angle between where the animal ended the trial and the center of the arena relative to the wrackline. Renamed to Final Orientation
• HeadingMean: The circular mean of each heading, weighted by the distance traveled relative to the wrackline
• StartHeading: The bearing of their first step, relative to the wrackline. Renamed to InitialBearing
• Velocity: Dist_Actual / Time in cm per s
• Start_X: Their starting X location in the original coordinate system
• Start_Y: Their starting Y location in the original coordinate system
• Start_X.T: Their starting X location transformed to a new coordinate system where the center of the pan is at 0,0, and the angle between the center of the pan and the beach is = 0 (rotated such that the beach is always to the right of all figures)
• Start_Y.T: Their starting Y location transformed to a new coordinate system where the center of the pan is at 0,0, and the angle between the center of the pan and the beach is = 0 (rotated such that the beach is always to the right of all figures)
• StartLocation: Their angle between where the animal started the trial and the center of the arena relative to the wrackline. Not equivalent to StartHeading which is the angle between their first and second step. Renamed to InitialOrientation
• End_X: Their final X location in the original coordinate system
• End_Y: Their final Y location in the original coordinate system
• Latency: The time between when the trial started (the cup was lifted) and when the animal began moving in s. This time is not included in their total time
• Animal: a designator for which animal was being digitized within each trial
• Dist_Diff: Dist_Actual-Dist_StrtLine. Renamed to ExcessDist
• TOD: Time of Day at the beginning of the trial
• Smoky: A designator for the days during which there was a nearby forest fire
• Year: The year the trial was conducted
• Latency2: log(Latency), used for fig S5
• GS.Trial: GS and Trial combined. Used for MEM's
• FinalPosition.Pan: The angle between where the animal ended the trial and the center of the arena relative to the arena. Renamed to FinalOrientation_pan
• FinalPosition.L: A linearized version of the Final position. Renamed to FinalOrientation.L
• StartHeading.L: A linearized version of the Start Heading. Renamed to InitialBearing.L

File: GSalltracks.csv

Description: The raw csv files output from DLTdv8a were compiled and tidied to use for further analysis using the code in the markdown document entitled "Code to Load and Wrangle Data.rmd". This markdown document also sources the following scripts which contain the functions needed to tidy the outputs from DLTdv8a and calculate the variables of interest: "Functions for GS1_Cleaned for Pub_V2.R", "Functions for GS5_Cleaned for Pub_V2.R", "Functions for GS10_Cleaned for Pub_V2.R". The output from this code is the two files: GSall.csv which contains the metadata and track summary statistics for each animal, and GSalltracks.csv which contains the raw path information for each animal.  More details about each of the columns of these files can be found in the beginning of the markdown document "Main Code.rmd".

Variables

• X: x of each step in the original coordinate system from the moment the animal starts moving till the trial ends
• Y: y of each step in the original coordinate system from the moment the animal starts moving till the trial ends
• X.Centered: x of each step rel. to CENTER OF ARENA (in this coordinate system, center=0,0 but the beach is still AngleBeach)
• Y.Centered: y of each step rel. to CENTER OF ARENA (in this coordinate system, center=0,0 but the beach is still AngleBeach)
• Dist_Center: distance between each step to CENTER OF ARENA (calculated after centering each x,y)
• AngleCenter: angle between each step and CENTER OF ARENA (calculated after centering each x,y)
• MinPrev_X: x of each step rel. to PREVIOUS STEP (calculated on the original x,y)
• MinPrev_Y: y of each step rel. to PREVIOUS STEP (calculated on the original x,y)
• Dist_Each: distance between EACH STEP (calculated on the original x,y)
• AngleCenter_Transformed: angle between each step and CENTER OF ARENA, transformed such that 0=beach (calculated as AngleCenter - AngleBeach)
• X.Transformed: new x locations transformed (in this coordinate system, center=0,0 and the angle towards the beach is = 0. Calculated as Dist_Center*cos or sin of AngleCenter_Transformed)
• Y.Transformed: new y locations transformed (in this coordinate system, center=0,0 and the angle towards the beach is = 0. Calculated as Dist_Center*cos or sin of AngleCenter_Transformed)
• Heading: the individuals actual bearing, relative to the beach. 0 = towards the beach (calculated using X/Y.Transformed)
• Frame: frame of that particular step
• Trial: Trial number
• Animal: Animal ID
• GS: Group Size
• All other variables are imported from GSall.csv and their description and explanation can be found under that file.

File: Bootstrapped_Underlying_Dist_for_GS5.csv

Description: The data from the non-parametric permutation test (i.e. "bootstrapped data"). This is done to reduce the time associated with re-running the bootstrap. These files are generated in the markdown document "Nonparametric Bootstrap Analysis.rmd". This is the bootstrapped data for group size 5

Variables

• Velocity: The mean of the bootstrapped version of Dist_Actual / Time in cm per s
• Sinuosity: The mean of the bootstrapped version of Dist_StrtLine/Dist/Actual.
• HeadingMean: The mean of the bootstrapped version of the circular mean of each heading, weighted by the distance traveled relative to the wrackline
• Latency: The mean of the bootstrapped version of the time between when the trial started (the cup was lifted) and when the animal began moving in s.
• rho_MeanHeading: The rho.circular of the circular mean of each heading, weighted by the distance traveled relative to the wrackline
• rho_FinalPosition: The rho.circular of the angle between where the animal ended the trial and the center of the arena relative to the wrackline.
• rho_StartHeading: The rho.circular of the bearing of their first step, relative to the wrackline
• Time: The mean of the bootstrapped version of the total number of seconds between when the animal started moving to when they finished the trial
• StartHeading: The circular mean of the bootstrapped version of the bearing of their first step, relative to the wrackline
• FinalPosition: The circular mean of the bootstrapped version of the angle between where the animal ended the trial and the center of the arena relative to the wrackline.
• StartPosition: The circular mean of the bootstrapped version of their angle between where the animal started the trial and the center of the arena relative to the wrackline
• rho_StartPosition: The rho.circular of their angle between where the animal started the trial and the center of the arena relative to the wrackline
• Dist_Actual: The mean of the bootstrapped version of the sum of the distance traveled between each step (cm)
• Dist_StrtLine: The mean of the bootstrapped version of the beeline distance between their starting and ending coordinates (cm)
• Dist_Diff: The mean of the bootstrapped version of Dist_Actual-Dist_StrtLine

File: Bootstrapped_Underlying_Dist_for_GS10.csv

Description: The data from the non-parametric permutation test (i.e. "bootstrapped data"). This is done to reduce the time associated with re-running the bootstrap. These files are generated in the markdown document "Nonparametric Bootstrap Analysis.rmd". This is the bootstrapped data for group size 10

Variables

These are the same as described for the file Bootstrapped_Underlying_Dist_for_GS5.csv

File: Averaged_GS1_Values.csv

Description: The data from the non-parametric permutation test (i.e. "bootstrapped data"). This is done to reduce the time associated with re-running the bootstrap. These files are generated in the markdown document "Nonparametric Bootstrap Analysis.rmd". This is the averaged data for group size 1

Variables

These are the same as described for the file Bootstrapped_Underlying_Dist_for_GS10.csv but instead of being the mean/rho over a bootstrapped version of the data, it is just the mean or rho over all GS 1 trials, with no bootstrapping performed.