Skip to main content
Dryad

Evidence for maintenance of key components of vocal learning in aging budgerigars despite diminished affiliative social interaction

Cite this dataset

Moussaoui, Bushra et al. (2024). Evidence for maintenance of key components of vocal learning in aging budgerigars despite diminished affiliative social interaction [Dataset]. Dryad. https://doi.org/10.5061/dryad.fttdz08xr

Abstract

In some species, the ability to acquire new vocalizations persists into adulthood and may be an important mediator of social interactions. While it is generally assumed that vocal learning persists undiminished throughout the lifespan of these open-ended learners, the stability of this trait remains largely unexplored. We hypothesize that vocal learning exhibits senescence, as is typical of complex cognitive traits, and that this decline may relate to age-dependent changes in sociality. The budgerigar (Melopsittacus undulatus), an open-ended learner which develops new contact call types that are shared with social associates upon joining new flocks, provides a robust assay for measuring the effects of aging on vocal learning ability. We formed captive flocks of 4 previously unfamiliar adult males of the same age class, either “young adults” (6 mo.-1 yr.) or “older adults” (≥ 3 yr.), and concurrently tracked changes in contact call structure and social interactions over time. Older adults exhibited decreased vocal diversity, which may be related to the sparser and weaker affiliative bonds observed in older adults. Older adults, however, displayed equivalent levels of vocal plasticity and vocal convergence compared to young adults, suggesting vocal learning ability is largely maintained into later adulthood in an open-ended learner.

README: Title of Dataset: Evidence for persistent vocal learning in aging budgerigars (Melopsittacus undulatus) despite diminished social connectedness


This dataset contains vocal and social data from an experiment investigating the effect of aging on vocal production learning and social integration in flocks of male budgerigars (Melopsittacus undulatus). This open-ended learner develops new contact call types that are shared with social associates upon joining new social groups. We formed captive flocks of 4 previously unfamiliar adult male birds belonging to the same age class, either “young adults” (6 mo-1 yr) or “older adults” (≥ 3 yr), and concurrently tracked changes in contact call structure and affiliative, agonistic, and proximity-based social interactions over time. Contact calls were isolated from audio recordings using an automated selection procedure and they were characterized in two-dimensional 'acoustic space' based on various extracted spectrographic features to measure vocal output, change in vocal diversity, vocal plasticity, and vocal convergence. Video recordings of flock social behavior were coded for inter-individual proximity, and dyadic affiliative & agonistic interactions to build social networks and measure flock connectedness and strength of social ties.This dataset contains dataframes (.csv & .xlsx) of both vocal and social raw data and the corresponding code used to analyze this data in the form of R scripts(.R & .Rmd). A description of each individual file can be found below.

Description of the Data and file structure

Vocal Analysis Files:

Raw Data:
    - sp_selections_analysis.csv: Contact call selection tables including spectrographic measurements.
        - selec: number of the sound selection in the sequence of automatically detected sounds from each audio filemthat was visual confirmed to be a budgerigar contact call.
        - duration: length of call in seconds.
        - meanfreq: mean of frequency spectrum in kHz.
        - sd: standard deviation of frequency in kHz.
        - freq.median: median frequency in kHz.
        - freq.IQR: interquartile frequency range in kHz.
        - time.IQR: interquartile time range in seconds.
        - skew: skewness.
        - kurt: kurtosis.
        - sp.ent: spectral entropy (Pure tone is ~ 0; noisy is ~ 1).
        - time.ent: time entropy (Pure tone is ~ 0; noisy is ~ 1).
        - entropy: spectrographic entropy (product of time and spectral entropy).
        - meandom: average of dominant frequency in kHz.
        - mindom: minimum of dominant frequency in kHz. 
        - maxdom: maximum of dominant frequency in kHz.
        - dfrange: range of dominant frequency in kHz.
        - modindx: modulation index, computed as the cumulative absolute difference between adjacent measurements of dominant frequencies divided by dominant frequency range where a value of 1 indicates no modulation.
        - meanpeakf: mean peak frequency in kHz.
        - ID: bird individual ID, where the number indicates the experimental flock the bird was assigned to and the letter indicates the source population.
        - date: date of audio recording.
    - sp_selections_analysis_metadata.csv: Contact call selection tables including acoustic measurements and metadata.
        - see previous file description for definitions of each acoustic variable.
        - treatment: adult age class (either Young or Old).
        - round: experimental round (1, 2, or 3).
        - source.room: source room in which the individual bird was housed prior to novel flock formation (A, B, C, or D); each experimental flock is composed from 1 bird from each of the four separate source rooms.
        - record.group: experimental flock (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12).
        - record.block: recording block corresponding to the date of audio recording. There were 5 audio-recording blocks (block 1 contains recordings of birds' baseline repertoire and blocks 2-5 contain recordings of birds' while in their novel flocks). Blocks were 4 days long and separated by a break day.

Scripts:
    - optimized_auto_detec_BM.R: R code for autodetection of contact calls from collected audio files. Autodetection was run in batched on audio files grouped by experimental flock for computing efficiency. Autodetected calls were then visually confirmed in Raven Pro before spectrographic measurements.
    - spectrogram_analysis_prep_BM.R: R code for spectrographic measurements of final contact calls. 
    - Spectrogram_analysis_BM.Rmd: R markdown code for visualizing spectrographic measurements of contact calls & corresponding statistical analysis.
    - Call_analysis_BM_VSCODE.Rmd: R markdown code for projecting contact calls in 'acoustic space' and computing vocal learning measurements.
    - CallAnalysis_STATS.r: R code for statistics of vocal learning measurements.

Social Analysis Files:

Raw Data:
    - Social_Interaction_Video_Examples.mp4 : Annotated video of examples of the types of budgerigar affiliative and agonistic social interactions measured in the study.
    - BehaviorCodingRawData_EarlyMorning.csv : Raw data file for social interaction coding from collected videos used to generate affiliative and agonistic social networks and metrics.
        - Treatment: adult age class (either Young or Old).
        - Group: experimental flock (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12).
        - Video date: date of video-recording (month/day/year).
        - Time of day: time of day at which video-recording took place. This data set only contains recordings taken during "Early Morning" defined as 0700-0900 since birds were most active at this time.
        - Actor: individual bird that initiated the social interaction.
        - Recipient: individual bird that the interaction was directed towards.
        - Behavior: specific social interaction that was observed.
            - KS: kiss (affiliative)
            - HB: head bob (affiliative)
            - AF: allofeed (affiliative)
            - AP: allopreen (affiliative)
            - BT: bill thrust (agonistic)
            - KI: kick (agonistic)
            - KS-RJ: rejected kiss (agonistic)
            - HB-RJ: rehected head bob (agonistic)
            - AF-RJ: rejected allofeed (agonistic)
            - AP-RJ: rejected allopreen (agonistic)
        - Start time: time in the video recording at which the social interaction began.
        - End time: time in the video recording at which the social interaction ended.
        - Duration: duration of the social interaction in h:mm:ss (for allofeeding and allopreening interactions only).
    - BehaviorCodingRawData_EarlyMorning_ForTotals.csv : Raw data file identical to as above but including individual social interactions categorized as affiliative or agonistic and including recording block metadata.
        - see previous file description for definitions of other variables.
        - Category: AFF for affiliative or AGO for agonistic.
        - Block: recording block corresponding to the date of video recording. There were 4 video-recording blocks. Blocks were 4 days long and separated by a break day.
        - Half: recordings classified as being taken during the first or second half of the experimental round. Half 1 contains recordings from blocks 1 and 2 and half 2 contains recordings from blocks 3 and 4.
    - Proximity Raw Data_R.csv: Raw data file for proximity coding from collected videos used to generate proximity social networks and metrics.
        - Treatment: adult age class (either Young or Old).
        - Group: experimental flock (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12).
        - Video_date: date of video-recording (day-month-year).
        - Snapshot_#: snapshot number; snapshots were taken at 4-minute intervals starting from 30 minutes after the experimenter left the room and for a duration of 40 minutes, yielding 11 snapshots per flock for each daily video recording.
        - Neighbor_1 and Neighbor_2: the two individual birds between which the inter-individual distance has been measured in cm; distance was measured using an in-frame ruler as a scale and the line tool in Image J.
        - block: recording block corresponding to the date of video recording. There were 4 video-recording blocks. Blocks were 4 days long and separated by a break day.
        - half: recordings classified as being taken during the first or second half of the experimental round. Half 1 contains recordings from blocks 1 and 2 and half 2 contains recordings from blocks 3 and 4.
    - Interobserver Assessment.xlsx: Raw data for interobserver assessment of social interaction coding.
        - Sheet 1 (BM Recoding): Recoding by BM of a subset of videos coded by observer SO for assessing interobserver reliability.
            - see description of variable definitions for the file BehaviorCodingRawData_EarlyMorning.csv.
        - Sheet 2 (BM Recoded Videos): checklist of videos that were recoded.
        - Sheet 3 (Total Affiliative): comparison of affiliative behavior coding between observers SO and BM.
            - SO: total number of affiliative behaviors coded by SO.
            - BM: total number of affiliative behaviors coded by BM.
            - BM_AP: total number of allopreening interactions coded by BM.
            - SO_AP: total number of allopreening interactions coded by SO.
            - BM_AF: total number of allofeeding interactions coded by BM.
            - SO_AF: total number of allofeeding interactions coded by SO.
            - BM_KS: total number of kiss interactions coded by BM.
            - SO_KS: total number of kiss interactions coded by SO.
            - BM_HB: total number of head bob interactions coded by BM.
            - SO_HB: total number of head bob interactions coded by SO.
        - Sheet 4 (Total Agonistic): comparison of agonistic behavior coding between observers SO and BM.
            - SO: total number of agonistic behaviors coded by SO.
            - BM: total number of agonistic behaviors coded by BM.
            - BM_BT: total number of bill thrust interactions coded by BM.
            - SO_BT: total number of bill thrust interactions coded by SO.
            - BM_KI: total number of kick interactions coded by BM.
            - SO_KI: total number of kick interactions coded by SO.
            - BM_KS-RJ: total number of rejected kiss interactions coded by BM.
            - SO_KS-RJ: total number of rejected kiss interactions coded by SO.
            - BM_HB-RJ: total number of rejected head bob interactions coded by BM.
            - SO_HB-RJ: total number of rejected head bob interactions coded by SO.
            - BM_AP-RJ: total number of rejected allopreening interactions coded by BM.
            - SO_AP-RJ: total number of rejected allopreening interactions coded by SO.
            - BM_AF-RJ: total number of rejected allofeeding interactions coded by BM.
            - SO_AF-RJ: total number of rejected allofeeding interactions coded by SO.
    - Interobsv_SO_BM_PooledAffiliativeAgonistic.csv:  Pooled raw data for interobserver assessment of social interaction coding.
        - Video: Video file with a prefix of either AFF for affiliative or AGO for agonistic to indicated which category of behavior the totals in the following two columns correspond to.
        - SO: total number of interactions coded by observer SO.
        - BM: total number of interactions coded by observer BM.
    - Edgelists of affiliative interactions with one edgelist per social flock per recording block:
        - These files are named using the following format: aff_group#_edgelist_recordingblock#.csv ((File example: aff_g1_edgelist_block1.csv))
            - the files with the suffix "overall" rather than a recording block number, contain all affiliative interactions pooled over the four recording blocks
        - Actor: individual ID of bird initiating an observed affiliative interaction
        - Recipient: individual ID of bird receiving an observed affiliative interaction
    - Edgelists of agonistic interactions with one edgelist per social group per recording block:
        - These files are named using the following format: ago_group#_edgelist_recordingblock#.csv ((File example: ago_g1_edgelist_block1.csv))
          - the files with the suffix "overall" rather than a recording block number, contain all agonistic interactions pooled over the four recording blocks
        - Actor: individual ID of bird initiating an observed agonistic interaction
        - Recipient: individual ID of bird receiving an observed agonistic interaction

Scripts:
    - aff_socialnetworkmetrics.R: R code for generating affiliative social networks and metrics.
        - aff_socialnetwork_stats.R : R code for statistical analysis of affiliative social metrics.
    - ago_socialnetworkmetrics.R : R code for generating agonistic social networks and metrics.
        - ago_socialnetwork_stats.R: R code for statistical analysis of agonistic social metrics.
    - Interobsv_SO_BM.R: R code for performing interobserver reliability assessment of social interaction coding between two observers. 
    - Proximity_Proportional.R: R code for generating proximity social networks, metrics, and statistics.
    - TotalInteractionSummary.R: R code for summarizing social interaction totals.

Usage notes

This dataset contains dataframes (.csv & .xlsx) of both vocal and social raw data and the corresponding code used in analyzing this data in the form of R scripts (.R & .Rmd). A description of the contents of each data or script file can be found in the README file provided. The dataset also contains example video recordings of the types of budgerigar affiliative and agonistic social interactions measured in the study.

Funding

Animal Behavior Society, Award: Student Research Grant

National Institute of General Medical Sciences, Award: P20GM103451

National Institute of Neurological Disorders and Stroke, Award: 1R21NS126079