Evidence for individual vocal recognition in a pair-bonding poison frog, Ranitomeya imitator
Data files
Feb 07, 2024 version files 130.40 KB
-
No_Choice_Scoring_Data.csv
-
No_Choice_Trial_Info.csv
-
README.md
-
Two_Choice_IDs.csv
-
Two_Choice_Scoring_Data.csv
-
Two_Choice_Trial_Info.csv
Abstract
Individually distinctive vocalizations are widespread in nature, although the ability of receivers to discriminate these signals has only been explored through limited taxonomic and social lenses. Here, we asked whether anuran advertisement calls, typically studied for their role in territory defense and mate attraction, facilitate recognition and preferential association with partners in a pair-bonding poison frog (Ranitomeya imitator). Combining no- and two-stimulus choice playback experiments, we evaluated behavioral responses of females to male acoustic stimuli. Virgin females oriented to and approached speakers broadcasting male calls independent of caller identity, implying that females are generally attracted to male acoustic stimuli outside the context of a pair bond. When pair-bonded females were presented with calls of a mate and a stranger, they showed significant preference for calls of their mate. Moreover, behavioral responses varied with breeding status: females with eggs were faster to approach stimuli than females that were pair-bonded but did not currently have eggs. Our study suggests a potential role for individual vocal recognition in the formation and maintenance of pair bonds in a poison frog and raises new questions about how acoustic signals are perceived in the context of monogamy and biparental care.
README: Evidence for individual vocal recognition in a pair-bonding poison frog, Ranitomeya imitator
---
This dataset contains all raw video scoring tabulations and relevant trial information used in the quantification and comparison of phonotaxis behavior in virgin and pair-bonded Ranitomeya imitator, contextualized in experimental procedures and results. Five files (described in detail below) collectively serve as inputs for the script, Phonotaxis_Final.R, which performs the statistical analyses.
Description of the data and file structure
The file No Choice Trial Info.csv contains all information pertaining to the 40 no-choice trials conducted on virgin female R. imitator. Variables are as follows:
- Trial Number: A numeric designation given to the trial at the time of completion, in the order of completion.
- Female ID: ID of the female subject ('Ri.####')
- Focal Female: Coded experimental ID given to each of 9 subjects ('j1-9')
- Male ID: ID of male from which the stimulus call was recorded
- Side: Side of arena from which stimulus was projected (L or R, randomized for each trial)
- Date: Date on which trial was conducted
- Time: Time at which trial was conducted
- Acclimation Start: Time at which subject was placed under plastic cup to begin acclimation period
- Acclimation End: Time at which the plastic cup was lifted, marking the end of the acclimation period (five minutes) and the start of the trial.
- Trial Start: Time at which the trial began
- Trial End: Time at which the trial ended (ten minutes later)
The file Two Choice Trial Info.csv contains all information pertaining to the 28 two-choice trials conducted on pair-bonded female R. imitator. Variables are as follows:
- Focal Male: Coded experimental (letter) ID given to each of the pair-bonded male partners from which the stimulus calls were recorded.
- Focal Female: Coded experimental (letter) ID given to each of 9 female subjects. Each female letter designation mirrors the letter of her male partner, but in lower-case to distinguish.
- Trial Number: A numeric designation given to the trial at the time of completion, in the order of completion. Note, the original study design included additional trial types which were excluded for the purpose of this analysis, leaving gaps in the numbering.
- Block: Time block (1 or 2) during which the trial was conducted. Blocks are spaced nine weeks apart and facilitate repeated measures of some individuals.
- Trial Type: 'Mate vs Stranger' or 'Stranger vs Mate' - two trial types signifying whether the speaker projecting the mate stimulus was on the left or right side, respectively.
- Left Speaker: Letter ID of male stimulus presented from the left side speaker
- Right Speaker: Letter ID of male stimulus presented from the right side speaker.
- Date: Date on which trial was conducted
- Time: Time at which trial was conducted
- Acclimation Start: Time at which subject was placed under plastic cup to begin acclimation period.
- Acclimation End: Time at which the plastic cup was lifted, marking the end of the acclimation period (five minutes) and the start of the trial.
- Trial Start: Time at which the trial began
- Trial End: Time at which the trial ended (ten minutes later)
- Humidity (%): Level of humidity measured within the experimental arena, when recorded.
- Temp (Farenheit): Temperature in degrees farenheit measured within the experimental arena, when recorded.
- Breeding: Letter designation signifying whether the focal pair was breeding (i.e., had eggs and/or tadpoles; Y) or not (N) at the time when the trial was conducted.
The file Two Choice IDs.csv deciphers the code used for individual females across time blocks in each time block for the two-choice experiments. Note that repeated females did not necessarily retain the same coded letter ID across time blocks. Variables are as follows:
- Block: Time block (1 or 2) during which the trial was conducted. Blocks are spaced nine weeks apart and facilitate repeated measures of some individuals.
- Male ID: ID of male from which the stimulus call was recorded.
- Focal Male: Coded experimental (letter) ID given to each of the pair-bonded male partners from which the stimulus calls were recorded.
- Female ID: ID of the female subject ('Ri.####')
- Focal female: Coded experimental (letter) ID given to each of 9 female subjects. Each female letter designation mirrors the letter of her male partner, but in lower-case to distinguish.
The files No Choice Scoring Data.csv and Two Choice Scoring Data.csv contain the tabulated 'time spent' data obtained from scoring trial videos with the behavioral analysis software, BORIS. Trials were scored without audio, and the observer was blind to side of the arena from which the stimulus/stimuli were presented. Variables are as follows:
- Trial Number: A numeric designation given to the trial at the time of completion, in the order of completion.
- Block: Time block (1 or 2) during which the trial was conducted. Blocks are spaced nine weeks apart and facilitate repeated measures of some individuals.
- Observation Date: The date and time at which the analysis was performed.
- Description: An automatically generated field in BORIS, which we left blank.
- Media file: File path to the input video.
- Total length: Total length of the video, in seconds. Videos longer than 600 seconds were analyzed only up to the 600 s mark.
- FPS: Frames per second
- Focal Female: Coded experimental ID given to each of 9 subjects ('j1-9')
- Behavior: The location of the frog within the arena - either 'center' (central starting square), 'R' (right arm), or 'L' (left arm). All frogs start the trial in the center and can move freely into the left or right arm.
- Behavior Type: A categorical variable within BORIS to signify whether behaviors were instantaneous or continuous. All of our behaviors were recorded as 'State' behaviors, meaning once started they continued until the user specified 'stop'. The frog begins the trial in center by default, and with the first click the user records the time of first movement, which is always within center because the central starting square spans ~10 frog body lengths. Thus, the first two behaviors are always 'center,' whereas the third denotes the first time the frog crosses into the left or right arm. Each trial may have as few as 1-2 rows (meaning the frog never left center) or >2 rows (meaning the frogs explored the arena).
- Start (s): Time at which the focal behavior commenced, expressed in seconds into the video.
- Stop (s): Time at which the focal behavior ceased, expressed in seconds into the video.
- Duration (s): Total duration of the behavior, calculated as Stop (s) - Start (s).
Code/Software
All statistical analyses were performed in R Studio v. 2023.03.0+386 (R Core Development Team, 2020). Analyses are divided into two sections for the two experimental types: No Choice and Two Choice experiments. We first read in and merge associated trial info and blind scoring data and expand data frames to include rows for behaviors not observed. We then summarize time spent data for each trial by summing the total time spent in each behavior, or section of the arena, and recoding BORISÕs coded behaviors (center, L, R) in the context of the experiment: For No-Choice experiments, L and R are recoded as 'Toward Stimulus' and 'Away from Stimulus', contingent on the side of the arena from which the stimulus was presented. For Two-Choice experiments, L and R are recoded as 'Toward Mate' and 'Toward Stranger' contingent on the Trial Type, which designates whether the Mate stimulus was presented on the left or right side.
Following data wrangling steps, we compare total time spent on each side of the arena using Type III ANOVAs. Model testing was carried out using the packages 'lmerTest', with breeding status specified as covariates (for two-choice trials) and female ID, speaker position, and block (for two-choice trials) specified as random effects in all models. For no-choice experiments, we also ran a model including Male ID as a fixed effect.
For each trial type, we next evaluated the latency of subjects to exit center (if they exited center) and the overall temporal pattern of space use within the arena. These resulted in additional models including latency to approach as an independent and dependent variable (see script and methods for details of model structure).
This script requires the following packages: lme4, lmerTest, plyr, dplyr, stats, ggplot2, and ggpubr.
Methods
This was a laboratory investigation of phonotaxis behavior and individual vocal recognition in a pair-bonding poison frog. Frogs were bred in a laboratory colony and virgin and pair-bonded (breeding) female subjects were assayed in an experimental arena within a soundproof chamber while male advertisement calls (also recorded in the laboratory) were broadcast from speakers. Video recordings of trials were scored blind in the behavioral analysis software, BORIS, and we quantified and compared the time spent in each arm of the arena to evaluate (1) how virgin females respond to a single acoustic stimulus (whether they preferentially spend time near the speaker), and (2) whether pair-bonded females show a preference for the speaker broadcasting calls of their mate as opposed to calls of a stranger. All behavioral data were analyzed in R.