Data for: Collective signalling is shaped by feedbacks between signaller variation, receiver perception, and acoustic environment in a simulated communication network
Cite this dataset
Reichert, Michael; Luttbeg, Barney; Hobson, Elizabeth (2023). Data for: Collective signalling is shaped by feedbacks between signaller variation, receiver perception, and acoustic environment in a simulated communication network [Dataset]. Dryad. https://doi.org/10.5061/dryad.n8pk0p338
Abstract
Communication takes place within a network of multiple signallers and receivers. Social network analysis provides tools to quantify how an individual’s social positioning affects group dynamics, and the subsequent biological consequences. However, network analysis is rarely applied to animal communication, likely due to the logistical difficulties of monitoring natural communication networks. We generated a simulated communication network to investigate how variation in individual communication behaviours generates network effects, and how this communication network’s structure feeds back to affect future signalling interactions. We simulated competitive acoustic signalling interactions among chorusing individuals and varied several parameters related to communication and chorus size to examine their effects on calling output and social connections. Larger choruses had higher noise levels, and this reduced network density and altered the relationships between individual traits and communication network position. Hearing sensitivity interacted with chorus size to affect both individuals’ positions in the network and the acoustic output of the chorus. Physical proximity to competitors influenced signalling, but a distinctive communication network structure emerged when signal active space was limited. Our model raises novel predictions about communication networks that could be tested experimentally, and identifies aspects of information processing in complex environments that remain to be investigated.
README: Data for: Collective signalling is shaped by feedbacks between signaller variation, receiver perception, and acoustic environment in a simulated communication network
https://doi.org/10.5061/dryad.n8pk0p338
By Michael Reichert - 12/05/2023. Contact michael.reichert@okstate.edu for questions
Data and code used for simulations and analyses.
There are 7 files (in addition to this readme file). Data files uploaded to Dryad. Code files available on Zenodo.
1. Simulation_Code.R
This is the R code that was used to run the simulations. Includes the actual simulation function plus some lines to manipulate and save the output. Note that the function is slow and is set up for parallel processing on a supercomputer, so use with caution.
2. IndividualData.csv
This is output showing characteristics of each simulated individual in each simulation run. This is a large dataset, in total there were 2.2 million simulated individuals across the simulations. Details on the variables contained in this file below. These variables include both attributes assigned to the individual prior to the simulation as well as results of the simulation.
3. NetworkData.csv
This is output containing summary statistics of the communication networks as well as the network parameters for each simulation run (40000 runs in total). Details on variables below.
4. Figure2ANetwork.Rdata
An exemplar communication network. This is an igraph object so needs to be loaded after igraph package is loaded in R. This network was from a simulation run with the parameters: Chorus size = 10, no selective attention, hearing sensitivity = - 12 dB, no noise responsiveness.
5. Figure2BNetwork.Rdata
Another exemplar communication network. This is an igraph object so needs to be loaded after igraph package is loaded in R. This network was from a simulation run with the parameters: Chorus size = 100, one competitor attended to, hearing sensitivity = 6 dB, no noise responsiveness.
6. NoiseLevelsInChorus.csv
Dataset containing the sound pressure level in the center of the chorus for each simulation run. Details on variables below.
7. Figure Preparation Script.Rmd
The R script used to prepare the figures. This loads up the data files listed above and processes them to produce the figures that are in both the main text and the supplement. This is an R markdown file, so the user may need to configure markdown to get this to work, and note that this is assuming the user is working in a Project in RStudio with all data files saved in the project folder. (see here for details on how to do that: <https://support.posit.co/hc/en-us/articles/200526207-Using-RStudio-Projects>)
Note that none of these files are the "raw" raw data. Raw output from the simulation runs include every single call given by every single individual. The total size of all of these files is 72.5 GB. We are unaware of any existing repository that can handle such a large file size. All of the files above contain summary statistics calculated from the raw data files. If any raw files are truly needed, please contact the corresponding author.
Details for each data file.
2. IndividualData.csv
-Each row represents an individual simulated in one simulation run
Columns:
FrogID - The individual's ID code within that simulation (note that codes were reused across simulations)
XCor - The x-coordinate location of that individual on the chorus arena, in m (minimum value = 0, maximum value = 50).
YCor - The y-coordinate location of that individual on the chorus arena, in m (minimum value = 0, maximum value = 50).
BaselineCD - That individual's baseline call duration, in s.
Amplitude - That individual's call amplitude, in dB.
SNR - That individual's hearing sensitivity, in dB.
InStrength - That individual's in-strength, based on the simulated communication network. Note that for analyses, we used the NormalizedInStrength, see below.
OutStrength - That individual's in-strength, based on the simulated communication network. Note that for analyses, we used the NormalizedOutStrength, see below.
EigenCentrality - That individual's eigenvector centrality in the simulated communication network (scaled to a maximum value of one).
MeanObservedCD - That individual's mean call duration, averaged across all calls it gave during the simulation.
NumberFrogs - Chorus size, number of individuals in the simulation.
NumberAttended - Number of individuals attended to. Note that when there was no selective attention, this equalled the size of the chorus.
MeanSignalNoiseRatio - The mean hearing sensitivity for that simulation run (-12, -6, 0, 6 dB). Calls had to be at least as loud as the background noise amplitude plus the hearing sensitivity to be heard and potentially influential on the receiver (therefore more negative values indicate more sensitive hearing).
DistNetworkCentrality - That individual's eigenvector centrality in the spatial network (scaled to a maximum value of one).
DeltaCentrality - The difference between the focal individual's eigenvector centrality in the communication network and its eigenvector centrality in the spatial network. Note that for graphs we multiplied this value by -1 in the graphing script to convert this as the centrality in the spatial network minus the centrality in the communication network (because it's more intuitive to interpret the graphs that way).
NormalizedInStrength - The focal individual's in-strength, normalized by dividing edge weights by the average edge weight of the network, to allow for comparison across different simulated network sizes.
NormalizedOutStrength - The focal individual's out-strength, normalized by dividing edge weights by the average edge weight of the network, to allow for comparison across different simulated network sizes.
NoiseEffect - Whether or not the individual was responsive to noise (i.e. adjusted its calling behavior to the level of the background noise, regardless of whether any actual calls were heard).
3. NetworkData.csv
-Each row represents a simulation run.
Columns:
NumberFrogs - Chorus size, number of individuals in the simulation.
NumberAttended - Number of individuals attended to. Note that when there was no selective attention, this equalled the size of the chorus.
MeanSignalNoiseRatio - The mean hearing sensitivity for that simulation run (-12, -6, 0, 6 dB). Calls had to be at least as loud as the background noise amplitude plus the hearing sensitivity to be heard and potentially influential on the receiver (therefore more negative values indicate more sensitive hearing).
NoiseEffect - Whether or not the individual was responsive to noise (i.e. adjusted its calling behavior to the level of the background noise, regardless of whether any actual calls were heard).
SelectiveAttention - If TRUE, individuals attended to only a subset of the most influential competitors who called in the interval since the focal individual's previous call, where the number of competitors attended to was given by the NumberAttended variable.
EdgeDensity - The mean edge density in the network (the number of edges divided by all possible edges).
MeanPathLength - the average of the shortest paths between all pairs of individuals in the chorus. Note that we used the NormalizedPathLength variable in analyses, not this one. See below.
GrandMeanCallDuration - The average call duration across all calls given during the simulation run, in s.
NormalizedPathLength - The path length calculated from a network where edge weights were normalized by the average edge weight in the network.
NumberAttendedCat - The number of individuals attended to, but for simulations where there was no selective attention, these were given a value of "5" indicating that all competitors were attended to while accounting for the fact that different simulation runs had different maximum possible individuals attended to due to differences in chorus size.
UniqueParameters - A helper variable for summary statistics, this is the chorus size, hearing sensitivity, number attended and noise responsiveness pasted together.
6. NoiseLevelsInChorus.csv
-Each row represents a simulation run.
ChorusSize - Chorus size, number of individuals in the simulation.
NumberFrogsAttended - Number of individuals attended to. Note that when there was no selective attention, this equalled the size of the chorus.
SNR - The mean hearing sensitivity for that simulation run (-12, -6, 0, 6 dB). Calls had to be at least as loud as the background noise amplitude plus the hearing sensitivity to be heard and potentially influential on the receiver (therefore more negative values indicate more sensitive hearing).
NoiseResponsiveness - Whether or not the individual was responsive to noise (i.e. adjusted its calling behavior to the level of the background noise, regardless of whether any actual calls were heard).
NoiseSPL - The noise amplitude, in dB SPL, calculated from the centre of the chorus by sampling the noise level every 10 s and taking the root-mean-square of these values across the entire duration of the chorus.
Funding